Tue, 16 Aug 2011 11:53:57 -0700
7079317: Incorrect branch's destination block in PrintoOptoAssembly output
Summary: save/restore label and block in scratch_emit_size()
Reviewed-by: never
1 /*
2 * Copyright (c) 1997, 2011, Oracle and/or its affiliates. All rights reserved.
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4 *
5 * This code is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU General Public License version 2 only, as
7 * published by the Free Software Foundation.
8 *
9 * This code is distributed in the hope that it will be useful, but WITHOUT
10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
12 * version 2 for more details (a copy is included in the LICENSE file that
13 * accompanied this code).
14 *
15 * You should have received a copy of the GNU General Public License version
16 * 2 along with this work; if not, write to the Free Software Foundation,
17 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
18 *
19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
20 * or visit www.oracle.com if you need additional information or have any
21 * questions.
22 *
23 */
25 #include "precompiled.hpp"
26 #include "asm/assembler.hpp"
27 #include "classfile/systemDictionary.hpp"
28 #include "code/exceptionHandlerTable.hpp"
29 #include "code/nmethod.hpp"
30 #include "compiler/compileLog.hpp"
31 #include "compiler/oopMap.hpp"
32 #include "opto/addnode.hpp"
33 #include "opto/block.hpp"
34 #include "opto/c2compiler.hpp"
35 #include "opto/callGenerator.hpp"
36 #include "opto/callnode.hpp"
37 #include "opto/cfgnode.hpp"
38 #include "opto/chaitin.hpp"
39 #include "opto/compile.hpp"
40 #include "opto/connode.hpp"
41 #include "opto/divnode.hpp"
42 #include "opto/escape.hpp"
43 #include "opto/idealGraphPrinter.hpp"
44 #include "opto/loopnode.hpp"
45 #include "opto/machnode.hpp"
46 #include "opto/macro.hpp"
47 #include "opto/matcher.hpp"
48 #include "opto/memnode.hpp"
49 #include "opto/mulnode.hpp"
50 #include "opto/node.hpp"
51 #include "opto/opcodes.hpp"
52 #include "opto/output.hpp"
53 #include "opto/parse.hpp"
54 #include "opto/phaseX.hpp"
55 #include "opto/rootnode.hpp"
56 #include "opto/runtime.hpp"
57 #include "opto/stringopts.hpp"
58 #include "opto/type.hpp"
59 #include "opto/vectornode.hpp"
60 #include "runtime/arguments.hpp"
61 #include "runtime/signature.hpp"
62 #include "runtime/stubRoutines.hpp"
63 #include "runtime/timer.hpp"
64 #include "utilities/copy.hpp"
65 #ifdef TARGET_ARCH_MODEL_x86_32
66 # include "adfiles/ad_x86_32.hpp"
67 #endif
68 #ifdef TARGET_ARCH_MODEL_x86_64
69 # include "adfiles/ad_x86_64.hpp"
70 #endif
71 #ifdef TARGET_ARCH_MODEL_sparc
72 # include "adfiles/ad_sparc.hpp"
73 #endif
74 #ifdef TARGET_ARCH_MODEL_zero
75 # include "adfiles/ad_zero.hpp"
76 #endif
77 #ifdef TARGET_ARCH_MODEL_arm
78 # include "adfiles/ad_arm.hpp"
79 #endif
80 #ifdef TARGET_ARCH_MODEL_ppc
81 # include "adfiles/ad_ppc.hpp"
82 #endif
85 // -------------------- Compile::mach_constant_base_node -----------------------
86 // Constant table base node singleton.
87 MachConstantBaseNode* Compile::mach_constant_base_node() {
88 if (_mach_constant_base_node == NULL) {
89 _mach_constant_base_node = new (C) MachConstantBaseNode();
90 _mach_constant_base_node->add_req(C->root());
91 }
92 return _mach_constant_base_node;
93 }
96 /// Support for intrinsics.
98 // Return the index at which m must be inserted (or already exists).
99 // The sort order is by the address of the ciMethod, with is_virtual as minor key.
100 int Compile::intrinsic_insertion_index(ciMethod* m, bool is_virtual) {
101 #ifdef ASSERT
102 for (int i = 1; i < _intrinsics->length(); i++) {
103 CallGenerator* cg1 = _intrinsics->at(i-1);
104 CallGenerator* cg2 = _intrinsics->at(i);
105 assert(cg1->method() != cg2->method()
106 ? cg1->method() < cg2->method()
107 : cg1->is_virtual() < cg2->is_virtual(),
108 "compiler intrinsics list must stay sorted");
109 }
110 #endif
111 // Binary search sorted list, in decreasing intervals [lo, hi].
112 int lo = 0, hi = _intrinsics->length()-1;
113 while (lo <= hi) {
114 int mid = (uint)(hi + lo) / 2;
115 ciMethod* mid_m = _intrinsics->at(mid)->method();
116 if (m < mid_m) {
117 hi = mid-1;
118 } else if (m > mid_m) {
119 lo = mid+1;
120 } else {
121 // look at minor sort key
122 bool mid_virt = _intrinsics->at(mid)->is_virtual();
123 if (is_virtual < mid_virt) {
124 hi = mid-1;
125 } else if (is_virtual > mid_virt) {
126 lo = mid+1;
127 } else {
128 return mid; // exact match
129 }
130 }
131 }
132 return lo; // inexact match
133 }
135 void Compile::register_intrinsic(CallGenerator* cg) {
136 if (_intrinsics == NULL) {
137 _intrinsics = new GrowableArray<CallGenerator*>(60);
138 }
139 // This code is stolen from ciObjectFactory::insert.
140 // Really, GrowableArray should have methods for
141 // insert_at, remove_at, and binary_search.
142 int len = _intrinsics->length();
143 int index = intrinsic_insertion_index(cg->method(), cg->is_virtual());
144 if (index == len) {
145 _intrinsics->append(cg);
146 } else {
147 #ifdef ASSERT
148 CallGenerator* oldcg = _intrinsics->at(index);
149 assert(oldcg->method() != cg->method() || oldcg->is_virtual() != cg->is_virtual(), "don't register twice");
150 #endif
151 _intrinsics->append(_intrinsics->at(len-1));
152 int pos;
153 for (pos = len-2; pos >= index; pos--) {
154 _intrinsics->at_put(pos+1,_intrinsics->at(pos));
155 }
156 _intrinsics->at_put(index, cg);
157 }
158 assert(find_intrinsic(cg->method(), cg->is_virtual()) == cg, "registration worked");
159 }
161 CallGenerator* Compile::find_intrinsic(ciMethod* m, bool is_virtual) {
162 assert(m->is_loaded(), "don't try this on unloaded methods");
163 if (_intrinsics != NULL) {
164 int index = intrinsic_insertion_index(m, is_virtual);
165 if (index < _intrinsics->length()
166 && _intrinsics->at(index)->method() == m
167 && _intrinsics->at(index)->is_virtual() == is_virtual) {
168 return _intrinsics->at(index);
169 }
170 }
171 // Lazily create intrinsics for intrinsic IDs well-known in the runtime.
172 if (m->intrinsic_id() != vmIntrinsics::_none &&
173 m->intrinsic_id() <= vmIntrinsics::LAST_COMPILER_INLINE) {
174 CallGenerator* cg = make_vm_intrinsic(m, is_virtual);
175 if (cg != NULL) {
176 // Save it for next time:
177 register_intrinsic(cg);
178 return cg;
179 } else {
180 gather_intrinsic_statistics(m->intrinsic_id(), is_virtual, _intrinsic_disabled);
181 }
182 }
183 return NULL;
184 }
186 // Compile:: register_library_intrinsics and make_vm_intrinsic are defined
187 // in library_call.cpp.
190 #ifndef PRODUCT
191 // statistics gathering...
193 juint Compile::_intrinsic_hist_count[vmIntrinsics::ID_LIMIT] = {0};
194 jubyte Compile::_intrinsic_hist_flags[vmIntrinsics::ID_LIMIT] = {0};
196 bool Compile::gather_intrinsic_statistics(vmIntrinsics::ID id, bool is_virtual, int flags) {
197 assert(id > vmIntrinsics::_none && id < vmIntrinsics::ID_LIMIT, "oob");
198 int oflags = _intrinsic_hist_flags[id];
199 assert(flags != 0, "what happened?");
200 if (is_virtual) {
201 flags |= _intrinsic_virtual;
202 }
203 bool changed = (flags != oflags);
204 if ((flags & _intrinsic_worked) != 0) {
205 juint count = (_intrinsic_hist_count[id] += 1);
206 if (count == 1) {
207 changed = true; // first time
208 }
209 // increment the overall count also:
210 _intrinsic_hist_count[vmIntrinsics::_none] += 1;
211 }
212 if (changed) {
213 if (((oflags ^ flags) & _intrinsic_virtual) != 0) {
214 // Something changed about the intrinsic's virtuality.
215 if ((flags & _intrinsic_virtual) != 0) {
216 // This is the first use of this intrinsic as a virtual call.
217 if (oflags != 0) {
218 // We already saw it as a non-virtual, so note both cases.
219 flags |= _intrinsic_both;
220 }
221 } else if ((oflags & _intrinsic_both) == 0) {
222 // This is the first use of this intrinsic as a non-virtual
223 flags |= _intrinsic_both;
224 }
225 }
226 _intrinsic_hist_flags[id] = (jubyte) (oflags | flags);
227 }
228 // update the overall flags also:
229 _intrinsic_hist_flags[vmIntrinsics::_none] |= (jubyte) flags;
230 return changed;
231 }
233 static char* format_flags(int flags, char* buf) {
234 buf[0] = 0;
235 if ((flags & Compile::_intrinsic_worked) != 0) strcat(buf, ",worked");
236 if ((flags & Compile::_intrinsic_failed) != 0) strcat(buf, ",failed");
237 if ((flags & Compile::_intrinsic_disabled) != 0) strcat(buf, ",disabled");
238 if ((flags & Compile::_intrinsic_virtual) != 0) strcat(buf, ",virtual");
239 if ((flags & Compile::_intrinsic_both) != 0) strcat(buf, ",nonvirtual");
240 if (buf[0] == 0) strcat(buf, ",");
241 assert(buf[0] == ',', "must be");
242 return &buf[1];
243 }
245 void Compile::print_intrinsic_statistics() {
246 char flagsbuf[100];
247 ttyLocker ttyl;
248 if (xtty != NULL) xtty->head("statistics type='intrinsic'");
249 tty->print_cr("Compiler intrinsic usage:");
250 juint total = _intrinsic_hist_count[vmIntrinsics::_none];
251 if (total == 0) total = 1; // avoid div0 in case of no successes
252 #define PRINT_STAT_LINE(name, c, f) \
253 tty->print_cr(" %4d (%4.1f%%) %s (%s)", (int)(c), ((c) * 100.0) / total, name, f);
254 for (int index = 1 + (int)vmIntrinsics::_none; index < (int)vmIntrinsics::ID_LIMIT; index++) {
255 vmIntrinsics::ID id = (vmIntrinsics::ID) index;
256 int flags = _intrinsic_hist_flags[id];
257 juint count = _intrinsic_hist_count[id];
258 if ((flags | count) != 0) {
259 PRINT_STAT_LINE(vmIntrinsics::name_at(id), count, format_flags(flags, flagsbuf));
260 }
261 }
262 PRINT_STAT_LINE("total", total, format_flags(_intrinsic_hist_flags[vmIntrinsics::_none], flagsbuf));
263 if (xtty != NULL) xtty->tail("statistics");
264 }
266 void Compile::print_statistics() {
267 { ttyLocker ttyl;
268 if (xtty != NULL) xtty->head("statistics type='opto'");
269 Parse::print_statistics();
270 PhaseCCP::print_statistics();
271 PhaseRegAlloc::print_statistics();
272 Scheduling::print_statistics();
273 PhasePeephole::print_statistics();
274 PhaseIdealLoop::print_statistics();
275 if (xtty != NULL) xtty->tail("statistics");
276 }
277 if (_intrinsic_hist_flags[vmIntrinsics::_none] != 0) {
278 // put this under its own <statistics> element.
279 print_intrinsic_statistics();
280 }
281 }
282 #endif //PRODUCT
284 // Support for bundling info
285 Bundle* Compile::node_bundling(const Node *n) {
286 assert(valid_bundle_info(n), "oob");
287 return &_node_bundling_base[n->_idx];
288 }
290 bool Compile::valid_bundle_info(const Node *n) {
291 return (_node_bundling_limit > n->_idx);
292 }
295 void Compile::gvn_replace_by(Node* n, Node* nn) {
296 for (DUIterator_Last imin, i = n->last_outs(imin); i >= imin; ) {
297 Node* use = n->last_out(i);
298 bool is_in_table = initial_gvn()->hash_delete(use);
299 uint uses_found = 0;
300 for (uint j = 0; j < use->len(); j++) {
301 if (use->in(j) == n) {
302 if (j < use->req())
303 use->set_req(j, nn);
304 else
305 use->set_prec(j, nn);
306 uses_found++;
307 }
308 }
309 if (is_in_table) {
310 // reinsert into table
311 initial_gvn()->hash_find_insert(use);
312 }
313 record_for_igvn(use);
314 i -= uses_found; // we deleted 1 or more copies of this edge
315 }
316 }
321 // Identify all nodes that are reachable from below, useful.
322 // Use breadth-first pass that records state in a Unique_Node_List,
323 // recursive traversal is slower.
324 void Compile::identify_useful_nodes(Unique_Node_List &useful) {
325 int estimated_worklist_size = unique();
326 useful.map( estimated_worklist_size, NULL ); // preallocate space
328 // Initialize worklist
329 if (root() != NULL) { useful.push(root()); }
330 // If 'top' is cached, declare it useful to preserve cached node
331 if( cached_top_node() ) { useful.push(cached_top_node()); }
333 // Push all useful nodes onto the list, breadthfirst
334 for( uint next = 0; next < useful.size(); ++next ) {
335 assert( next < unique(), "Unique useful nodes < total nodes");
336 Node *n = useful.at(next);
337 uint max = n->len();
338 for( uint i = 0; i < max; ++i ) {
339 Node *m = n->in(i);
340 if( m == NULL ) continue;
341 useful.push(m);
342 }
343 }
344 }
346 // Disconnect all useless nodes by disconnecting those at the boundary.
347 void Compile::remove_useless_nodes(Unique_Node_List &useful) {
348 uint next = 0;
349 while( next < useful.size() ) {
350 Node *n = useful.at(next++);
351 // Use raw traversal of out edges since this code removes out edges
352 int max = n->outcnt();
353 for (int j = 0; j < max; ++j ) {
354 Node* child = n->raw_out(j);
355 if( ! useful.member(child) ) {
356 assert( !child->is_top() || child != top(),
357 "If top is cached in Compile object it is in useful list");
358 // Only need to remove this out-edge to the useless node
359 n->raw_del_out(j);
360 --j;
361 --max;
362 }
363 }
364 if (n->outcnt() == 1 && n->has_special_unique_user()) {
365 record_for_igvn( n->unique_out() );
366 }
367 }
368 debug_only(verify_graph_edges(true/*check for no_dead_code*/);)
369 }
371 //------------------------------frame_size_in_words-----------------------------
372 // frame_slots in units of words
373 int Compile::frame_size_in_words() const {
374 // shift is 0 in LP32 and 1 in LP64
375 const int shift = (LogBytesPerWord - LogBytesPerInt);
376 int words = _frame_slots >> shift;
377 assert( words << shift == _frame_slots, "frame size must be properly aligned in LP64" );
378 return words;
379 }
381 // ============================================================================
382 //------------------------------CompileWrapper---------------------------------
383 class CompileWrapper : public StackObj {
384 Compile *const _compile;
385 public:
386 CompileWrapper(Compile* compile);
388 ~CompileWrapper();
389 };
391 CompileWrapper::CompileWrapper(Compile* compile) : _compile(compile) {
392 // the Compile* pointer is stored in the current ciEnv:
393 ciEnv* env = compile->env();
394 assert(env == ciEnv::current(), "must already be a ciEnv active");
395 assert(env->compiler_data() == NULL, "compile already active?");
396 env->set_compiler_data(compile);
397 assert(compile == Compile::current(), "sanity");
399 compile->set_type_dict(NULL);
400 compile->set_type_hwm(NULL);
401 compile->set_type_last_size(0);
402 compile->set_last_tf(NULL, NULL);
403 compile->set_indexSet_arena(NULL);
404 compile->set_indexSet_free_block_list(NULL);
405 compile->init_type_arena();
406 Type::Initialize(compile);
407 _compile->set_scratch_buffer_blob(NULL);
408 _compile->begin_method();
409 }
410 CompileWrapper::~CompileWrapper() {
411 _compile->end_method();
412 if (_compile->scratch_buffer_blob() != NULL)
413 BufferBlob::free(_compile->scratch_buffer_blob());
414 _compile->env()->set_compiler_data(NULL);
415 }
418 //----------------------------print_compile_messages---------------------------
419 void Compile::print_compile_messages() {
420 #ifndef PRODUCT
421 // Check if recompiling
422 if (_subsume_loads == false && PrintOpto) {
423 // Recompiling without allowing machine instructions to subsume loads
424 tty->print_cr("*********************************************************");
425 tty->print_cr("** Bailout: Recompile without subsuming loads **");
426 tty->print_cr("*********************************************************");
427 }
428 if (_do_escape_analysis != DoEscapeAnalysis && PrintOpto) {
429 // Recompiling without escape analysis
430 tty->print_cr("*********************************************************");
431 tty->print_cr("** Bailout: Recompile without escape analysis **");
432 tty->print_cr("*********************************************************");
433 }
434 if (env()->break_at_compile()) {
435 // Open the debugger when compiling this method.
436 tty->print("### Breaking when compiling: ");
437 method()->print_short_name();
438 tty->cr();
439 BREAKPOINT;
440 }
442 if( PrintOpto ) {
443 if (is_osr_compilation()) {
444 tty->print("[OSR]%3d", _compile_id);
445 } else {
446 tty->print("%3d", _compile_id);
447 }
448 }
449 #endif
450 }
453 //-----------------------init_scratch_buffer_blob------------------------------
454 // Construct a temporary BufferBlob and cache it for this compile.
455 void Compile::init_scratch_buffer_blob(int const_size) {
456 // If there is already a scratch buffer blob allocated and the
457 // constant section is big enough, use it. Otherwise free the
458 // current and allocate a new one.
459 BufferBlob* blob = scratch_buffer_blob();
460 if ((blob != NULL) && (const_size <= _scratch_const_size)) {
461 // Use the current blob.
462 } else {
463 if (blob != NULL) {
464 BufferBlob::free(blob);
465 }
467 ResourceMark rm;
468 _scratch_const_size = const_size;
469 int size = (MAX_inst_size + MAX_stubs_size + _scratch_const_size);
470 blob = BufferBlob::create("Compile::scratch_buffer", size);
471 // Record the buffer blob for next time.
472 set_scratch_buffer_blob(blob);
473 // Have we run out of code space?
474 if (scratch_buffer_blob() == NULL) {
475 // Let CompilerBroker disable further compilations.
476 record_failure("Not enough space for scratch buffer in CodeCache");
477 return;
478 }
479 }
481 // Initialize the relocation buffers
482 relocInfo* locs_buf = (relocInfo*) blob->content_end() - MAX_locs_size;
483 set_scratch_locs_memory(locs_buf);
484 }
487 //-----------------------scratch_emit_size-------------------------------------
488 // Helper function that computes size by emitting code
489 uint Compile::scratch_emit_size(const Node* n) {
490 // Start scratch_emit_size section.
491 set_in_scratch_emit_size(true);
493 // Emit into a trash buffer and count bytes emitted.
494 // This is a pretty expensive way to compute a size,
495 // but it works well enough if seldom used.
496 // All common fixed-size instructions are given a size
497 // method by the AD file.
498 // Note that the scratch buffer blob and locs memory are
499 // allocated at the beginning of the compile task, and
500 // may be shared by several calls to scratch_emit_size.
501 // The allocation of the scratch buffer blob is particularly
502 // expensive, since it has to grab the code cache lock.
503 BufferBlob* blob = this->scratch_buffer_blob();
504 assert(blob != NULL, "Initialize BufferBlob at start");
505 assert(blob->size() > MAX_inst_size, "sanity");
506 relocInfo* locs_buf = scratch_locs_memory();
507 address blob_begin = blob->content_begin();
508 address blob_end = (address)locs_buf;
509 assert(blob->content_contains(blob_end), "sanity");
510 CodeBuffer buf(blob_begin, blob_end - blob_begin);
511 buf.initialize_consts_size(_scratch_const_size);
512 buf.initialize_stubs_size(MAX_stubs_size);
513 assert(locs_buf != NULL, "sanity");
514 int lsize = MAX_locs_size / 3;
515 buf.consts()->initialize_shared_locs(&locs_buf[lsize * 0], lsize);
516 buf.insts()->initialize_shared_locs( &locs_buf[lsize * 1], lsize);
517 buf.stubs()->initialize_shared_locs( &locs_buf[lsize * 2], lsize);
519 // Do the emission.
521 Label fakeL; // Fake label for branch instructions.
522 Label* saveL = NULL;
523 uint save_bnum = 0;
524 bool is_branch = n->is_MachBranch();
525 if (is_branch) {
526 MacroAssembler masm(&buf);
527 masm.bind(fakeL);
528 n->as_MachBranch()->save_label(&saveL, &save_bnum);
529 n->as_MachBranch()->label_set(&fakeL, 0);
530 }
531 n->emit(buf, this->regalloc());
532 if (is_branch) // Restore label.
533 n->as_MachBranch()->label_set(saveL, save_bnum);
535 // End scratch_emit_size section.
536 set_in_scratch_emit_size(false);
538 return buf.insts_size();
539 }
542 // ============================================================================
543 //------------------------------Compile standard-------------------------------
544 debug_only( int Compile::_debug_idx = 100000; )
546 // Compile a method. entry_bci is -1 for normal compilations and indicates
547 // the continuation bci for on stack replacement.
550 Compile::Compile( ciEnv* ci_env, C2Compiler* compiler, ciMethod* target, int osr_bci, bool subsume_loads, bool do_escape_analysis )
551 : Phase(Compiler),
552 _env(ci_env),
553 _log(ci_env->log()),
554 _compile_id(ci_env->compile_id()),
555 _save_argument_registers(false),
556 _stub_name(NULL),
557 _stub_function(NULL),
558 _stub_entry_point(NULL),
559 _method(target),
560 _entry_bci(osr_bci),
561 _initial_gvn(NULL),
562 _for_igvn(NULL),
563 _warm_calls(NULL),
564 _subsume_loads(subsume_loads),
565 _do_escape_analysis(do_escape_analysis),
566 _failure_reason(NULL),
567 _code_buffer("Compile::Fill_buffer"),
568 _orig_pc_slot(0),
569 _orig_pc_slot_offset_in_bytes(0),
570 _has_method_handle_invokes(false),
571 _mach_constant_base_node(NULL),
572 _node_bundling_limit(0),
573 _node_bundling_base(NULL),
574 _java_calls(0),
575 _inner_loops(0),
576 _scratch_const_size(-1),
577 _in_scratch_emit_size(false),
578 #ifndef PRODUCT
579 _trace_opto_output(TraceOptoOutput || method()->has_option("TraceOptoOutput")),
580 _printer(IdealGraphPrinter::printer()),
581 #endif
582 _congraph(NULL) {
583 C = this;
585 CompileWrapper cw(this);
586 #ifndef PRODUCT
587 if (TimeCompiler2) {
588 tty->print(" ");
589 target->holder()->name()->print();
590 tty->print(".");
591 target->print_short_name();
592 tty->print(" ");
593 }
594 TraceTime t1("Total compilation time", &_t_totalCompilation, TimeCompiler, TimeCompiler2);
595 TraceTime t2(NULL, &_t_methodCompilation, TimeCompiler, false);
596 bool print_opto_assembly = PrintOptoAssembly || _method->has_option("PrintOptoAssembly");
597 if (!print_opto_assembly) {
598 bool print_assembly = (PrintAssembly || _method->should_print_assembly());
599 if (print_assembly && !Disassembler::can_decode()) {
600 tty->print_cr("PrintAssembly request changed to PrintOptoAssembly");
601 print_opto_assembly = true;
602 }
603 }
604 set_print_assembly(print_opto_assembly);
605 set_parsed_irreducible_loop(false);
606 #endif
608 if (ProfileTraps) {
609 // Make sure the method being compiled gets its own MDO,
610 // so we can at least track the decompile_count().
611 method()->ensure_method_data();
612 }
614 Init(::AliasLevel);
617 print_compile_messages();
619 if (UseOldInlining || PrintCompilation NOT_PRODUCT( || PrintOpto) )
620 _ilt = InlineTree::build_inline_tree_root();
621 else
622 _ilt = NULL;
624 // Even if NO memory addresses are used, MergeMem nodes must have at least 1 slice
625 assert(num_alias_types() >= AliasIdxRaw, "");
627 #define MINIMUM_NODE_HASH 1023
628 // Node list that Iterative GVN will start with
629 Unique_Node_List for_igvn(comp_arena());
630 set_for_igvn(&for_igvn);
632 // GVN that will be run immediately on new nodes
633 uint estimated_size = method()->code_size()*4+64;
634 estimated_size = (estimated_size < MINIMUM_NODE_HASH ? MINIMUM_NODE_HASH : estimated_size);
635 PhaseGVN gvn(node_arena(), estimated_size);
636 set_initial_gvn(&gvn);
638 { // Scope for timing the parser
639 TracePhase t3("parse", &_t_parser, true);
641 // Put top into the hash table ASAP.
642 initial_gvn()->transform_no_reclaim(top());
644 // Set up tf(), start(), and find a CallGenerator.
645 CallGenerator* cg = NULL;
646 if (is_osr_compilation()) {
647 const TypeTuple *domain = StartOSRNode::osr_domain();
648 const TypeTuple *range = TypeTuple::make_range(method()->signature());
649 init_tf(TypeFunc::make(domain, range));
650 StartNode* s = new (this, 2) StartOSRNode(root(), domain);
651 initial_gvn()->set_type_bottom(s);
652 init_start(s);
653 cg = CallGenerator::for_osr(method(), entry_bci());
654 } else {
655 // Normal case.
656 init_tf(TypeFunc::make(method()));
657 StartNode* s = new (this, 2) StartNode(root(), tf()->domain());
658 initial_gvn()->set_type_bottom(s);
659 init_start(s);
660 if (method()->intrinsic_id() == vmIntrinsics::_Reference_get && UseG1GC) {
661 // With java.lang.ref.reference.get() we must go through the
662 // intrinsic when G1 is enabled - even when get() is the root
663 // method of the compile - so that, if necessary, the value in
664 // the referent field of the reference object gets recorded by
665 // the pre-barrier code.
666 // Specifically, if G1 is enabled, the value in the referent
667 // field is recorded by the G1 SATB pre barrier. This will
668 // result in the referent being marked live and the reference
669 // object removed from the list of discovered references during
670 // reference processing.
671 cg = find_intrinsic(method(), false);
672 }
673 if (cg == NULL) {
674 float past_uses = method()->interpreter_invocation_count();
675 float expected_uses = past_uses;
676 cg = CallGenerator::for_inline(method(), expected_uses);
677 }
678 }
679 if (failing()) return;
680 if (cg == NULL) {
681 record_method_not_compilable_all_tiers("cannot parse method");
682 return;
683 }
684 JVMState* jvms = build_start_state(start(), tf());
685 if ((jvms = cg->generate(jvms)) == NULL) {
686 record_method_not_compilable("method parse failed");
687 return;
688 }
689 GraphKit kit(jvms);
691 if (!kit.stopped()) {
692 // Accept return values, and transfer control we know not where.
693 // This is done by a special, unique ReturnNode bound to root.
694 return_values(kit.jvms());
695 }
697 if (kit.has_exceptions()) {
698 // Any exceptions that escape from this call must be rethrown
699 // to whatever caller is dynamically above us on the stack.
700 // This is done by a special, unique RethrowNode bound to root.
701 rethrow_exceptions(kit.transfer_exceptions_into_jvms());
702 }
704 if (!failing() && has_stringbuilder()) {
705 {
706 // remove useless nodes to make the usage analysis simpler
707 ResourceMark rm;
708 PhaseRemoveUseless pru(initial_gvn(), &for_igvn);
709 }
711 {
712 ResourceMark rm;
713 print_method("Before StringOpts", 3);
714 PhaseStringOpts pso(initial_gvn(), &for_igvn);
715 print_method("After StringOpts", 3);
716 }
718 // now inline anything that we skipped the first time around
719 while (_late_inlines.length() > 0) {
720 CallGenerator* cg = _late_inlines.pop();
721 cg->do_late_inline();
722 }
723 }
724 assert(_late_inlines.length() == 0, "should have been processed");
726 print_method("Before RemoveUseless", 3);
728 // Remove clutter produced by parsing.
729 if (!failing()) {
730 ResourceMark rm;
731 PhaseRemoveUseless pru(initial_gvn(), &for_igvn);
732 }
733 }
735 // Note: Large methods are capped off in do_one_bytecode().
736 if (failing()) return;
738 // After parsing, node notes are no longer automagic.
739 // They must be propagated by register_new_node_with_optimizer(),
740 // clone(), or the like.
741 set_default_node_notes(NULL);
743 for (;;) {
744 int successes = Inline_Warm();
745 if (failing()) return;
746 if (successes == 0) break;
747 }
749 // Drain the list.
750 Finish_Warm();
751 #ifndef PRODUCT
752 if (_printer) {
753 _printer->print_inlining(this);
754 }
755 #endif
757 if (failing()) return;
758 NOT_PRODUCT( verify_graph_edges(); )
760 // Now optimize
761 Optimize();
762 if (failing()) return;
763 NOT_PRODUCT( verify_graph_edges(); )
765 #ifndef PRODUCT
766 if (PrintIdeal) {
767 ttyLocker ttyl; // keep the following output all in one block
768 // This output goes directly to the tty, not the compiler log.
769 // To enable tools to match it up with the compilation activity,
770 // be sure to tag this tty output with the compile ID.
771 if (xtty != NULL) {
772 xtty->head("ideal compile_id='%d'%s", compile_id(),
773 is_osr_compilation() ? " compile_kind='osr'" :
774 "");
775 }
776 root()->dump(9999);
777 if (xtty != NULL) {
778 xtty->tail("ideal");
779 }
780 }
781 #endif
783 // Now that we know the size of all the monitors we can add a fixed slot
784 // for the original deopt pc.
786 _orig_pc_slot = fixed_slots();
787 int next_slot = _orig_pc_slot + (sizeof(address) / VMRegImpl::stack_slot_size);
788 set_fixed_slots(next_slot);
790 // Now generate code
791 Code_Gen();
792 if (failing()) return;
794 // Check if we want to skip execution of all compiled code.
795 {
796 #ifndef PRODUCT
797 if (OptoNoExecute) {
798 record_method_not_compilable("+OptoNoExecute"); // Flag as failed
799 return;
800 }
801 TracePhase t2("install_code", &_t_registerMethod, TimeCompiler);
802 #endif
804 if (is_osr_compilation()) {
805 _code_offsets.set_value(CodeOffsets::Verified_Entry, 0);
806 _code_offsets.set_value(CodeOffsets::OSR_Entry, _first_block_size);
807 } else {
808 _code_offsets.set_value(CodeOffsets::Verified_Entry, _first_block_size);
809 _code_offsets.set_value(CodeOffsets::OSR_Entry, 0);
810 }
812 env()->register_method(_method, _entry_bci,
813 &_code_offsets,
814 _orig_pc_slot_offset_in_bytes,
815 code_buffer(),
816 frame_size_in_words(), _oop_map_set,
817 &_handler_table, &_inc_table,
818 compiler,
819 env()->comp_level(),
820 true, /*has_debug_info*/
821 has_unsafe_access()
822 );
823 }
824 }
826 //------------------------------Compile----------------------------------------
827 // Compile a runtime stub
828 Compile::Compile( ciEnv* ci_env,
829 TypeFunc_generator generator,
830 address stub_function,
831 const char *stub_name,
832 int is_fancy_jump,
833 bool pass_tls,
834 bool save_arg_registers,
835 bool return_pc )
836 : Phase(Compiler),
837 _env(ci_env),
838 _log(ci_env->log()),
839 _compile_id(-1),
840 _save_argument_registers(save_arg_registers),
841 _method(NULL),
842 _stub_name(stub_name),
843 _stub_function(stub_function),
844 _stub_entry_point(NULL),
845 _entry_bci(InvocationEntryBci),
846 _initial_gvn(NULL),
847 _for_igvn(NULL),
848 _warm_calls(NULL),
849 _orig_pc_slot(0),
850 _orig_pc_slot_offset_in_bytes(0),
851 _subsume_loads(true),
852 _do_escape_analysis(false),
853 _failure_reason(NULL),
854 _code_buffer("Compile::Fill_buffer"),
855 _has_method_handle_invokes(false),
856 _mach_constant_base_node(NULL),
857 _node_bundling_limit(0),
858 _node_bundling_base(NULL),
859 _java_calls(0),
860 _inner_loops(0),
861 #ifndef PRODUCT
862 _trace_opto_output(TraceOptoOutput),
863 _printer(NULL),
864 #endif
865 _congraph(NULL) {
866 C = this;
868 #ifndef PRODUCT
869 TraceTime t1(NULL, &_t_totalCompilation, TimeCompiler, false);
870 TraceTime t2(NULL, &_t_stubCompilation, TimeCompiler, false);
871 set_print_assembly(PrintFrameConverterAssembly);
872 set_parsed_irreducible_loop(false);
873 #endif
874 CompileWrapper cw(this);
875 Init(/*AliasLevel=*/ 0);
876 init_tf((*generator)());
878 {
879 // The following is a dummy for the sake of GraphKit::gen_stub
880 Unique_Node_List for_igvn(comp_arena());
881 set_for_igvn(&for_igvn); // not used, but some GraphKit guys push on this
882 PhaseGVN gvn(Thread::current()->resource_area(),255);
883 set_initial_gvn(&gvn); // not significant, but GraphKit guys use it pervasively
884 gvn.transform_no_reclaim(top());
886 GraphKit kit;
887 kit.gen_stub(stub_function, stub_name, is_fancy_jump, pass_tls, return_pc);
888 }
890 NOT_PRODUCT( verify_graph_edges(); )
891 Code_Gen();
892 if (failing()) return;
895 // Entry point will be accessed using compile->stub_entry_point();
896 if (code_buffer() == NULL) {
897 Matcher::soft_match_failure();
898 } else {
899 if (PrintAssembly && (WizardMode || Verbose))
900 tty->print_cr("### Stub::%s", stub_name);
902 if (!failing()) {
903 assert(_fixed_slots == 0, "no fixed slots used for runtime stubs");
905 // Make the NMethod
906 // For now we mark the frame as never safe for profile stackwalking
907 RuntimeStub *rs = RuntimeStub::new_runtime_stub(stub_name,
908 code_buffer(),
909 CodeOffsets::frame_never_safe,
910 // _code_offsets.value(CodeOffsets::Frame_Complete),
911 frame_size_in_words(),
912 _oop_map_set,
913 save_arg_registers);
914 assert(rs != NULL && rs->is_runtime_stub(), "sanity check");
916 _stub_entry_point = rs->entry_point();
917 }
918 }
919 }
921 #ifndef PRODUCT
922 void print_opto_verbose_signature( const TypeFunc *j_sig, const char *stub_name ) {
923 if(PrintOpto && Verbose) {
924 tty->print("%s ", stub_name); j_sig->print_flattened(); tty->cr();
925 }
926 }
927 #endif
929 void Compile::print_codes() {
930 }
932 //------------------------------Init-------------------------------------------
933 // Prepare for a single compilation
934 void Compile::Init(int aliaslevel) {
935 _unique = 0;
936 _regalloc = NULL;
938 _tf = NULL; // filled in later
939 _top = NULL; // cached later
940 _matcher = NULL; // filled in later
941 _cfg = NULL; // filled in later
943 set_24_bit_selection_and_mode(Use24BitFP, false);
945 _node_note_array = NULL;
946 _default_node_notes = NULL;
948 _immutable_memory = NULL; // filled in at first inquiry
950 // Globally visible Nodes
951 // First set TOP to NULL to give safe behavior during creation of RootNode
952 set_cached_top_node(NULL);
953 set_root(new (this, 3) RootNode());
954 // Now that you have a Root to point to, create the real TOP
955 set_cached_top_node( new (this, 1) ConNode(Type::TOP) );
956 set_recent_alloc(NULL, NULL);
958 // Create Debug Information Recorder to record scopes, oopmaps, etc.
959 env()->set_oop_recorder(new OopRecorder(comp_arena()));
960 env()->set_debug_info(new DebugInformationRecorder(env()->oop_recorder()));
961 env()->set_dependencies(new Dependencies(env()));
963 _fixed_slots = 0;
964 set_has_split_ifs(false);
965 set_has_loops(has_method() && method()->has_loops()); // first approximation
966 set_has_stringbuilder(false);
967 _trap_can_recompile = false; // no traps emitted yet
968 _major_progress = true; // start out assuming good things will happen
969 set_has_unsafe_access(false);
970 Copy::zero_to_bytes(_trap_hist, sizeof(_trap_hist));
971 set_decompile_count(0);
973 set_do_freq_based_layout(BlockLayoutByFrequency || method_has_option("BlockLayoutByFrequency"));
974 set_num_loop_opts(LoopOptsCount);
975 set_do_inlining(Inline);
976 set_max_inline_size(MaxInlineSize);
977 set_freq_inline_size(FreqInlineSize);
978 set_do_scheduling(OptoScheduling);
979 set_do_count_invocations(false);
980 set_do_method_data_update(false);
982 if (debug_info()->recording_non_safepoints()) {
983 set_node_note_array(new(comp_arena()) GrowableArray<Node_Notes*>
984 (comp_arena(), 8, 0, NULL));
985 set_default_node_notes(Node_Notes::make(this));
986 }
988 // // -- Initialize types before each compile --
989 // // Update cached type information
990 // if( _method && _method->constants() )
991 // Type::update_loaded_types(_method, _method->constants());
993 // Init alias_type map.
994 if (!_do_escape_analysis && aliaslevel == 3)
995 aliaslevel = 2; // No unique types without escape analysis
996 _AliasLevel = aliaslevel;
997 const int grow_ats = 16;
998 _max_alias_types = grow_ats;
999 _alias_types = NEW_ARENA_ARRAY(comp_arena(), AliasType*, grow_ats);
1000 AliasType* ats = NEW_ARENA_ARRAY(comp_arena(), AliasType, grow_ats);
1001 Copy::zero_to_bytes(ats, sizeof(AliasType)*grow_ats);
1002 {
1003 for (int i = 0; i < grow_ats; i++) _alias_types[i] = &ats[i];
1004 }
1005 // Initialize the first few types.
1006 _alias_types[AliasIdxTop]->Init(AliasIdxTop, NULL);
1007 _alias_types[AliasIdxBot]->Init(AliasIdxBot, TypePtr::BOTTOM);
1008 _alias_types[AliasIdxRaw]->Init(AliasIdxRaw, TypeRawPtr::BOTTOM);
1009 _num_alias_types = AliasIdxRaw+1;
1010 // Zero out the alias type cache.
1011 Copy::zero_to_bytes(_alias_cache, sizeof(_alias_cache));
1012 // A NULL adr_type hits in the cache right away. Preload the right answer.
1013 probe_alias_cache(NULL)->_index = AliasIdxTop;
1015 _intrinsics = NULL;
1016 _macro_nodes = new(comp_arena()) GrowableArray<Node*>(comp_arena(), 8, 0, NULL);
1017 _predicate_opaqs = new(comp_arena()) GrowableArray<Node*>(comp_arena(), 8, 0, NULL);
1018 register_library_intrinsics();
1019 }
1021 //---------------------------init_start----------------------------------------
1022 // Install the StartNode on this compile object.
1023 void Compile::init_start(StartNode* s) {
1024 if (failing())
1025 return; // already failing
1026 assert(s == start(), "");
1027 }
1029 StartNode* Compile::start() const {
1030 assert(!failing(), "");
1031 for (DUIterator_Fast imax, i = root()->fast_outs(imax); i < imax; i++) {
1032 Node* start = root()->fast_out(i);
1033 if( start->is_Start() )
1034 return start->as_Start();
1035 }
1036 ShouldNotReachHere();
1037 return NULL;
1038 }
1040 //-------------------------------immutable_memory-------------------------------------
1041 // Access immutable memory
1042 Node* Compile::immutable_memory() {
1043 if (_immutable_memory != NULL) {
1044 return _immutable_memory;
1045 }
1046 StartNode* s = start();
1047 for (DUIterator_Fast imax, i = s->fast_outs(imax); true; i++) {
1048 Node *p = s->fast_out(i);
1049 if (p != s && p->as_Proj()->_con == TypeFunc::Memory) {
1050 _immutable_memory = p;
1051 return _immutable_memory;
1052 }
1053 }
1054 ShouldNotReachHere();
1055 return NULL;
1056 }
1058 //----------------------set_cached_top_node------------------------------------
1059 // Install the cached top node, and make sure Node::is_top works correctly.
1060 void Compile::set_cached_top_node(Node* tn) {
1061 if (tn != NULL) verify_top(tn);
1062 Node* old_top = _top;
1063 _top = tn;
1064 // Calling Node::setup_is_top allows the nodes the chance to adjust
1065 // their _out arrays.
1066 if (_top != NULL) _top->setup_is_top();
1067 if (old_top != NULL) old_top->setup_is_top();
1068 assert(_top == NULL || top()->is_top(), "");
1069 }
1071 #ifndef PRODUCT
1072 void Compile::verify_top(Node* tn) const {
1073 if (tn != NULL) {
1074 assert(tn->is_Con(), "top node must be a constant");
1075 assert(((ConNode*)tn)->type() == Type::TOP, "top node must have correct type");
1076 assert(tn->in(0) != NULL, "must have live top node");
1077 }
1078 }
1079 #endif
1082 ///-------------------Managing Per-Node Debug & Profile Info-------------------
1084 void Compile::grow_node_notes(GrowableArray<Node_Notes*>* arr, int grow_by) {
1085 guarantee(arr != NULL, "");
1086 int num_blocks = arr->length();
1087 if (grow_by < num_blocks) grow_by = num_blocks;
1088 int num_notes = grow_by * _node_notes_block_size;
1089 Node_Notes* notes = NEW_ARENA_ARRAY(node_arena(), Node_Notes, num_notes);
1090 Copy::zero_to_bytes(notes, num_notes * sizeof(Node_Notes));
1091 while (num_notes > 0) {
1092 arr->append(notes);
1093 notes += _node_notes_block_size;
1094 num_notes -= _node_notes_block_size;
1095 }
1096 assert(num_notes == 0, "exact multiple, please");
1097 }
1099 bool Compile::copy_node_notes_to(Node* dest, Node* source) {
1100 if (source == NULL || dest == NULL) return false;
1102 if (dest->is_Con())
1103 return false; // Do not push debug info onto constants.
1105 #ifdef ASSERT
1106 // Leave a bread crumb trail pointing to the original node:
1107 if (dest != NULL && dest != source && dest->debug_orig() == NULL) {
1108 dest->set_debug_orig(source);
1109 }
1110 #endif
1112 if (node_note_array() == NULL)
1113 return false; // Not collecting any notes now.
1115 // This is a copy onto a pre-existing node, which may already have notes.
1116 // If both nodes have notes, do not overwrite any pre-existing notes.
1117 Node_Notes* source_notes = node_notes_at(source->_idx);
1118 if (source_notes == NULL || source_notes->is_clear()) return false;
1119 Node_Notes* dest_notes = node_notes_at(dest->_idx);
1120 if (dest_notes == NULL || dest_notes->is_clear()) {
1121 return set_node_notes_at(dest->_idx, source_notes);
1122 }
1124 Node_Notes merged_notes = (*source_notes);
1125 // The order of operations here ensures that dest notes will win...
1126 merged_notes.update_from(dest_notes);
1127 return set_node_notes_at(dest->_idx, &merged_notes);
1128 }
1131 //--------------------------allow_range_check_smearing-------------------------
1132 // Gating condition for coalescing similar range checks.
1133 // Sometimes we try 'speculatively' replacing a series of a range checks by a
1134 // single covering check that is at least as strong as any of them.
1135 // If the optimization succeeds, the simplified (strengthened) range check
1136 // will always succeed. If it fails, we will deopt, and then give up
1137 // on the optimization.
1138 bool Compile::allow_range_check_smearing() const {
1139 // If this method has already thrown a range-check,
1140 // assume it was because we already tried range smearing
1141 // and it failed.
1142 uint already_trapped = trap_count(Deoptimization::Reason_range_check);
1143 return !already_trapped;
1144 }
1147 //------------------------------flatten_alias_type-----------------------------
1148 const TypePtr *Compile::flatten_alias_type( const TypePtr *tj ) const {
1149 int offset = tj->offset();
1150 TypePtr::PTR ptr = tj->ptr();
1152 // Known instance (scalarizable allocation) alias only with itself.
1153 bool is_known_inst = tj->isa_oopptr() != NULL &&
1154 tj->is_oopptr()->is_known_instance();
1156 // Process weird unsafe references.
1157 if (offset == Type::OffsetBot && (tj->isa_instptr() /*|| tj->isa_klassptr()*/)) {
1158 assert(InlineUnsafeOps, "indeterminate pointers come only from unsafe ops");
1159 assert(!is_known_inst, "scalarizable allocation should not have unsafe references");
1160 tj = TypeOopPtr::BOTTOM;
1161 ptr = tj->ptr();
1162 offset = tj->offset();
1163 }
1165 // Array pointers need some flattening
1166 const TypeAryPtr *ta = tj->isa_aryptr();
1167 if( ta && is_known_inst ) {
1168 if ( offset != Type::OffsetBot &&
1169 offset > arrayOopDesc::length_offset_in_bytes() ) {
1170 offset = Type::OffsetBot; // Flatten constant access into array body only
1171 tj = ta = TypeAryPtr::make(ptr, ta->ary(), ta->klass(), true, offset, ta->instance_id());
1172 }
1173 } else if( ta && _AliasLevel >= 2 ) {
1174 // For arrays indexed by constant indices, we flatten the alias
1175 // space to include all of the array body. Only the header, klass
1176 // and array length can be accessed un-aliased.
1177 if( offset != Type::OffsetBot ) {
1178 if( ta->const_oop() ) { // methodDataOop or methodOop
1179 offset = Type::OffsetBot; // Flatten constant access into array body
1180 tj = ta = TypeAryPtr::make(ptr,ta->const_oop(),ta->ary(),ta->klass(),false,offset);
1181 } else if( offset == arrayOopDesc::length_offset_in_bytes() ) {
1182 // range is OK as-is.
1183 tj = ta = TypeAryPtr::RANGE;
1184 } else if( offset == oopDesc::klass_offset_in_bytes() ) {
1185 tj = TypeInstPtr::KLASS; // all klass loads look alike
1186 ta = TypeAryPtr::RANGE; // generic ignored junk
1187 ptr = TypePtr::BotPTR;
1188 } else if( offset == oopDesc::mark_offset_in_bytes() ) {
1189 tj = TypeInstPtr::MARK;
1190 ta = TypeAryPtr::RANGE; // generic ignored junk
1191 ptr = TypePtr::BotPTR;
1192 } else { // Random constant offset into array body
1193 offset = Type::OffsetBot; // Flatten constant access into array body
1194 tj = ta = TypeAryPtr::make(ptr,ta->ary(),ta->klass(),false,offset);
1195 }
1196 }
1197 // Arrays of fixed size alias with arrays of unknown size.
1198 if (ta->size() != TypeInt::POS) {
1199 const TypeAry *tary = TypeAry::make(ta->elem(), TypeInt::POS);
1200 tj = ta = TypeAryPtr::make(ptr,ta->const_oop(),tary,ta->klass(),false,offset);
1201 }
1202 // Arrays of known objects become arrays of unknown objects.
1203 if (ta->elem()->isa_narrowoop() && ta->elem() != TypeNarrowOop::BOTTOM) {
1204 const TypeAry *tary = TypeAry::make(TypeNarrowOop::BOTTOM, ta->size());
1205 tj = ta = TypeAryPtr::make(ptr,ta->const_oop(),tary,NULL,false,offset);
1206 }
1207 if (ta->elem()->isa_oopptr() && ta->elem() != TypeInstPtr::BOTTOM) {
1208 const TypeAry *tary = TypeAry::make(TypeInstPtr::BOTTOM, ta->size());
1209 tj = ta = TypeAryPtr::make(ptr,ta->const_oop(),tary,NULL,false,offset);
1210 }
1211 // Arrays of bytes and of booleans both use 'bastore' and 'baload' so
1212 // cannot be distinguished by bytecode alone.
1213 if (ta->elem() == TypeInt::BOOL) {
1214 const TypeAry *tary = TypeAry::make(TypeInt::BYTE, ta->size());
1215 ciKlass* aklass = ciTypeArrayKlass::make(T_BYTE);
1216 tj = ta = TypeAryPtr::make(ptr,ta->const_oop(),tary,aklass,false,offset);
1217 }
1218 // During the 2nd round of IterGVN, NotNull castings are removed.
1219 // Make sure the Bottom and NotNull variants alias the same.
1220 // Also, make sure exact and non-exact variants alias the same.
1221 if( ptr == TypePtr::NotNull || ta->klass_is_exact() ) {
1222 tj = ta = TypeAryPtr::make(TypePtr::BotPTR,ta->ary(),ta->klass(),false,offset);
1223 }
1224 }
1226 // Oop pointers need some flattening
1227 const TypeInstPtr *to = tj->isa_instptr();
1228 if( to && _AliasLevel >= 2 && to != TypeOopPtr::BOTTOM ) {
1229 ciInstanceKlass *k = to->klass()->as_instance_klass();
1230 if( ptr == TypePtr::Constant ) {
1231 if (to->klass() != ciEnv::current()->Class_klass() ||
1232 offset < k->size_helper() * wordSize) {
1233 // No constant oop pointers (such as Strings); they alias with
1234 // unknown strings.
1235 assert(!is_known_inst, "not scalarizable allocation");
1236 tj = to = TypeInstPtr::make(TypePtr::BotPTR,to->klass(),false,0,offset);
1237 }
1238 } else if( is_known_inst ) {
1239 tj = to; // Keep NotNull and klass_is_exact for instance type
1240 } else if( ptr == TypePtr::NotNull || to->klass_is_exact() ) {
1241 // During the 2nd round of IterGVN, NotNull castings are removed.
1242 // Make sure the Bottom and NotNull variants alias the same.
1243 // Also, make sure exact and non-exact variants alias the same.
1244 tj = to = TypeInstPtr::make(TypePtr::BotPTR,to->klass(),false,0,offset);
1245 }
1246 // Canonicalize the holder of this field
1247 if (offset >= 0 && offset < instanceOopDesc::base_offset_in_bytes()) {
1248 // First handle header references such as a LoadKlassNode, even if the
1249 // object's klass is unloaded at compile time (4965979).
1250 if (!is_known_inst) { // Do it only for non-instance types
1251 tj = to = TypeInstPtr::make(TypePtr::BotPTR, env()->Object_klass(), false, NULL, offset);
1252 }
1253 } else if (offset < 0 || offset >= k->size_helper() * wordSize) {
1254 // Static fields are in the space above the normal instance
1255 // fields in the java.lang.Class instance.
1256 if (to->klass() != ciEnv::current()->Class_klass()) {
1257 to = NULL;
1258 tj = TypeOopPtr::BOTTOM;
1259 offset = tj->offset();
1260 }
1261 } else {
1262 ciInstanceKlass *canonical_holder = k->get_canonical_holder(offset);
1263 if (!k->equals(canonical_holder) || tj->offset() != offset) {
1264 if( is_known_inst ) {
1265 tj = to = TypeInstPtr::make(to->ptr(), canonical_holder, true, NULL, offset, to->instance_id());
1266 } else {
1267 tj = to = TypeInstPtr::make(to->ptr(), canonical_holder, false, NULL, offset);
1268 }
1269 }
1270 }
1271 }
1273 // Klass pointers to object array klasses need some flattening
1274 const TypeKlassPtr *tk = tj->isa_klassptr();
1275 if( tk ) {
1276 // If we are referencing a field within a Klass, we need
1277 // to assume the worst case of an Object. Both exact and
1278 // inexact types must flatten to the same alias class.
1279 // Since the flattened result for a klass is defined to be
1280 // precisely java.lang.Object, use a constant ptr.
1281 if ( offset == Type::OffsetBot || (offset >= 0 && (size_t)offset < sizeof(Klass)) ) {
1283 tj = tk = TypeKlassPtr::make(TypePtr::Constant,
1284 TypeKlassPtr::OBJECT->klass(),
1285 offset);
1286 }
1288 ciKlass* klass = tk->klass();
1289 if( klass->is_obj_array_klass() ) {
1290 ciKlass* k = TypeAryPtr::OOPS->klass();
1291 if( !k || !k->is_loaded() ) // Only fails for some -Xcomp runs
1292 k = TypeInstPtr::BOTTOM->klass();
1293 tj = tk = TypeKlassPtr::make( TypePtr::NotNull, k, offset );
1294 }
1296 // Check for precise loads from the primary supertype array and force them
1297 // to the supertype cache alias index. Check for generic array loads from
1298 // the primary supertype array and also force them to the supertype cache
1299 // alias index. Since the same load can reach both, we need to merge
1300 // these 2 disparate memories into the same alias class. Since the
1301 // primary supertype array is read-only, there's no chance of confusion
1302 // where we bypass an array load and an array store.
1303 uint off2 = offset - Klass::primary_supers_offset_in_bytes();
1304 if( offset == Type::OffsetBot ||
1305 off2 < Klass::primary_super_limit()*wordSize ) {
1306 offset = sizeof(oopDesc) +Klass::secondary_super_cache_offset_in_bytes();
1307 tj = tk = TypeKlassPtr::make( TypePtr::NotNull, tk->klass(), offset );
1308 }
1309 }
1311 // Flatten all Raw pointers together.
1312 if (tj->base() == Type::RawPtr)
1313 tj = TypeRawPtr::BOTTOM;
1315 if (tj->base() == Type::AnyPtr)
1316 tj = TypePtr::BOTTOM; // An error, which the caller must check for.
1318 // Flatten all to bottom for now
1319 switch( _AliasLevel ) {
1320 case 0:
1321 tj = TypePtr::BOTTOM;
1322 break;
1323 case 1: // Flatten to: oop, static, field or array
1324 switch (tj->base()) {
1325 //case Type::AryPtr: tj = TypeAryPtr::RANGE; break;
1326 case Type::RawPtr: tj = TypeRawPtr::BOTTOM; break;
1327 case Type::AryPtr: // do not distinguish arrays at all
1328 case Type::InstPtr: tj = TypeInstPtr::BOTTOM; break;
1329 case Type::KlassPtr: tj = TypeKlassPtr::OBJECT; break;
1330 case Type::AnyPtr: tj = TypePtr::BOTTOM; break; // caller checks it
1331 default: ShouldNotReachHere();
1332 }
1333 break;
1334 case 2: // No collapsing at level 2; keep all splits
1335 case 3: // No collapsing at level 3; keep all splits
1336 break;
1337 default:
1338 Unimplemented();
1339 }
1341 offset = tj->offset();
1342 assert( offset != Type::OffsetTop, "Offset has fallen from constant" );
1344 assert( (offset != Type::OffsetBot && tj->base() != Type::AryPtr) ||
1345 (offset == Type::OffsetBot && tj->base() == Type::AryPtr) ||
1346 (offset == Type::OffsetBot && tj == TypeOopPtr::BOTTOM) ||
1347 (offset == Type::OffsetBot && tj == TypePtr::BOTTOM) ||
1348 (offset == oopDesc::mark_offset_in_bytes() && tj->base() == Type::AryPtr) ||
1349 (offset == oopDesc::klass_offset_in_bytes() && tj->base() == Type::AryPtr) ||
1350 (offset == arrayOopDesc::length_offset_in_bytes() && tj->base() == Type::AryPtr) ,
1351 "For oops, klasses, raw offset must be constant; for arrays the offset is never known" );
1352 assert( tj->ptr() != TypePtr::TopPTR &&
1353 tj->ptr() != TypePtr::AnyNull &&
1354 tj->ptr() != TypePtr::Null, "No imprecise addresses" );
1355 // assert( tj->ptr() != TypePtr::Constant ||
1356 // tj->base() == Type::RawPtr ||
1357 // tj->base() == Type::KlassPtr, "No constant oop addresses" );
1359 return tj;
1360 }
1362 void Compile::AliasType::Init(int i, const TypePtr* at) {
1363 _index = i;
1364 _adr_type = at;
1365 _field = NULL;
1366 _is_rewritable = true; // default
1367 const TypeOopPtr *atoop = (at != NULL) ? at->isa_oopptr() : NULL;
1368 if (atoop != NULL && atoop->is_known_instance()) {
1369 const TypeOopPtr *gt = atoop->cast_to_instance_id(TypeOopPtr::InstanceBot);
1370 _general_index = Compile::current()->get_alias_index(gt);
1371 } else {
1372 _general_index = 0;
1373 }
1374 }
1376 //---------------------------------print_on------------------------------------
1377 #ifndef PRODUCT
1378 void Compile::AliasType::print_on(outputStream* st) {
1379 if (index() < 10)
1380 st->print("@ <%d> ", index());
1381 else st->print("@ <%d>", index());
1382 st->print(is_rewritable() ? " " : " RO");
1383 int offset = adr_type()->offset();
1384 if (offset == Type::OffsetBot)
1385 st->print(" +any");
1386 else st->print(" +%-3d", offset);
1387 st->print(" in ");
1388 adr_type()->dump_on(st);
1389 const TypeOopPtr* tjp = adr_type()->isa_oopptr();
1390 if (field() != NULL && tjp) {
1391 if (tjp->klass() != field()->holder() ||
1392 tjp->offset() != field()->offset_in_bytes()) {
1393 st->print(" != ");
1394 field()->print();
1395 st->print(" ***");
1396 }
1397 }
1398 }
1400 void print_alias_types() {
1401 Compile* C = Compile::current();
1402 tty->print_cr("--- Alias types, AliasIdxBot .. %d", C->num_alias_types()-1);
1403 for (int idx = Compile::AliasIdxBot; idx < C->num_alias_types(); idx++) {
1404 C->alias_type(idx)->print_on(tty);
1405 tty->cr();
1406 }
1407 }
1408 #endif
1411 //----------------------------probe_alias_cache--------------------------------
1412 Compile::AliasCacheEntry* Compile::probe_alias_cache(const TypePtr* adr_type) {
1413 intptr_t key = (intptr_t) adr_type;
1414 key ^= key >> logAliasCacheSize;
1415 return &_alias_cache[key & right_n_bits(logAliasCacheSize)];
1416 }
1419 //-----------------------------grow_alias_types--------------------------------
1420 void Compile::grow_alias_types() {
1421 const int old_ats = _max_alias_types; // how many before?
1422 const int new_ats = old_ats; // how many more?
1423 const int grow_ats = old_ats+new_ats; // how many now?
1424 _max_alias_types = grow_ats;
1425 _alias_types = REALLOC_ARENA_ARRAY(comp_arena(), AliasType*, _alias_types, old_ats, grow_ats);
1426 AliasType* ats = NEW_ARENA_ARRAY(comp_arena(), AliasType, new_ats);
1427 Copy::zero_to_bytes(ats, sizeof(AliasType)*new_ats);
1428 for (int i = 0; i < new_ats; i++) _alias_types[old_ats+i] = &ats[i];
1429 }
1432 //--------------------------------find_alias_type------------------------------
1433 Compile::AliasType* Compile::find_alias_type(const TypePtr* adr_type, bool no_create, ciField* original_field) {
1434 if (_AliasLevel == 0)
1435 return alias_type(AliasIdxBot);
1437 AliasCacheEntry* ace = probe_alias_cache(adr_type);
1438 if (ace->_adr_type == adr_type) {
1439 return alias_type(ace->_index);
1440 }
1442 // Handle special cases.
1443 if (adr_type == NULL) return alias_type(AliasIdxTop);
1444 if (adr_type == TypePtr::BOTTOM) return alias_type(AliasIdxBot);
1446 // Do it the slow way.
1447 const TypePtr* flat = flatten_alias_type(adr_type);
1449 #ifdef ASSERT
1450 assert(flat == flatten_alias_type(flat), "idempotent");
1451 assert(flat != TypePtr::BOTTOM, "cannot alias-analyze an untyped ptr");
1452 if (flat->isa_oopptr() && !flat->isa_klassptr()) {
1453 const TypeOopPtr* foop = flat->is_oopptr();
1454 // Scalarizable allocations have exact klass always.
1455 bool exact = !foop->klass_is_exact() || foop->is_known_instance();
1456 const TypePtr* xoop = foop->cast_to_exactness(exact)->is_ptr();
1457 assert(foop == flatten_alias_type(xoop), "exactness must not affect alias type");
1458 }
1459 assert(flat == flatten_alias_type(flat), "exact bit doesn't matter");
1460 #endif
1462 int idx = AliasIdxTop;
1463 for (int i = 0; i < num_alias_types(); i++) {
1464 if (alias_type(i)->adr_type() == flat) {
1465 idx = i;
1466 break;
1467 }
1468 }
1470 if (idx == AliasIdxTop) {
1471 if (no_create) return NULL;
1472 // Grow the array if necessary.
1473 if (_num_alias_types == _max_alias_types) grow_alias_types();
1474 // Add a new alias type.
1475 idx = _num_alias_types++;
1476 _alias_types[idx]->Init(idx, flat);
1477 if (flat == TypeInstPtr::KLASS) alias_type(idx)->set_rewritable(false);
1478 if (flat == TypeAryPtr::RANGE) alias_type(idx)->set_rewritable(false);
1479 if (flat->isa_instptr()) {
1480 if (flat->offset() == java_lang_Class::klass_offset_in_bytes()
1481 && flat->is_instptr()->klass() == env()->Class_klass())
1482 alias_type(idx)->set_rewritable(false);
1483 }
1484 if (flat->isa_klassptr()) {
1485 if (flat->offset() == Klass::super_check_offset_offset_in_bytes() + (int)sizeof(oopDesc))
1486 alias_type(idx)->set_rewritable(false);
1487 if (flat->offset() == Klass::modifier_flags_offset_in_bytes() + (int)sizeof(oopDesc))
1488 alias_type(idx)->set_rewritable(false);
1489 if (flat->offset() == Klass::access_flags_offset_in_bytes() + (int)sizeof(oopDesc))
1490 alias_type(idx)->set_rewritable(false);
1491 if (flat->offset() == Klass::java_mirror_offset_in_bytes() + (int)sizeof(oopDesc))
1492 alias_type(idx)->set_rewritable(false);
1493 }
1494 // %%% (We would like to finalize JavaThread::threadObj_offset(),
1495 // but the base pointer type is not distinctive enough to identify
1496 // references into JavaThread.)
1498 // Check for final fields.
1499 const TypeInstPtr* tinst = flat->isa_instptr();
1500 if (tinst && tinst->offset() >= instanceOopDesc::base_offset_in_bytes()) {
1501 ciField* field;
1502 if (tinst->const_oop() != NULL &&
1503 tinst->klass() == ciEnv::current()->Class_klass() &&
1504 tinst->offset() >= (tinst->klass()->as_instance_klass()->size_helper() * wordSize)) {
1505 // static field
1506 ciInstanceKlass* k = tinst->const_oop()->as_instance()->java_lang_Class_klass()->as_instance_klass();
1507 field = k->get_field_by_offset(tinst->offset(), true);
1508 } else {
1509 ciInstanceKlass *k = tinst->klass()->as_instance_klass();
1510 field = k->get_field_by_offset(tinst->offset(), false);
1511 }
1512 assert(field == NULL ||
1513 original_field == NULL ||
1514 (field->holder() == original_field->holder() &&
1515 field->offset() == original_field->offset() &&
1516 field->is_static() == original_field->is_static()), "wrong field?");
1517 // Set field() and is_rewritable() attributes.
1518 if (field != NULL) alias_type(idx)->set_field(field);
1519 }
1520 }
1522 // Fill the cache for next time.
1523 ace->_adr_type = adr_type;
1524 ace->_index = idx;
1525 assert(alias_type(adr_type) == alias_type(idx), "type must be installed");
1527 // Might as well try to fill the cache for the flattened version, too.
1528 AliasCacheEntry* face = probe_alias_cache(flat);
1529 if (face->_adr_type == NULL) {
1530 face->_adr_type = flat;
1531 face->_index = idx;
1532 assert(alias_type(flat) == alias_type(idx), "flat type must work too");
1533 }
1535 return alias_type(idx);
1536 }
1539 Compile::AliasType* Compile::alias_type(ciField* field) {
1540 const TypeOopPtr* t;
1541 if (field->is_static())
1542 t = TypeInstPtr::make(field->holder()->java_mirror());
1543 else
1544 t = TypeOopPtr::make_from_klass_raw(field->holder());
1545 AliasType* atp = alias_type(t->add_offset(field->offset_in_bytes()), field);
1546 assert(field->is_final() == !atp->is_rewritable(), "must get the rewritable bits correct");
1547 return atp;
1548 }
1551 //------------------------------have_alias_type--------------------------------
1552 bool Compile::have_alias_type(const TypePtr* adr_type) {
1553 AliasCacheEntry* ace = probe_alias_cache(adr_type);
1554 if (ace->_adr_type == adr_type) {
1555 return true;
1556 }
1558 // Handle special cases.
1559 if (adr_type == NULL) return true;
1560 if (adr_type == TypePtr::BOTTOM) return true;
1562 return find_alias_type(adr_type, true, NULL) != NULL;
1563 }
1565 //-----------------------------must_alias--------------------------------------
1566 // True if all values of the given address type are in the given alias category.
1567 bool Compile::must_alias(const TypePtr* adr_type, int alias_idx) {
1568 if (alias_idx == AliasIdxBot) return true; // the universal category
1569 if (adr_type == NULL) return true; // NULL serves as TypePtr::TOP
1570 if (alias_idx == AliasIdxTop) return false; // the empty category
1571 if (adr_type->base() == Type::AnyPtr) return false; // TypePtr::BOTTOM or its twins
1573 // the only remaining possible overlap is identity
1574 int adr_idx = get_alias_index(adr_type);
1575 assert(adr_idx != AliasIdxBot && adr_idx != AliasIdxTop, "");
1576 assert(adr_idx == alias_idx ||
1577 (alias_type(alias_idx)->adr_type() != TypeOopPtr::BOTTOM
1578 && adr_type != TypeOopPtr::BOTTOM),
1579 "should not be testing for overlap with an unsafe pointer");
1580 return adr_idx == alias_idx;
1581 }
1583 //------------------------------can_alias--------------------------------------
1584 // True if any values of the given address type are in the given alias category.
1585 bool Compile::can_alias(const TypePtr* adr_type, int alias_idx) {
1586 if (alias_idx == AliasIdxTop) return false; // the empty category
1587 if (adr_type == NULL) return false; // NULL serves as TypePtr::TOP
1588 if (alias_idx == AliasIdxBot) return true; // the universal category
1589 if (adr_type->base() == Type::AnyPtr) return true; // TypePtr::BOTTOM or its twins
1591 // the only remaining possible overlap is identity
1592 int adr_idx = get_alias_index(adr_type);
1593 assert(adr_idx != AliasIdxBot && adr_idx != AliasIdxTop, "");
1594 return adr_idx == alias_idx;
1595 }
1599 //---------------------------pop_warm_call-------------------------------------
1600 WarmCallInfo* Compile::pop_warm_call() {
1601 WarmCallInfo* wci = _warm_calls;
1602 if (wci != NULL) _warm_calls = wci->remove_from(wci);
1603 return wci;
1604 }
1606 //----------------------------Inline_Warm--------------------------------------
1607 int Compile::Inline_Warm() {
1608 // If there is room, try to inline some more warm call sites.
1609 // %%% Do a graph index compaction pass when we think we're out of space?
1610 if (!InlineWarmCalls) return 0;
1612 int calls_made_hot = 0;
1613 int room_to_grow = NodeCountInliningCutoff - unique();
1614 int amount_to_grow = MIN2(room_to_grow, (int)NodeCountInliningStep);
1615 int amount_grown = 0;
1616 WarmCallInfo* call;
1617 while (amount_to_grow > 0 && (call = pop_warm_call()) != NULL) {
1618 int est_size = (int)call->size();
1619 if (est_size > (room_to_grow - amount_grown)) {
1620 // This one won't fit anyway. Get rid of it.
1621 call->make_cold();
1622 continue;
1623 }
1624 call->make_hot();
1625 calls_made_hot++;
1626 amount_grown += est_size;
1627 amount_to_grow -= est_size;
1628 }
1630 if (calls_made_hot > 0) set_major_progress();
1631 return calls_made_hot;
1632 }
1635 //----------------------------Finish_Warm--------------------------------------
1636 void Compile::Finish_Warm() {
1637 if (!InlineWarmCalls) return;
1638 if (failing()) return;
1639 if (warm_calls() == NULL) return;
1641 // Clean up loose ends, if we are out of space for inlining.
1642 WarmCallInfo* call;
1643 while ((call = pop_warm_call()) != NULL) {
1644 call->make_cold();
1645 }
1646 }
1648 //---------------------cleanup_loop_predicates-----------------------
1649 // Remove the opaque nodes that protect the predicates so that all unused
1650 // checks and uncommon_traps will be eliminated from the ideal graph
1651 void Compile::cleanup_loop_predicates(PhaseIterGVN &igvn) {
1652 if (predicate_count()==0) return;
1653 for (int i = predicate_count(); i > 0; i--) {
1654 Node * n = predicate_opaque1_node(i-1);
1655 assert(n->Opcode() == Op_Opaque1, "must be");
1656 igvn.replace_node(n, n->in(1));
1657 }
1658 assert(predicate_count()==0, "should be clean!");
1659 }
1661 //------------------------------Optimize---------------------------------------
1662 // Given a graph, optimize it.
1663 void Compile::Optimize() {
1664 TracePhase t1("optimizer", &_t_optimizer, true);
1666 #ifndef PRODUCT
1667 if (env()->break_at_compile()) {
1668 BREAKPOINT;
1669 }
1671 #endif
1673 ResourceMark rm;
1674 int loop_opts_cnt;
1676 NOT_PRODUCT( verify_graph_edges(); )
1678 print_method("After Parsing");
1680 {
1681 // Iterative Global Value Numbering, including ideal transforms
1682 // Initialize IterGVN with types and values from parse-time GVN
1683 PhaseIterGVN igvn(initial_gvn());
1684 {
1685 NOT_PRODUCT( TracePhase t2("iterGVN", &_t_iterGVN, TimeCompiler); )
1686 igvn.optimize();
1687 }
1689 print_method("Iter GVN 1", 2);
1691 if (failing()) return;
1693 // Perform escape analysis
1694 if (_do_escape_analysis && ConnectionGraph::has_candidates(this)) {
1695 TracePhase t2("escapeAnalysis", &_t_escapeAnalysis, true);
1696 ConnectionGraph::do_analysis(this, &igvn);
1698 if (failing()) return;
1700 igvn.optimize();
1701 print_method("Iter GVN 3", 2);
1703 if (failing()) return;
1705 }
1707 // Loop transforms on the ideal graph. Range Check Elimination,
1708 // peeling, unrolling, etc.
1710 // Set loop opts counter
1711 loop_opts_cnt = num_loop_opts();
1712 if((loop_opts_cnt > 0) && (has_loops() || has_split_ifs())) {
1713 {
1714 TracePhase t2("idealLoop", &_t_idealLoop, true);
1715 PhaseIdealLoop ideal_loop( igvn, true );
1716 loop_opts_cnt--;
1717 if (major_progress()) print_method("PhaseIdealLoop 1", 2);
1718 if (failing()) return;
1719 }
1720 // Loop opts pass if partial peeling occurred in previous pass
1721 if(PartialPeelLoop && major_progress() && (loop_opts_cnt > 0)) {
1722 TracePhase t3("idealLoop", &_t_idealLoop, true);
1723 PhaseIdealLoop ideal_loop( igvn, false );
1724 loop_opts_cnt--;
1725 if (major_progress()) print_method("PhaseIdealLoop 2", 2);
1726 if (failing()) return;
1727 }
1728 // Loop opts pass for loop-unrolling before CCP
1729 if(major_progress() && (loop_opts_cnt > 0)) {
1730 TracePhase t4("idealLoop", &_t_idealLoop, true);
1731 PhaseIdealLoop ideal_loop( igvn, false );
1732 loop_opts_cnt--;
1733 if (major_progress()) print_method("PhaseIdealLoop 3", 2);
1734 }
1735 if (!failing()) {
1736 // Verify that last round of loop opts produced a valid graph
1737 NOT_PRODUCT( TracePhase t2("idealLoopVerify", &_t_idealLoopVerify, TimeCompiler); )
1738 PhaseIdealLoop::verify(igvn);
1739 }
1740 }
1741 if (failing()) return;
1743 // Conditional Constant Propagation;
1744 PhaseCCP ccp( &igvn );
1745 assert( true, "Break here to ccp.dump_nodes_and_types(_root,999,1)");
1746 {
1747 TracePhase t2("ccp", &_t_ccp, true);
1748 ccp.do_transform();
1749 }
1750 print_method("PhaseCPP 1", 2);
1752 assert( true, "Break here to ccp.dump_old2new_map()");
1754 // Iterative Global Value Numbering, including ideal transforms
1755 {
1756 NOT_PRODUCT( TracePhase t2("iterGVN2", &_t_iterGVN2, TimeCompiler); )
1757 igvn = ccp;
1758 igvn.optimize();
1759 }
1761 print_method("Iter GVN 2", 2);
1763 if (failing()) return;
1765 // Loop transforms on the ideal graph. Range Check Elimination,
1766 // peeling, unrolling, etc.
1767 if(loop_opts_cnt > 0) {
1768 debug_only( int cnt = 0; );
1769 while(major_progress() && (loop_opts_cnt > 0)) {
1770 TracePhase t2("idealLoop", &_t_idealLoop, true);
1771 assert( cnt++ < 40, "infinite cycle in loop optimization" );
1772 PhaseIdealLoop ideal_loop( igvn, true);
1773 loop_opts_cnt--;
1774 if (major_progress()) print_method("PhaseIdealLoop iterations", 2);
1775 if (failing()) return;
1776 }
1777 }
1779 {
1780 // Verify that all previous optimizations produced a valid graph
1781 // at least to this point, even if no loop optimizations were done.
1782 NOT_PRODUCT( TracePhase t2("idealLoopVerify", &_t_idealLoopVerify, TimeCompiler); )
1783 PhaseIdealLoop::verify(igvn);
1784 }
1786 {
1787 NOT_PRODUCT( TracePhase t2("macroExpand", &_t_macroExpand, TimeCompiler); )
1788 PhaseMacroExpand mex(igvn);
1789 if (mex.expand_macro_nodes()) {
1790 assert(failing(), "must bail out w/ explicit message");
1791 return;
1792 }
1793 }
1795 } // (End scope of igvn; run destructor if necessary for asserts.)
1797 // A method with only infinite loops has no edges entering loops from root
1798 {
1799 NOT_PRODUCT( TracePhase t2("graphReshape", &_t_graphReshaping, TimeCompiler); )
1800 if (final_graph_reshaping()) {
1801 assert(failing(), "must bail out w/ explicit message");
1802 return;
1803 }
1804 }
1806 print_method("Optimize finished", 2);
1807 }
1810 //------------------------------Code_Gen---------------------------------------
1811 // Given a graph, generate code for it
1812 void Compile::Code_Gen() {
1813 if (failing()) return;
1815 // Perform instruction selection. You might think we could reclaim Matcher
1816 // memory PDQ, but actually the Matcher is used in generating spill code.
1817 // Internals of the Matcher (including some VectorSets) must remain live
1818 // for awhile - thus I cannot reclaim Matcher memory lest a VectorSet usage
1819 // set a bit in reclaimed memory.
1821 // In debug mode can dump m._nodes.dump() for mapping of ideal to machine
1822 // nodes. Mapping is only valid at the root of each matched subtree.
1823 NOT_PRODUCT( verify_graph_edges(); )
1825 Node_List proj_list;
1826 Matcher m(proj_list);
1827 _matcher = &m;
1828 {
1829 TracePhase t2("matcher", &_t_matcher, true);
1830 m.match();
1831 }
1832 // In debug mode can dump m._nodes.dump() for mapping of ideal to machine
1833 // nodes. Mapping is only valid at the root of each matched subtree.
1834 NOT_PRODUCT( verify_graph_edges(); )
1836 // If you have too many nodes, or if matching has failed, bail out
1837 check_node_count(0, "out of nodes matching instructions");
1838 if (failing()) return;
1840 // Build a proper-looking CFG
1841 PhaseCFG cfg(node_arena(), root(), m);
1842 _cfg = &cfg;
1843 {
1844 NOT_PRODUCT( TracePhase t2("scheduler", &_t_scheduler, TimeCompiler); )
1845 cfg.Dominators();
1846 if (failing()) return;
1848 NOT_PRODUCT( verify_graph_edges(); )
1850 cfg.Estimate_Block_Frequency();
1851 cfg.GlobalCodeMotion(m,unique(),proj_list);
1853 print_method("Global code motion", 2);
1855 if (failing()) return;
1856 NOT_PRODUCT( verify_graph_edges(); )
1858 debug_only( cfg.verify(); )
1859 }
1860 NOT_PRODUCT( verify_graph_edges(); )
1862 PhaseChaitin regalloc(unique(),cfg,m);
1863 _regalloc = ®alloc;
1864 {
1865 TracePhase t2("regalloc", &_t_registerAllocation, true);
1866 // Perform any platform dependent preallocation actions. This is used,
1867 // for example, to avoid taking an implicit null pointer exception
1868 // using the frame pointer on win95.
1869 _regalloc->pd_preallocate_hook();
1871 // Perform register allocation. After Chaitin, use-def chains are
1872 // no longer accurate (at spill code) and so must be ignored.
1873 // Node->LRG->reg mappings are still accurate.
1874 _regalloc->Register_Allocate();
1876 // Bail out if the allocator builds too many nodes
1877 if (failing()) return;
1878 }
1880 // Prior to register allocation we kept empty basic blocks in case the
1881 // the allocator needed a place to spill. After register allocation we
1882 // are not adding any new instructions. If any basic block is empty, we
1883 // can now safely remove it.
1884 {
1885 NOT_PRODUCT( TracePhase t2("blockOrdering", &_t_blockOrdering, TimeCompiler); )
1886 cfg.remove_empty();
1887 if (do_freq_based_layout()) {
1888 PhaseBlockLayout layout(cfg);
1889 } else {
1890 cfg.set_loop_alignment();
1891 }
1892 cfg.fixup_flow();
1893 }
1895 // Perform any platform dependent postallocation verifications.
1896 debug_only( _regalloc->pd_postallocate_verify_hook(); )
1898 // Apply peephole optimizations
1899 if( OptoPeephole ) {
1900 NOT_PRODUCT( TracePhase t2("peephole", &_t_peephole, TimeCompiler); )
1901 PhasePeephole peep( _regalloc, cfg);
1902 peep.do_transform();
1903 }
1905 // Convert Nodes to instruction bits in a buffer
1906 {
1907 // %%%% workspace merge brought two timers together for one job
1908 TracePhase t2a("output", &_t_output, true);
1909 NOT_PRODUCT( TraceTime t2b(NULL, &_t_codeGeneration, TimeCompiler, false); )
1910 Output();
1911 }
1913 print_method("Final Code");
1915 // He's dead, Jim.
1916 _cfg = (PhaseCFG*)0xdeadbeef;
1917 _regalloc = (PhaseChaitin*)0xdeadbeef;
1918 }
1921 //------------------------------dump_asm---------------------------------------
1922 // Dump formatted assembly
1923 #ifndef PRODUCT
1924 void Compile::dump_asm(int *pcs, uint pc_limit) {
1925 bool cut_short = false;
1926 tty->print_cr("#");
1927 tty->print("# "); _tf->dump(); tty->cr();
1928 tty->print_cr("#");
1930 // For all blocks
1931 int pc = 0x0; // Program counter
1932 char starts_bundle = ' ';
1933 _regalloc->dump_frame();
1935 Node *n = NULL;
1936 for( uint i=0; i<_cfg->_num_blocks; i++ ) {
1937 if (VMThread::should_terminate()) { cut_short = true; break; }
1938 Block *b = _cfg->_blocks[i];
1939 if (b->is_connector() && !Verbose) continue;
1940 n = b->_nodes[0];
1941 if (pcs && n->_idx < pc_limit)
1942 tty->print("%3.3x ", pcs[n->_idx]);
1943 else
1944 tty->print(" ");
1945 b->dump_head( &_cfg->_bbs );
1946 if (b->is_connector()) {
1947 tty->print_cr(" # Empty connector block");
1948 } else if (b->num_preds() == 2 && b->pred(1)->is_CatchProj() && b->pred(1)->as_CatchProj()->_con == CatchProjNode::fall_through_index) {
1949 tty->print_cr(" # Block is sole successor of call");
1950 }
1952 // For all instructions
1953 Node *delay = NULL;
1954 for( uint j = 0; j<b->_nodes.size(); j++ ) {
1955 if (VMThread::should_terminate()) { cut_short = true; break; }
1956 n = b->_nodes[j];
1957 if (valid_bundle_info(n)) {
1958 Bundle *bundle = node_bundling(n);
1959 if (bundle->used_in_unconditional_delay()) {
1960 delay = n;
1961 continue;
1962 }
1963 if (bundle->starts_bundle())
1964 starts_bundle = '+';
1965 }
1967 if (WizardMode) n->dump();
1969 if( !n->is_Region() && // Dont print in the Assembly
1970 !n->is_Phi() && // a few noisely useless nodes
1971 !n->is_Proj() &&
1972 !n->is_MachTemp() &&
1973 !n->is_SafePointScalarObject() &&
1974 !n->is_Catch() && // Would be nice to print exception table targets
1975 !n->is_MergeMem() && // Not very interesting
1976 !n->is_top() && // Debug info table constants
1977 !(n->is_Con() && !n->is_Mach())// Debug info table constants
1978 ) {
1979 if (pcs && n->_idx < pc_limit)
1980 tty->print("%3.3x", pcs[n->_idx]);
1981 else
1982 tty->print(" ");
1983 tty->print(" %c ", starts_bundle);
1984 starts_bundle = ' ';
1985 tty->print("\t");
1986 n->format(_regalloc, tty);
1987 tty->cr();
1988 }
1990 // If we have an instruction with a delay slot, and have seen a delay,
1991 // then back up and print it
1992 if (valid_bundle_info(n) && node_bundling(n)->use_unconditional_delay()) {
1993 assert(delay != NULL, "no unconditional delay instruction");
1994 if (WizardMode) delay->dump();
1996 if (node_bundling(delay)->starts_bundle())
1997 starts_bundle = '+';
1998 if (pcs && n->_idx < pc_limit)
1999 tty->print("%3.3x", pcs[n->_idx]);
2000 else
2001 tty->print(" ");
2002 tty->print(" %c ", starts_bundle);
2003 starts_bundle = ' ';
2004 tty->print("\t");
2005 delay->format(_regalloc, tty);
2006 tty->print_cr("");
2007 delay = NULL;
2008 }
2010 // Dump the exception table as well
2011 if( n->is_Catch() && (Verbose || WizardMode) ) {
2012 // Print the exception table for this offset
2013 _handler_table.print_subtable_for(pc);
2014 }
2015 }
2017 if (pcs && n->_idx < pc_limit)
2018 tty->print_cr("%3.3x", pcs[n->_idx]);
2019 else
2020 tty->print_cr("");
2022 assert(cut_short || delay == NULL, "no unconditional delay branch");
2024 } // End of per-block dump
2025 tty->print_cr("");
2027 if (cut_short) tty->print_cr("*** disassembly is cut short ***");
2028 }
2029 #endif
2031 //------------------------------Final_Reshape_Counts---------------------------
2032 // This class defines counters to help identify when a method
2033 // may/must be executed using hardware with only 24-bit precision.
2034 struct Final_Reshape_Counts : public StackObj {
2035 int _call_count; // count non-inlined 'common' calls
2036 int _float_count; // count float ops requiring 24-bit precision
2037 int _double_count; // count double ops requiring more precision
2038 int _java_call_count; // count non-inlined 'java' calls
2039 int _inner_loop_count; // count loops which need alignment
2040 VectorSet _visited; // Visitation flags
2041 Node_List _tests; // Set of IfNodes & PCTableNodes
2043 Final_Reshape_Counts() :
2044 _call_count(0), _float_count(0), _double_count(0),
2045 _java_call_count(0), _inner_loop_count(0),
2046 _visited( Thread::current()->resource_area() ) { }
2048 void inc_call_count () { _call_count ++; }
2049 void inc_float_count () { _float_count ++; }
2050 void inc_double_count() { _double_count++; }
2051 void inc_java_call_count() { _java_call_count++; }
2052 void inc_inner_loop_count() { _inner_loop_count++; }
2054 int get_call_count () const { return _call_count ; }
2055 int get_float_count () const { return _float_count ; }
2056 int get_double_count() const { return _double_count; }
2057 int get_java_call_count() const { return _java_call_count; }
2058 int get_inner_loop_count() const { return _inner_loop_count; }
2059 };
2061 static bool oop_offset_is_sane(const TypeInstPtr* tp) {
2062 ciInstanceKlass *k = tp->klass()->as_instance_klass();
2063 // Make sure the offset goes inside the instance layout.
2064 return k->contains_field_offset(tp->offset());
2065 // Note that OffsetBot and OffsetTop are very negative.
2066 }
2068 // Eliminate trivially redundant StoreCMs and accumulate their
2069 // precedence edges.
2070 static void eliminate_redundant_card_marks(Node* n) {
2071 assert(n->Opcode() == Op_StoreCM, "expected StoreCM");
2072 if (n->in(MemNode::Address)->outcnt() > 1) {
2073 // There are multiple users of the same address so it might be
2074 // possible to eliminate some of the StoreCMs
2075 Node* mem = n->in(MemNode::Memory);
2076 Node* adr = n->in(MemNode::Address);
2077 Node* val = n->in(MemNode::ValueIn);
2078 Node* prev = n;
2079 bool done = false;
2080 // Walk the chain of StoreCMs eliminating ones that match. As
2081 // long as it's a chain of single users then the optimization is
2082 // safe. Eliminating partially redundant StoreCMs would require
2083 // cloning copies down the other paths.
2084 while (mem->Opcode() == Op_StoreCM && mem->outcnt() == 1 && !done) {
2085 if (adr == mem->in(MemNode::Address) &&
2086 val == mem->in(MemNode::ValueIn)) {
2087 // redundant StoreCM
2088 if (mem->req() > MemNode::OopStore) {
2089 // Hasn't been processed by this code yet.
2090 n->add_prec(mem->in(MemNode::OopStore));
2091 } else {
2092 // Already converted to precedence edge
2093 for (uint i = mem->req(); i < mem->len(); i++) {
2094 // Accumulate any precedence edges
2095 if (mem->in(i) != NULL) {
2096 n->add_prec(mem->in(i));
2097 }
2098 }
2099 // Everything above this point has been processed.
2100 done = true;
2101 }
2102 // Eliminate the previous StoreCM
2103 prev->set_req(MemNode::Memory, mem->in(MemNode::Memory));
2104 assert(mem->outcnt() == 0, "should be dead");
2105 mem->disconnect_inputs(NULL);
2106 } else {
2107 prev = mem;
2108 }
2109 mem = prev->in(MemNode::Memory);
2110 }
2111 }
2112 }
2114 //------------------------------final_graph_reshaping_impl----------------------
2115 // Implement items 1-5 from final_graph_reshaping below.
2116 static void final_graph_reshaping_impl( Node *n, Final_Reshape_Counts &frc ) {
2118 if ( n->outcnt() == 0 ) return; // dead node
2119 uint nop = n->Opcode();
2121 // Check for 2-input instruction with "last use" on right input.
2122 // Swap to left input. Implements item (2).
2123 if( n->req() == 3 && // two-input instruction
2124 n->in(1)->outcnt() > 1 && // left use is NOT a last use
2125 (!n->in(1)->is_Phi() || n->in(1)->in(2) != n) && // it is not data loop
2126 n->in(2)->outcnt() == 1 &&// right use IS a last use
2127 !n->in(2)->is_Con() ) { // right use is not a constant
2128 // Check for commutative opcode
2129 switch( nop ) {
2130 case Op_AddI: case Op_AddF: case Op_AddD: case Op_AddL:
2131 case Op_MaxI: case Op_MinI:
2132 case Op_MulI: case Op_MulF: case Op_MulD: case Op_MulL:
2133 case Op_AndL: case Op_XorL: case Op_OrL:
2134 case Op_AndI: case Op_XorI: case Op_OrI: {
2135 // Move "last use" input to left by swapping inputs
2136 n->swap_edges(1, 2);
2137 break;
2138 }
2139 default:
2140 break;
2141 }
2142 }
2144 #ifdef ASSERT
2145 if( n->is_Mem() ) {
2146 Compile* C = Compile::current();
2147 int alias_idx = C->get_alias_index(n->as_Mem()->adr_type());
2148 assert( n->in(0) != NULL || alias_idx != Compile::AliasIdxRaw ||
2149 // oop will be recorded in oop map if load crosses safepoint
2150 n->is_Load() && (n->as_Load()->bottom_type()->isa_oopptr() ||
2151 LoadNode::is_immutable_value(n->in(MemNode::Address))),
2152 "raw memory operations should have control edge");
2153 }
2154 #endif
2155 // Count FPU ops and common calls, implements item (3)
2156 switch( nop ) {
2157 // Count all float operations that may use FPU
2158 case Op_AddF:
2159 case Op_SubF:
2160 case Op_MulF:
2161 case Op_DivF:
2162 case Op_NegF:
2163 case Op_ModF:
2164 case Op_ConvI2F:
2165 case Op_ConF:
2166 case Op_CmpF:
2167 case Op_CmpF3:
2168 // case Op_ConvL2F: // longs are split into 32-bit halves
2169 frc.inc_float_count();
2170 break;
2172 case Op_ConvF2D:
2173 case Op_ConvD2F:
2174 frc.inc_float_count();
2175 frc.inc_double_count();
2176 break;
2178 // Count all double operations that may use FPU
2179 case Op_AddD:
2180 case Op_SubD:
2181 case Op_MulD:
2182 case Op_DivD:
2183 case Op_NegD:
2184 case Op_ModD:
2185 case Op_ConvI2D:
2186 case Op_ConvD2I:
2187 // case Op_ConvL2D: // handled by leaf call
2188 // case Op_ConvD2L: // handled by leaf call
2189 case Op_ConD:
2190 case Op_CmpD:
2191 case Op_CmpD3:
2192 frc.inc_double_count();
2193 break;
2194 case Op_Opaque1: // Remove Opaque Nodes before matching
2195 case Op_Opaque2: // Remove Opaque Nodes before matching
2196 n->subsume_by(n->in(1));
2197 break;
2198 case Op_CallStaticJava:
2199 case Op_CallJava:
2200 case Op_CallDynamicJava:
2201 frc.inc_java_call_count(); // Count java call site;
2202 case Op_CallRuntime:
2203 case Op_CallLeaf:
2204 case Op_CallLeafNoFP: {
2205 assert( n->is_Call(), "" );
2206 CallNode *call = n->as_Call();
2207 // Count call sites where the FP mode bit would have to be flipped.
2208 // Do not count uncommon runtime calls:
2209 // uncommon_trap, _complete_monitor_locking, _complete_monitor_unlocking,
2210 // _new_Java, _new_typeArray, _new_objArray, _rethrow_Java, ...
2211 if( !call->is_CallStaticJava() || !call->as_CallStaticJava()->_name ) {
2212 frc.inc_call_count(); // Count the call site
2213 } else { // See if uncommon argument is shared
2214 Node *n = call->in(TypeFunc::Parms);
2215 int nop = n->Opcode();
2216 // Clone shared simple arguments to uncommon calls, item (1).
2217 if( n->outcnt() > 1 &&
2218 !n->is_Proj() &&
2219 nop != Op_CreateEx &&
2220 nop != Op_CheckCastPP &&
2221 nop != Op_DecodeN &&
2222 !n->is_Mem() ) {
2223 Node *x = n->clone();
2224 call->set_req( TypeFunc::Parms, x );
2225 }
2226 }
2227 break;
2228 }
2230 case Op_StoreD:
2231 case Op_LoadD:
2232 case Op_LoadD_unaligned:
2233 frc.inc_double_count();
2234 goto handle_mem;
2235 case Op_StoreF:
2236 case Op_LoadF:
2237 frc.inc_float_count();
2238 goto handle_mem;
2240 case Op_StoreCM:
2241 {
2242 // Convert OopStore dependence into precedence edge
2243 Node* prec = n->in(MemNode::OopStore);
2244 n->del_req(MemNode::OopStore);
2245 n->add_prec(prec);
2246 eliminate_redundant_card_marks(n);
2247 }
2249 // fall through
2251 case Op_StoreB:
2252 case Op_StoreC:
2253 case Op_StorePConditional:
2254 case Op_StoreI:
2255 case Op_StoreL:
2256 case Op_StoreIConditional:
2257 case Op_StoreLConditional:
2258 case Op_CompareAndSwapI:
2259 case Op_CompareAndSwapL:
2260 case Op_CompareAndSwapP:
2261 case Op_CompareAndSwapN:
2262 case Op_StoreP:
2263 case Op_StoreN:
2264 case Op_LoadB:
2265 case Op_LoadUB:
2266 case Op_LoadUS:
2267 case Op_LoadI:
2268 case Op_LoadUI2L:
2269 case Op_LoadKlass:
2270 case Op_LoadNKlass:
2271 case Op_LoadL:
2272 case Op_LoadL_unaligned:
2273 case Op_LoadPLocked:
2274 case Op_LoadLLocked:
2275 case Op_LoadP:
2276 case Op_LoadN:
2277 case Op_LoadRange:
2278 case Op_LoadS: {
2279 handle_mem:
2280 #ifdef ASSERT
2281 if( VerifyOptoOopOffsets ) {
2282 assert( n->is_Mem(), "" );
2283 MemNode *mem = (MemNode*)n;
2284 // Check to see if address types have grounded out somehow.
2285 const TypeInstPtr *tp = mem->in(MemNode::Address)->bottom_type()->isa_instptr();
2286 assert( !tp || oop_offset_is_sane(tp), "" );
2287 }
2288 #endif
2289 break;
2290 }
2292 case Op_AddP: { // Assert sane base pointers
2293 Node *addp = n->in(AddPNode::Address);
2294 assert( !addp->is_AddP() ||
2295 addp->in(AddPNode::Base)->is_top() || // Top OK for allocation
2296 addp->in(AddPNode::Base) == n->in(AddPNode::Base),
2297 "Base pointers must match" );
2298 #ifdef _LP64
2299 if (UseCompressedOops &&
2300 addp->Opcode() == Op_ConP &&
2301 addp == n->in(AddPNode::Base) &&
2302 n->in(AddPNode::Offset)->is_Con()) {
2303 // Use addressing with narrow klass to load with offset on x86.
2304 // On sparc loading 32-bits constant and decoding it have less
2305 // instructions (4) then load 64-bits constant (7).
2306 // Do this transformation here since IGVN will convert ConN back to ConP.
2307 const Type* t = addp->bottom_type();
2308 if (t->isa_oopptr()) {
2309 Node* nn = NULL;
2311 // Look for existing ConN node of the same exact type.
2312 Compile* C = Compile::current();
2313 Node* r = C->root();
2314 uint cnt = r->outcnt();
2315 for (uint i = 0; i < cnt; i++) {
2316 Node* m = r->raw_out(i);
2317 if (m!= NULL && m->Opcode() == Op_ConN &&
2318 m->bottom_type()->make_ptr() == t) {
2319 nn = m;
2320 break;
2321 }
2322 }
2323 if (nn != NULL) {
2324 // Decode a narrow oop to match address
2325 // [R12 + narrow_oop_reg<<3 + offset]
2326 nn = new (C, 2) DecodeNNode(nn, t);
2327 n->set_req(AddPNode::Base, nn);
2328 n->set_req(AddPNode::Address, nn);
2329 if (addp->outcnt() == 0) {
2330 addp->disconnect_inputs(NULL);
2331 }
2332 }
2333 }
2334 }
2335 #endif
2336 break;
2337 }
2339 #ifdef _LP64
2340 case Op_CastPP:
2341 if (n->in(1)->is_DecodeN() && Matcher::gen_narrow_oop_implicit_null_checks()) {
2342 Compile* C = Compile::current();
2343 Node* in1 = n->in(1);
2344 const Type* t = n->bottom_type();
2345 Node* new_in1 = in1->clone();
2346 new_in1->as_DecodeN()->set_type(t);
2348 if (!Matcher::narrow_oop_use_complex_address()) {
2349 //
2350 // x86, ARM and friends can handle 2 adds in addressing mode
2351 // and Matcher can fold a DecodeN node into address by using
2352 // a narrow oop directly and do implicit NULL check in address:
2353 //
2354 // [R12 + narrow_oop_reg<<3 + offset]
2355 // NullCheck narrow_oop_reg
2356 //
2357 // On other platforms (Sparc) we have to keep new DecodeN node and
2358 // use it to do implicit NULL check in address:
2359 //
2360 // decode_not_null narrow_oop_reg, base_reg
2361 // [base_reg + offset]
2362 // NullCheck base_reg
2363 //
2364 // Pin the new DecodeN node to non-null path on these platform (Sparc)
2365 // to keep the information to which NULL check the new DecodeN node
2366 // corresponds to use it as value in implicit_null_check().
2367 //
2368 new_in1->set_req(0, n->in(0));
2369 }
2371 n->subsume_by(new_in1);
2372 if (in1->outcnt() == 0) {
2373 in1->disconnect_inputs(NULL);
2374 }
2375 }
2376 break;
2378 case Op_CmpP:
2379 // Do this transformation here to preserve CmpPNode::sub() and
2380 // other TypePtr related Ideal optimizations (for example, ptr nullness).
2381 if (n->in(1)->is_DecodeN() || n->in(2)->is_DecodeN()) {
2382 Node* in1 = n->in(1);
2383 Node* in2 = n->in(2);
2384 if (!in1->is_DecodeN()) {
2385 in2 = in1;
2386 in1 = n->in(2);
2387 }
2388 assert(in1->is_DecodeN(), "sanity");
2390 Compile* C = Compile::current();
2391 Node* new_in2 = NULL;
2392 if (in2->is_DecodeN()) {
2393 new_in2 = in2->in(1);
2394 } else if (in2->Opcode() == Op_ConP) {
2395 const Type* t = in2->bottom_type();
2396 if (t == TypePtr::NULL_PTR) {
2397 // Don't convert CmpP null check into CmpN if compressed
2398 // oops implicit null check is not generated.
2399 // This will allow to generate normal oop implicit null check.
2400 if (Matcher::gen_narrow_oop_implicit_null_checks())
2401 new_in2 = ConNode::make(C, TypeNarrowOop::NULL_PTR);
2402 //
2403 // This transformation together with CastPP transformation above
2404 // will generated code for implicit NULL checks for compressed oops.
2405 //
2406 // The original code after Optimize()
2407 //
2408 // LoadN memory, narrow_oop_reg
2409 // decode narrow_oop_reg, base_reg
2410 // CmpP base_reg, NULL
2411 // CastPP base_reg // NotNull
2412 // Load [base_reg + offset], val_reg
2413 //
2414 // after these transformations will be
2415 //
2416 // LoadN memory, narrow_oop_reg
2417 // CmpN narrow_oop_reg, NULL
2418 // decode_not_null narrow_oop_reg, base_reg
2419 // Load [base_reg + offset], val_reg
2420 //
2421 // and the uncommon path (== NULL) will use narrow_oop_reg directly
2422 // since narrow oops can be used in debug info now (see the code in
2423 // final_graph_reshaping_walk()).
2424 //
2425 // At the end the code will be matched to
2426 // on x86:
2427 //
2428 // Load_narrow_oop memory, narrow_oop_reg
2429 // Load [R12 + narrow_oop_reg<<3 + offset], val_reg
2430 // NullCheck narrow_oop_reg
2431 //
2432 // and on sparc:
2433 //
2434 // Load_narrow_oop memory, narrow_oop_reg
2435 // decode_not_null narrow_oop_reg, base_reg
2436 // Load [base_reg + offset], val_reg
2437 // NullCheck base_reg
2438 //
2439 } else if (t->isa_oopptr()) {
2440 new_in2 = ConNode::make(C, t->make_narrowoop());
2441 }
2442 }
2443 if (new_in2 != NULL) {
2444 Node* cmpN = new (C, 3) CmpNNode(in1->in(1), new_in2);
2445 n->subsume_by( cmpN );
2446 if (in1->outcnt() == 0) {
2447 in1->disconnect_inputs(NULL);
2448 }
2449 if (in2->outcnt() == 0) {
2450 in2->disconnect_inputs(NULL);
2451 }
2452 }
2453 }
2454 break;
2456 case Op_DecodeN:
2457 assert(!n->in(1)->is_EncodeP(), "should be optimized out");
2458 // DecodeN could be pinned when it can't be fold into
2459 // an address expression, see the code for Op_CastPP above.
2460 assert(n->in(0) == NULL || !Matcher::narrow_oop_use_complex_address(), "no control");
2461 break;
2463 case Op_EncodeP: {
2464 Node* in1 = n->in(1);
2465 if (in1->is_DecodeN()) {
2466 n->subsume_by(in1->in(1));
2467 } else if (in1->Opcode() == Op_ConP) {
2468 Compile* C = Compile::current();
2469 const Type* t = in1->bottom_type();
2470 if (t == TypePtr::NULL_PTR) {
2471 n->subsume_by(ConNode::make(C, TypeNarrowOop::NULL_PTR));
2472 } else if (t->isa_oopptr()) {
2473 n->subsume_by(ConNode::make(C, t->make_narrowoop()));
2474 }
2475 }
2476 if (in1->outcnt() == 0) {
2477 in1->disconnect_inputs(NULL);
2478 }
2479 break;
2480 }
2482 case Op_Proj: {
2483 if (OptimizeStringConcat) {
2484 ProjNode* p = n->as_Proj();
2485 if (p->_is_io_use) {
2486 // Separate projections were used for the exception path which
2487 // are normally removed by a late inline. If it wasn't inlined
2488 // then they will hang around and should just be replaced with
2489 // the original one.
2490 Node* proj = NULL;
2491 // Replace with just one
2492 for (SimpleDUIterator i(p->in(0)); i.has_next(); i.next()) {
2493 Node *use = i.get();
2494 if (use->is_Proj() && p != use && use->as_Proj()->_con == p->_con) {
2495 proj = use;
2496 break;
2497 }
2498 }
2499 assert(p != NULL, "must be found");
2500 p->subsume_by(proj);
2501 }
2502 }
2503 break;
2504 }
2506 case Op_Phi:
2507 if (n->as_Phi()->bottom_type()->isa_narrowoop()) {
2508 // The EncodeP optimization may create Phi with the same edges
2509 // for all paths. It is not handled well by Register Allocator.
2510 Node* unique_in = n->in(1);
2511 assert(unique_in != NULL, "");
2512 uint cnt = n->req();
2513 for (uint i = 2; i < cnt; i++) {
2514 Node* m = n->in(i);
2515 assert(m != NULL, "");
2516 if (unique_in != m)
2517 unique_in = NULL;
2518 }
2519 if (unique_in != NULL) {
2520 n->subsume_by(unique_in);
2521 }
2522 }
2523 break;
2525 #endif
2527 case Op_ModI:
2528 if (UseDivMod) {
2529 // Check if a%b and a/b both exist
2530 Node* d = n->find_similar(Op_DivI);
2531 if (d) {
2532 // Replace them with a fused divmod if supported
2533 Compile* C = Compile::current();
2534 if (Matcher::has_match_rule(Op_DivModI)) {
2535 DivModINode* divmod = DivModINode::make(C, n);
2536 d->subsume_by(divmod->div_proj());
2537 n->subsume_by(divmod->mod_proj());
2538 } else {
2539 // replace a%b with a-((a/b)*b)
2540 Node* mult = new (C, 3) MulINode(d, d->in(2));
2541 Node* sub = new (C, 3) SubINode(d->in(1), mult);
2542 n->subsume_by( sub );
2543 }
2544 }
2545 }
2546 break;
2548 case Op_ModL:
2549 if (UseDivMod) {
2550 // Check if a%b and a/b both exist
2551 Node* d = n->find_similar(Op_DivL);
2552 if (d) {
2553 // Replace them with a fused divmod if supported
2554 Compile* C = Compile::current();
2555 if (Matcher::has_match_rule(Op_DivModL)) {
2556 DivModLNode* divmod = DivModLNode::make(C, n);
2557 d->subsume_by(divmod->div_proj());
2558 n->subsume_by(divmod->mod_proj());
2559 } else {
2560 // replace a%b with a-((a/b)*b)
2561 Node* mult = new (C, 3) MulLNode(d, d->in(2));
2562 Node* sub = new (C, 3) SubLNode(d->in(1), mult);
2563 n->subsume_by( sub );
2564 }
2565 }
2566 }
2567 break;
2569 case Op_Load16B:
2570 case Op_Load8B:
2571 case Op_Load4B:
2572 case Op_Load8S:
2573 case Op_Load4S:
2574 case Op_Load2S:
2575 case Op_Load8C:
2576 case Op_Load4C:
2577 case Op_Load2C:
2578 case Op_Load4I:
2579 case Op_Load2I:
2580 case Op_Load2L:
2581 case Op_Load4F:
2582 case Op_Load2F:
2583 case Op_Load2D:
2584 case Op_Store16B:
2585 case Op_Store8B:
2586 case Op_Store4B:
2587 case Op_Store8C:
2588 case Op_Store4C:
2589 case Op_Store2C:
2590 case Op_Store4I:
2591 case Op_Store2I:
2592 case Op_Store2L:
2593 case Op_Store4F:
2594 case Op_Store2F:
2595 case Op_Store2D:
2596 break;
2598 case Op_PackB:
2599 case Op_PackS:
2600 case Op_PackC:
2601 case Op_PackI:
2602 case Op_PackF:
2603 case Op_PackL:
2604 case Op_PackD:
2605 if (n->req()-1 > 2) {
2606 // Replace many operand PackNodes with a binary tree for matching
2607 PackNode* p = (PackNode*) n;
2608 Node* btp = p->binaryTreePack(Compile::current(), 1, n->req());
2609 n->subsume_by(btp);
2610 }
2611 break;
2612 case Op_Loop:
2613 case Op_CountedLoop:
2614 if (n->as_Loop()->is_inner_loop()) {
2615 frc.inc_inner_loop_count();
2616 }
2617 break;
2618 case Op_LShiftI:
2619 case Op_RShiftI:
2620 case Op_URShiftI:
2621 case Op_LShiftL:
2622 case Op_RShiftL:
2623 case Op_URShiftL:
2624 if (Matcher::need_masked_shift_count) {
2625 // The cpu's shift instructions don't restrict the count to the
2626 // lower 5/6 bits. We need to do the masking ourselves.
2627 Node* in2 = n->in(2);
2628 juint mask = (n->bottom_type() == TypeInt::INT) ? (BitsPerInt - 1) : (BitsPerLong - 1);
2629 const TypeInt* t = in2->find_int_type();
2630 if (t != NULL && t->is_con()) {
2631 juint shift = t->get_con();
2632 if (shift > mask) { // Unsigned cmp
2633 Compile* C = Compile::current();
2634 n->set_req(2, ConNode::make(C, TypeInt::make(shift & mask)));
2635 }
2636 } else {
2637 if (t == NULL || t->_lo < 0 || t->_hi > (int)mask) {
2638 Compile* C = Compile::current();
2639 Node* shift = new (C, 3) AndINode(in2, ConNode::make(C, TypeInt::make(mask)));
2640 n->set_req(2, shift);
2641 }
2642 }
2643 if (in2->outcnt() == 0) { // Remove dead node
2644 in2->disconnect_inputs(NULL);
2645 }
2646 }
2647 break;
2648 default:
2649 assert( !n->is_Call(), "" );
2650 assert( !n->is_Mem(), "" );
2651 break;
2652 }
2654 // Collect CFG split points
2655 if (n->is_MultiBranch())
2656 frc._tests.push(n);
2657 }
2659 //------------------------------final_graph_reshaping_walk---------------------
2660 // Replacing Opaque nodes with their input in final_graph_reshaping_impl(),
2661 // requires that the walk visits a node's inputs before visiting the node.
2662 static void final_graph_reshaping_walk( Node_Stack &nstack, Node *root, Final_Reshape_Counts &frc ) {
2663 ResourceArea *area = Thread::current()->resource_area();
2664 Unique_Node_List sfpt(area);
2666 frc._visited.set(root->_idx); // first, mark node as visited
2667 uint cnt = root->req();
2668 Node *n = root;
2669 uint i = 0;
2670 while (true) {
2671 if (i < cnt) {
2672 // Place all non-visited non-null inputs onto stack
2673 Node* m = n->in(i);
2674 ++i;
2675 if (m != NULL && !frc._visited.test_set(m->_idx)) {
2676 if (m->is_SafePoint() && m->as_SafePoint()->jvms() != NULL)
2677 sfpt.push(m);
2678 cnt = m->req();
2679 nstack.push(n, i); // put on stack parent and next input's index
2680 n = m;
2681 i = 0;
2682 }
2683 } else {
2684 // Now do post-visit work
2685 final_graph_reshaping_impl( n, frc );
2686 if (nstack.is_empty())
2687 break; // finished
2688 n = nstack.node(); // Get node from stack
2689 cnt = n->req();
2690 i = nstack.index();
2691 nstack.pop(); // Shift to the next node on stack
2692 }
2693 }
2695 // Skip next transformation if compressed oops are not used.
2696 if (!UseCompressedOops || !Matcher::gen_narrow_oop_implicit_null_checks())
2697 return;
2699 // Go over safepoints nodes to skip DecodeN nodes for debug edges.
2700 // It could be done for an uncommon traps or any safepoints/calls
2701 // if the DecodeN node is referenced only in a debug info.
2702 while (sfpt.size() > 0) {
2703 n = sfpt.pop();
2704 JVMState *jvms = n->as_SafePoint()->jvms();
2705 assert(jvms != NULL, "sanity");
2706 int start = jvms->debug_start();
2707 int end = n->req();
2708 bool is_uncommon = (n->is_CallStaticJava() &&
2709 n->as_CallStaticJava()->uncommon_trap_request() != 0);
2710 for (int j = start; j < end; j++) {
2711 Node* in = n->in(j);
2712 if (in->is_DecodeN()) {
2713 bool safe_to_skip = true;
2714 if (!is_uncommon ) {
2715 // Is it safe to skip?
2716 for (uint i = 0; i < in->outcnt(); i++) {
2717 Node* u = in->raw_out(i);
2718 if (!u->is_SafePoint() ||
2719 u->is_Call() && u->as_Call()->has_non_debug_use(n)) {
2720 safe_to_skip = false;
2721 }
2722 }
2723 }
2724 if (safe_to_skip) {
2725 n->set_req(j, in->in(1));
2726 }
2727 if (in->outcnt() == 0) {
2728 in->disconnect_inputs(NULL);
2729 }
2730 }
2731 }
2732 }
2733 }
2735 //------------------------------final_graph_reshaping--------------------------
2736 // Final Graph Reshaping.
2737 //
2738 // (1) Clone simple inputs to uncommon calls, so they can be scheduled late
2739 // and not commoned up and forced early. Must come after regular
2740 // optimizations to avoid GVN undoing the cloning. Clone constant
2741 // inputs to Loop Phis; these will be split by the allocator anyways.
2742 // Remove Opaque nodes.
2743 // (2) Move last-uses by commutative operations to the left input to encourage
2744 // Intel update-in-place two-address operations and better register usage
2745 // on RISCs. Must come after regular optimizations to avoid GVN Ideal
2746 // calls canonicalizing them back.
2747 // (3) Count the number of double-precision FP ops, single-precision FP ops
2748 // and call sites. On Intel, we can get correct rounding either by
2749 // forcing singles to memory (requires extra stores and loads after each
2750 // FP bytecode) or we can set a rounding mode bit (requires setting and
2751 // clearing the mode bit around call sites). The mode bit is only used
2752 // if the relative frequency of single FP ops to calls is low enough.
2753 // This is a key transform for SPEC mpeg_audio.
2754 // (4) Detect infinite loops; blobs of code reachable from above but not
2755 // below. Several of the Code_Gen algorithms fail on such code shapes,
2756 // so we simply bail out. Happens a lot in ZKM.jar, but also happens
2757 // from time to time in other codes (such as -Xcomp finalizer loops, etc).
2758 // Detection is by looking for IfNodes where only 1 projection is
2759 // reachable from below or CatchNodes missing some targets.
2760 // (5) Assert for insane oop offsets in debug mode.
2762 bool Compile::final_graph_reshaping() {
2763 // an infinite loop may have been eliminated by the optimizer,
2764 // in which case the graph will be empty.
2765 if (root()->req() == 1) {
2766 record_method_not_compilable("trivial infinite loop");
2767 return true;
2768 }
2770 Final_Reshape_Counts frc;
2772 // Visit everybody reachable!
2773 // Allocate stack of size C->unique()/2 to avoid frequent realloc
2774 Node_Stack nstack(unique() >> 1);
2775 final_graph_reshaping_walk(nstack, root(), frc);
2777 // Check for unreachable (from below) code (i.e., infinite loops).
2778 for( uint i = 0; i < frc._tests.size(); i++ ) {
2779 MultiBranchNode *n = frc._tests[i]->as_MultiBranch();
2780 // Get number of CFG targets.
2781 // Note that PCTables include exception targets after calls.
2782 uint required_outcnt = n->required_outcnt();
2783 if (n->outcnt() != required_outcnt) {
2784 // Check for a few special cases. Rethrow Nodes never take the
2785 // 'fall-thru' path, so expected kids is 1 less.
2786 if (n->is_PCTable() && n->in(0) && n->in(0)->in(0)) {
2787 if (n->in(0)->in(0)->is_Call()) {
2788 CallNode *call = n->in(0)->in(0)->as_Call();
2789 if (call->entry_point() == OptoRuntime::rethrow_stub()) {
2790 required_outcnt--; // Rethrow always has 1 less kid
2791 } else if (call->req() > TypeFunc::Parms &&
2792 call->is_CallDynamicJava()) {
2793 // Check for null receiver. In such case, the optimizer has
2794 // detected that the virtual call will always result in a null
2795 // pointer exception. The fall-through projection of this CatchNode
2796 // will not be populated.
2797 Node *arg0 = call->in(TypeFunc::Parms);
2798 if (arg0->is_Type() &&
2799 arg0->as_Type()->type()->higher_equal(TypePtr::NULL_PTR)) {
2800 required_outcnt--;
2801 }
2802 } else if (call->entry_point() == OptoRuntime::new_array_Java() &&
2803 call->req() > TypeFunc::Parms+1 &&
2804 call->is_CallStaticJava()) {
2805 // Check for negative array length. In such case, the optimizer has
2806 // detected that the allocation attempt will always result in an
2807 // exception. There is no fall-through projection of this CatchNode .
2808 Node *arg1 = call->in(TypeFunc::Parms+1);
2809 if (arg1->is_Type() &&
2810 arg1->as_Type()->type()->join(TypeInt::POS)->empty()) {
2811 required_outcnt--;
2812 }
2813 }
2814 }
2815 }
2816 // Recheck with a better notion of 'required_outcnt'
2817 if (n->outcnt() != required_outcnt) {
2818 record_method_not_compilable("malformed control flow");
2819 return true; // Not all targets reachable!
2820 }
2821 }
2822 // Check that I actually visited all kids. Unreached kids
2823 // must be infinite loops.
2824 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++)
2825 if (!frc._visited.test(n->fast_out(j)->_idx)) {
2826 record_method_not_compilable("infinite loop");
2827 return true; // Found unvisited kid; must be unreach
2828 }
2829 }
2831 // If original bytecodes contained a mixture of floats and doubles
2832 // check if the optimizer has made it homogenous, item (3).
2833 if( Use24BitFPMode && Use24BitFP && UseSSE == 0 &&
2834 frc.get_float_count() > 32 &&
2835 frc.get_double_count() == 0 &&
2836 (10 * frc.get_call_count() < frc.get_float_count()) ) {
2837 set_24_bit_selection_and_mode( false, true );
2838 }
2840 set_java_calls(frc.get_java_call_count());
2841 set_inner_loops(frc.get_inner_loop_count());
2843 // No infinite loops, no reason to bail out.
2844 return false;
2845 }
2847 //-----------------------------too_many_traps----------------------------------
2848 // Report if there are too many traps at the current method and bci.
2849 // Return true if there was a trap, and/or PerMethodTrapLimit is exceeded.
2850 bool Compile::too_many_traps(ciMethod* method,
2851 int bci,
2852 Deoptimization::DeoptReason reason) {
2853 ciMethodData* md = method->method_data();
2854 if (md->is_empty()) {
2855 // Assume the trap has not occurred, or that it occurred only
2856 // because of a transient condition during start-up in the interpreter.
2857 return false;
2858 }
2859 if (md->has_trap_at(bci, reason) != 0) {
2860 // Assume PerBytecodeTrapLimit==0, for a more conservative heuristic.
2861 // Also, if there are multiple reasons, or if there is no per-BCI record,
2862 // assume the worst.
2863 if (log())
2864 log()->elem("observe trap='%s' count='%d'",
2865 Deoptimization::trap_reason_name(reason),
2866 md->trap_count(reason));
2867 return true;
2868 } else {
2869 // Ignore method/bci and see if there have been too many globally.
2870 return too_many_traps(reason, md);
2871 }
2872 }
2874 // Less-accurate variant which does not require a method and bci.
2875 bool Compile::too_many_traps(Deoptimization::DeoptReason reason,
2876 ciMethodData* logmd) {
2877 if (trap_count(reason) >= (uint)PerMethodTrapLimit) {
2878 // Too many traps globally.
2879 // Note that we use cumulative trap_count, not just md->trap_count.
2880 if (log()) {
2881 int mcount = (logmd == NULL)? -1: (int)logmd->trap_count(reason);
2882 log()->elem("observe trap='%s' count='0' mcount='%d' ccount='%d'",
2883 Deoptimization::trap_reason_name(reason),
2884 mcount, trap_count(reason));
2885 }
2886 return true;
2887 } else {
2888 // The coast is clear.
2889 return false;
2890 }
2891 }
2893 //--------------------------too_many_recompiles--------------------------------
2894 // Report if there are too many recompiles at the current method and bci.
2895 // Consults PerBytecodeRecompilationCutoff and PerMethodRecompilationCutoff.
2896 // Is not eager to return true, since this will cause the compiler to use
2897 // Action_none for a trap point, to avoid too many recompilations.
2898 bool Compile::too_many_recompiles(ciMethod* method,
2899 int bci,
2900 Deoptimization::DeoptReason reason) {
2901 ciMethodData* md = method->method_data();
2902 if (md->is_empty()) {
2903 // Assume the trap has not occurred, or that it occurred only
2904 // because of a transient condition during start-up in the interpreter.
2905 return false;
2906 }
2907 // Pick a cutoff point well within PerBytecodeRecompilationCutoff.
2908 uint bc_cutoff = (uint) PerBytecodeRecompilationCutoff / 8;
2909 uint m_cutoff = (uint) PerMethodRecompilationCutoff / 2 + 1; // not zero
2910 Deoptimization::DeoptReason per_bc_reason
2911 = Deoptimization::reason_recorded_per_bytecode_if_any(reason);
2912 if ((per_bc_reason == Deoptimization::Reason_none
2913 || md->has_trap_at(bci, reason) != 0)
2914 // The trap frequency measure we care about is the recompile count:
2915 && md->trap_recompiled_at(bci)
2916 && md->overflow_recompile_count() >= bc_cutoff) {
2917 // Do not emit a trap here if it has already caused recompilations.
2918 // Also, if there are multiple reasons, or if there is no per-BCI record,
2919 // assume the worst.
2920 if (log())
2921 log()->elem("observe trap='%s recompiled' count='%d' recompiles2='%d'",
2922 Deoptimization::trap_reason_name(reason),
2923 md->trap_count(reason),
2924 md->overflow_recompile_count());
2925 return true;
2926 } else if (trap_count(reason) != 0
2927 && decompile_count() >= m_cutoff) {
2928 // Too many recompiles globally, and we have seen this sort of trap.
2929 // Use cumulative decompile_count, not just md->decompile_count.
2930 if (log())
2931 log()->elem("observe trap='%s' count='%d' mcount='%d' decompiles='%d' mdecompiles='%d'",
2932 Deoptimization::trap_reason_name(reason),
2933 md->trap_count(reason), trap_count(reason),
2934 md->decompile_count(), decompile_count());
2935 return true;
2936 } else {
2937 // The coast is clear.
2938 return false;
2939 }
2940 }
2943 #ifndef PRODUCT
2944 //------------------------------verify_graph_edges---------------------------
2945 // Walk the Graph and verify that there is a one-to-one correspondence
2946 // between Use-Def edges and Def-Use edges in the graph.
2947 void Compile::verify_graph_edges(bool no_dead_code) {
2948 if (VerifyGraphEdges) {
2949 ResourceArea *area = Thread::current()->resource_area();
2950 Unique_Node_List visited(area);
2951 // Call recursive graph walk to check edges
2952 _root->verify_edges(visited);
2953 if (no_dead_code) {
2954 // Now make sure that no visited node is used by an unvisited node.
2955 bool dead_nodes = 0;
2956 Unique_Node_List checked(area);
2957 while (visited.size() > 0) {
2958 Node* n = visited.pop();
2959 checked.push(n);
2960 for (uint i = 0; i < n->outcnt(); i++) {
2961 Node* use = n->raw_out(i);
2962 if (checked.member(use)) continue; // already checked
2963 if (visited.member(use)) continue; // already in the graph
2964 if (use->is_Con()) continue; // a dead ConNode is OK
2965 // At this point, we have found a dead node which is DU-reachable.
2966 if (dead_nodes++ == 0)
2967 tty->print_cr("*** Dead nodes reachable via DU edges:");
2968 use->dump(2);
2969 tty->print_cr("---");
2970 checked.push(use); // No repeats; pretend it is now checked.
2971 }
2972 }
2973 assert(dead_nodes == 0, "using nodes must be reachable from root");
2974 }
2975 }
2976 }
2977 #endif
2979 // The Compile object keeps track of failure reasons separately from the ciEnv.
2980 // This is required because there is not quite a 1-1 relation between the
2981 // ciEnv and its compilation task and the Compile object. Note that one
2982 // ciEnv might use two Compile objects, if C2Compiler::compile_method decides
2983 // to backtrack and retry without subsuming loads. Other than this backtracking
2984 // behavior, the Compile's failure reason is quietly copied up to the ciEnv
2985 // by the logic in C2Compiler.
2986 void Compile::record_failure(const char* reason) {
2987 if (log() != NULL) {
2988 log()->elem("failure reason='%s' phase='compile'", reason);
2989 }
2990 if (_failure_reason == NULL) {
2991 // Record the first failure reason.
2992 _failure_reason = reason;
2993 }
2994 if (!C->failure_reason_is(C2Compiler::retry_no_subsuming_loads())) {
2995 C->print_method(_failure_reason);
2996 }
2997 _root = NULL; // flush the graph, too
2998 }
3000 Compile::TracePhase::TracePhase(const char* name, elapsedTimer* accumulator, bool dolog)
3001 : TraceTime(NULL, accumulator, false NOT_PRODUCT( || TimeCompiler ), false)
3002 {
3003 if (dolog) {
3004 C = Compile::current();
3005 _log = C->log();
3006 } else {
3007 C = NULL;
3008 _log = NULL;
3009 }
3010 if (_log != NULL) {
3011 _log->begin_head("phase name='%s' nodes='%d'", name, C->unique());
3012 _log->stamp();
3013 _log->end_head();
3014 }
3015 }
3017 Compile::TracePhase::~TracePhase() {
3018 if (_log != NULL) {
3019 _log->done("phase nodes='%d'", C->unique());
3020 }
3021 }
3023 //=============================================================================
3024 // Two Constant's are equal when the type and the value are equal.
3025 bool Compile::Constant::operator==(const Constant& other) {
3026 if (type() != other.type() ) return false;
3027 if (can_be_reused() != other.can_be_reused()) return false;
3028 // For floating point values we compare the bit pattern.
3029 switch (type()) {
3030 case T_FLOAT: return (_value.i == other._value.i);
3031 case T_LONG:
3032 case T_DOUBLE: return (_value.j == other._value.j);
3033 case T_OBJECT:
3034 case T_ADDRESS: return (_value.l == other._value.l);
3035 case T_VOID: return (_value.l == other._value.l); // jump-table entries
3036 default: ShouldNotReachHere();
3037 }
3038 return false;
3039 }
3041 // Emit constants grouped in the following order:
3042 static BasicType type_order[] = {
3043 T_FLOAT, // 32-bit
3044 T_OBJECT, // 32 or 64-bit
3045 T_ADDRESS, // 32 or 64-bit
3046 T_DOUBLE, // 64-bit
3047 T_LONG, // 64-bit
3048 T_VOID, // 32 or 64-bit (jump-tables are at the end of the constant table for code emission reasons)
3049 T_ILLEGAL
3050 };
3052 static int type_to_size_in_bytes(BasicType t) {
3053 switch (t) {
3054 case T_LONG: return sizeof(jlong );
3055 case T_FLOAT: return sizeof(jfloat );
3056 case T_DOUBLE: return sizeof(jdouble);
3057 // We use T_VOID as marker for jump-table entries (labels) which
3058 // need an interal word relocation.
3059 case T_VOID:
3060 case T_ADDRESS:
3061 case T_OBJECT: return sizeof(jobject);
3062 }
3064 ShouldNotReachHere();
3065 return -1;
3066 }
3068 void Compile::ConstantTable::calculate_offsets_and_size() {
3069 int size = 0;
3070 for (int t = 0; type_order[t] != T_ILLEGAL; t++) {
3071 BasicType type = type_order[t];
3073 for (int i = 0; i < _constants.length(); i++) {
3074 Constant con = _constants.at(i);
3075 if (con.type() != type) continue; // Skip other types.
3077 // Align size for type.
3078 int typesize = type_to_size_in_bytes(con.type());
3079 size = align_size_up(size, typesize);
3081 // Set offset.
3082 con.set_offset(size);
3083 _constants.at_put(i, con);
3085 // Add type size.
3086 size = size + typesize;
3087 }
3088 }
3090 // Align size up to the next section start (which is insts; see
3091 // CodeBuffer::align_at_start).
3092 assert(_size == -1, "already set?");
3093 _size = align_size_up(size, CodeEntryAlignment);
3095 if (Matcher::constant_table_absolute_addressing) {
3096 set_table_base_offset(0); // No table base offset required
3097 } else {
3098 if (UseRDPCForConstantTableBase) {
3099 // table base offset is set in MachConstantBaseNode::emit
3100 } else {
3101 // When RDPC is not used, the table base is set into the middle of
3102 // the constant table.
3103 int half_size = _size / 2;
3104 assert(half_size * 2 == _size, "sanity");
3105 set_table_base_offset(-half_size);
3106 }
3107 }
3108 }
3110 void Compile::ConstantTable::emit(CodeBuffer& cb) {
3111 MacroAssembler _masm(&cb);
3112 for (int t = 0; type_order[t] != T_ILLEGAL; t++) {
3113 BasicType type = type_order[t];
3115 for (int i = 0; i < _constants.length(); i++) {
3116 Constant con = _constants.at(i);
3117 if (con.type() != type) continue; // Skip other types.
3119 address constant_addr;
3120 switch (con.type()) {
3121 case T_LONG: constant_addr = _masm.long_constant( con.get_jlong() ); break;
3122 case T_FLOAT: constant_addr = _masm.float_constant( con.get_jfloat() ); break;
3123 case T_DOUBLE: constant_addr = _masm.double_constant(con.get_jdouble()); break;
3124 case T_OBJECT: {
3125 jobject obj = con.get_jobject();
3126 int oop_index = _masm.oop_recorder()->find_index(obj);
3127 constant_addr = _masm.address_constant((address) obj, oop_Relocation::spec(oop_index));
3128 break;
3129 }
3130 case T_ADDRESS: {
3131 address addr = (address) con.get_jobject();
3132 constant_addr = _masm.address_constant(addr);
3133 break;
3134 }
3135 // We use T_VOID as marker for jump-table entries (labels) which
3136 // need an interal word relocation.
3137 case T_VOID: {
3138 // Write a dummy word. The real value is filled in later
3139 // in fill_jump_table_in_constant_table.
3140 address addr = (address) con.get_jobject();
3141 constant_addr = _masm.address_constant(addr);
3142 break;
3143 }
3144 default: ShouldNotReachHere();
3145 }
3146 assert(constant_addr != NULL, "consts section too small");
3147 assert((constant_addr - _masm.code()->consts()->start()) == con.offset(), err_msg("must be: %d == %d", constant_addr - _masm.code()->consts()->start(), con.offset()));
3148 }
3149 }
3150 }
3152 int Compile::ConstantTable::find_offset(Constant& con) const {
3153 int idx = _constants.find(con);
3154 assert(idx != -1, "constant must be in constant table");
3155 int offset = _constants.at(idx).offset();
3156 assert(offset != -1, "constant table not emitted yet?");
3157 return offset;
3158 }
3160 void Compile::ConstantTable::add(Constant& con) {
3161 if (con.can_be_reused()) {
3162 int idx = _constants.find(con);
3163 if (idx != -1 && _constants.at(idx).can_be_reused()) {
3164 return;
3165 }
3166 }
3167 (void) _constants.append(con);
3168 }
3170 Compile::Constant Compile::ConstantTable::add(BasicType type, jvalue value) {
3171 Constant con(type, value);
3172 add(con);
3173 return con;
3174 }
3176 Compile::Constant Compile::ConstantTable::add(MachOper* oper) {
3177 jvalue value;
3178 BasicType type = oper->type()->basic_type();
3179 switch (type) {
3180 case T_LONG: value.j = oper->constantL(); break;
3181 case T_FLOAT: value.f = oper->constantF(); break;
3182 case T_DOUBLE: value.d = oper->constantD(); break;
3183 case T_OBJECT:
3184 case T_ADDRESS: value.l = (jobject) oper->constant(); break;
3185 default: ShouldNotReachHere();
3186 }
3187 return add(type, value);
3188 }
3190 Compile::Constant Compile::ConstantTable::allocate_jump_table(MachConstantNode* n) {
3191 jvalue value;
3192 // We can use the node pointer here to identify the right jump-table
3193 // as this method is called from Compile::Fill_buffer right before
3194 // the MachNodes are emitted and the jump-table is filled (means the
3195 // MachNode pointers do not change anymore).
3196 value.l = (jobject) n;
3197 Constant con(T_VOID, value, false); // Labels of a jump-table cannot be reused.
3198 for (uint i = 0; i < n->outcnt(); i++) {
3199 add(con);
3200 }
3201 return con;
3202 }
3204 void Compile::ConstantTable::fill_jump_table(CodeBuffer& cb, MachConstantNode* n, GrowableArray<Label*> labels) const {
3205 // If called from Compile::scratch_emit_size do nothing.
3206 if (Compile::current()->in_scratch_emit_size()) return;
3208 assert(labels.is_nonempty(), "must be");
3209 assert((uint) labels.length() == n->outcnt(), err_msg("must be equal: %d == %d", labels.length(), n->outcnt()));
3211 // Since MachConstantNode::constant_offset() also contains
3212 // table_base_offset() we need to subtract the table_base_offset()
3213 // to get the plain offset into the constant table.
3214 int offset = n->constant_offset() - table_base_offset();
3216 MacroAssembler _masm(&cb);
3217 address* jump_table_base = (address*) (_masm.code()->consts()->start() + offset);
3219 for (int i = 0; i < labels.length(); i++) {
3220 address* constant_addr = &jump_table_base[i];
3221 assert(*constant_addr == (address) n, "all jump-table entries must contain node pointer");
3222 *constant_addr = cb.consts()->target(*labels.at(i), (address) constant_addr);
3223 cb.consts()->relocate((address) constant_addr, relocInfo::internal_word_type);
3224 }
3225 }