// Copyright 2010 the V8 project authors. All rights reserved. // Redistribution and use in source and binary forms, with or without // modification, are permitted provided that the following conditions are // met: // // * Redistributions of source code must retain the above copyright // notice, this list of conditions and the following disclaimer. // * Redistributions in binary form must reproduce the above // copyright notice, this list of conditions and the following // disclaimer in the documentation and/or other materials provided // with the distribution. // * Neither the name of Google Inc. nor the names of its // contributors may be used to endorse or promote products derived // from this software without specific prior written permission. // // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. #include "v8.h" #include "codegen.h" #include "deoptimizer.h" #include "disasm.h" #include "full-codegen.h" #include "global-handles.h" #include "macro-assembler.h" #include "prettyprinter.h" namespace v8 { namespace internal { LargeObjectChunk* Deoptimizer::eager_deoptimization_entry_code_ = NULL; LargeObjectChunk* Deoptimizer::lazy_deoptimization_entry_code_ = NULL; Deoptimizer* Deoptimizer::current_ = NULL; DeoptimizingCodeListNode* Deoptimizer::deoptimizing_code_list_ = NULL; Deoptimizer* Deoptimizer::New(JSFunction* function, BailoutType type, unsigned bailout_id, Address from, int fp_to_sp_delta) { Deoptimizer* deoptimizer = new Deoptimizer(function, type, bailout_id, from, fp_to_sp_delta); ASSERT(current_ == NULL); current_ = deoptimizer; return deoptimizer; } Deoptimizer* Deoptimizer::Grab() { Deoptimizer* result = current_; ASSERT(result != NULL); result->DeleteFrameDescriptions(); current_ = NULL; return result; } void Deoptimizer::GenerateDeoptimizationEntries(MacroAssembler* masm, int count, BailoutType type) { TableEntryGenerator generator(masm, type, count); generator.Generate(); } class DeoptimizingVisitor : public OptimizedFunctionVisitor { public: virtual void EnterContext(Context* context) { if (FLAG_trace_deopt) { PrintF("[deoptimize context: %" V8PRIxPTR "]\n", reinterpret_cast(context)); } } virtual void VisitFunction(JSFunction* function) { Deoptimizer::DeoptimizeFunction(function); } virtual void LeaveContext(Context* context) { context->ClearOptimizedFunctions(); } }; void Deoptimizer::DeoptimizeAll() { AssertNoAllocation no_allocation; if (FLAG_trace_deopt) { PrintF("[deoptimize all contexts]\n"); } DeoptimizingVisitor visitor; VisitAllOptimizedFunctions(&visitor); } void Deoptimizer::DeoptimizeGlobalObject(JSObject* object) { AssertNoAllocation no_allocation; DeoptimizingVisitor visitor; VisitAllOptimizedFunctionsForGlobalObject(object, &visitor); } void Deoptimizer::VisitAllOptimizedFunctionsForContext( Context* context, OptimizedFunctionVisitor* visitor) { AssertNoAllocation no_allocation; ASSERT(context->IsGlobalContext()); visitor->EnterContext(context); // Run through the list of optimized functions and deoptimize them. Object* element = context->OptimizedFunctionsListHead(); while (!element->IsUndefined()) { JSFunction* element_function = JSFunction::cast(element); // Get the next link before deoptimizing as deoptimizing will clear the // next link. element = element_function->next_function_link(); visitor->VisitFunction(element_function); } visitor->LeaveContext(context); } void Deoptimizer::VisitAllOptimizedFunctionsForGlobalObject( JSObject* object, OptimizedFunctionVisitor* visitor) { AssertNoAllocation no_allocation; if (object->IsJSGlobalProxy()) { Object* proto = object->GetPrototype(); ASSERT(proto->IsJSGlobalObject()); VisitAllOptimizedFunctionsForContext( GlobalObject::cast(proto)->global_context(), visitor); } else if (object->IsGlobalObject()) { VisitAllOptimizedFunctionsForContext( GlobalObject::cast(object)->global_context(), visitor); } } void Deoptimizer::VisitAllOptimizedFunctions( OptimizedFunctionVisitor* visitor) { AssertNoAllocation no_allocation; // Run through the list of all global contexts and deoptimize. Object* global = Heap::global_contexts_list(); while (!global->IsUndefined()) { VisitAllOptimizedFunctionsForGlobalObject(Context::cast(global)->global(), visitor); global = Context::cast(global)->get(Context::NEXT_CONTEXT_LINK); } } void Deoptimizer::HandleWeakDeoptimizedCode( v8::Persistent obj, void* data) { DeoptimizingCodeListNode* node = reinterpret_cast(data); RemoveDeoptimizingCode(*node->code()); #ifdef DEBUG node = Deoptimizer::deoptimizing_code_list_; while (node != NULL) { ASSERT(node != reinterpret_cast(data)); node = node->next(); } #endif } void Deoptimizer::ComputeOutputFrames(Deoptimizer* deoptimizer) { deoptimizer->DoComputeOutputFrames(); } Deoptimizer::Deoptimizer(JSFunction* function, BailoutType type, unsigned bailout_id, Address from, int fp_to_sp_delta) : function_(function), bailout_id_(bailout_id), bailout_type_(type), from_(from), fp_to_sp_delta_(fp_to_sp_delta), output_count_(0), output_(NULL), integer32_values_(NULL), double_values_(NULL) { if (FLAG_trace_deopt && type != OSR) { PrintF("**** DEOPT: "); function->PrintName(); PrintF(" at bailout #%u, address 0x%" V8PRIxPTR ", frame size %d\n", bailout_id, reinterpret_cast(from), fp_to_sp_delta - (2 * kPointerSize)); } else if (FLAG_trace_osr && type == OSR) { PrintF("**** OSR: "); function->PrintName(); PrintF(" at ast id #%u, address 0x%" V8PRIxPTR ", frame size %d\n", bailout_id, reinterpret_cast(from), fp_to_sp_delta - (2 * kPointerSize)); } // Find the optimized code. if (type == EAGER) { ASSERT(from == NULL); optimized_code_ = function_->code(); } else if (type == LAZY) { optimized_code_ = FindDeoptimizingCodeFromAddress(from); ASSERT(optimized_code_ != NULL); } else if (type == OSR) { // The function has already been optimized and we're transitioning // from the unoptimized shared version to the optimized one in the // function. The return address (from) points to unoptimized code. optimized_code_ = function_->code(); ASSERT(optimized_code_->kind() == Code::OPTIMIZED_FUNCTION); ASSERT(!optimized_code_->contains(from)); } ASSERT(Heap::allow_allocation(false)); unsigned size = ComputeInputFrameSize(); input_ = new(size) FrameDescription(size, function); } Deoptimizer::~Deoptimizer() { ASSERT(input_ == NULL && output_ == NULL); delete[] integer32_values_; delete[] double_values_; } void Deoptimizer::DeleteFrameDescriptions() { delete input_; for (int i = 0; i < output_count_; ++i) { if (output_[i] != input_) delete output_[i]; } delete[] output_; input_ = NULL; output_ = NULL; ASSERT(!Heap::allow_allocation(true)); } Address Deoptimizer::GetDeoptimizationEntry(int id, BailoutType type) { ASSERT(id >= 0); if (id >= kNumberOfEntries) return NULL; LargeObjectChunk* base = NULL; if (type == EAGER) { if (eager_deoptimization_entry_code_ == NULL) { eager_deoptimization_entry_code_ = CreateCode(type); } base = eager_deoptimization_entry_code_; } else { if (lazy_deoptimization_entry_code_ == NULL) { lazy_deoptimization_entry_code_ = CreateCode(type); } base = lazy_deoptimization_entry_code_; } return static_cast
(base->GetStartAddress()) + (id * table_entry_size_); } int Deoptimizer::GetDeoptimizationId(Address addr, BailoutType type) { LargeObjectChunk* base = NULL; if (type == EAGER) { base = eager_deoptimization_entry_code_; } else { base = lazy_deoptimization_entry_code_; } if (base == NULL || addr < base->GetStartAddress() || addr >= base->GetStartAddress() + (kNumberOfEntries * table_entry_size_)) { return kNotDeoptimizationEntry; } ASSERT_EQ(0, static_cast(addr - base->GetStartAddress()) % table_entry_size_); return static_cast(addr - base->GetStartAddress()) / table_entry_size_; } void Deoptimizer::Setup() { // Do nothing yet. } void Deoptimizer::TearDown() { if (eager_deoptimization_entry_code_ != NULL) { eager_deoptimization_entry_code_->Free(EXECUTABLE); eager_deoptimization_entry_code_ = NULL; } if (lazy_deoptimization_entry_code_ != NULL) { lazy_deoptimization_entry_code_->Free(EXECUTABLE); lazy_deoptimization_entry_code_ = NULL; } } unsigned Deoptimizer::GetOutputInfo(DeoptimizationOutputData* data, unsigned id, SharedFunctionInfo* shared) { // TODO(kasperl): For now, we do a simple linear search for the PC // offset associated with the given node id. This should probably be // changed to a binary search. int length = data->DeoptPoints(); Smi* smi_id = Smi::FromInt(id); for (int i = 0; i < length; i++) { if (data->AstId(i) == smi_id) { return data->PcAndState(i)->value(); } } PrintF("[couldn't find pc offset for node=%u]\n", id); PrintF("[method: %s]\n", *shared->DebugName()->ToCString()); // Print the source code if available. HeapStringAllocator string_allocator; StringStream stream(&string_allocator); shared->SourceCodePrint(&stream, -1); PrintF("[source:\n%s\n]", *stream.ToCString()); UNREACHABLE(); return -1; } int Deoptimizer::GetDeoptimizedCodeCount() { int length = 0; DeoptimizingCodeListNode* node = Deoptimizer::deoptimizing_code_list_; while (node != NULL) { length++; node = node->next(); } return length; } void Deoptimizer::DoComputeOutputFrames() { if (bailout_type_ == OSR) { DoComputeOsrOutputFrame(); return; } // Print some helpful diagnostic information. int64_t start = OS::Ticks(); if (FLAG_trace_deopt) { PrintF("[deoptimizing%s: begin 0x%08" V8PRIxPTR " ", (bailout_type_ == LAZY ? " (lazy)" : ""), reinterpret_cast(function_)); function_->PrintName(); PrintF(" @%d]\n", bailout_id_); } // Determine basic deoptimization information. The optimized frame is // described by the input data. DeoptimizationInputData* input_data = DeoptimizationInputData::cast(optimized_code_->deoptimization_data()); unsigned node_id = input_data->AstId(bailout_id_)->value(); ByteArray* translations = input_data->TranslationByteArray(); unsigned translation_index = input_data->TranslationIndex(bailout_id_)->value(); // Do the input frame to output frame(s) translation. TranslationIterator iterator(translations, translation_index); Translation::Opcode opcode = static_cast(iterator.Next()); ASSERT(Translation::BEGIN == opcode); USE(opcode); // Read the number of output frames and allocate an array for their // descriptions. int count = iterator.Next(); ASSERT(output_ == NULL); output_ = new FrameDescription*[count]; // Per-frame lists of untagged and unboxed int32 and double values. integer32_values_ = new List[count]; double_values_ = new List[count]; for (int i = 0; i < count; ++i) { output_[i] = NULL; integer32_values_[i].Initialize(0); double_values_[i].Initialize(0); } output_count_ = count; // Translate each output frame. for (int i = 0; i < count; ++i) { DoComputeFrame(&iterator, i); } // Print some helpful diagnostic information. if (FLAG_trace_deopt) { double ms = static_cast(OS::Ticks() - start) / 1000; int index = output_count_ - 1; // Index of the topmost frame. JSFunction* function = output_[index]->GetFunction(); PrintF("[deoptimizing: end 0x%08" V8PRIxPTR " ", reinterpret_cast(function)); function->PrintName(); PrintF(" => node=%u, pc=0x%08" V8PRIxPTR ", state=%s, took %0.3f ms]\n", node_id, output_[index]->GetPc(), FullCodeGenerator::State2String( static_cast( output_[index]->GetState()->value())), ms); } } void Deoptimizer::InsertHeapNumberValues(int index, JavaScriptFrame* frame) { // We need to adjust the stack index by one for the top-most frame. int extra_slot_count = (index == output_count() - 1) ? 1 : 0; List* ints = &integer32_values_[index]; for (int i = 0; i < ints->length(); i++) { ValueDescriptionInteger32 value = ints->at(i); double val = static_cast(value.int32_value()); InsertHeapNumberValue(frame, value.stack_index(), val, extra_slot_count); } // Iterate over double values and convert them to a heap number. List* doubles = &double_values_[index]; for (int i = 0; i < doubles->length(); ++i) { ValueDescriptionDouble value = doubles->at(i); InsertHeapNumberValue(frame, value.stack_index(), value.double_value(), extra_slot_count); } } void Deoptimizer::InsertHeapNumberValue(JavaScriptFrame* frame, int stack_index, double val, int extra_slot_count) { // Add one to the TOS index to take the 'state' pushed before jumping // to the stub that calls Runtime::NotifyDeoptimized into account. int tos_index = stack_index + extra_slot_count; int index = (frame->ComputeExpressionsCount() - 1) - tos_index; if (FLAG_trace_deopt) PrintF("Allocating a new heap number: %e\n", val); Handle num = Factory::NewNumber(val); frame->SetExpression(index, *num); } void Deoptimizer::DoTranslateCommand(TranslationIterator* iterator, int frame_index, unsigned output_offset) { disasm::NameConverter converter; // A GC-safe temporary placeholder that we can put in the output frame. const intptr_t kPlaceholder = reinterpret_cast(Smi::FromInt(0)); // Ignore commands marked as duplicate and act on the first non-duplicate. Translation::Opcode opcode = static_cast(iterator->Next()); while (opcode == Translation::DUPLICATE) { opcode = static_cast(iterator->Next()); iterator->Skip(Translation::NumberOfOperandsFor(opcode)); opcode = static_cast(iterator->Next()); } switch (opcode) { case Translation::BEGIN: case Translation::FRAME: case Translation::DUPLICATE: UNREACHABLE(); return; case Translation::REGISTER: { int input_reg = iterator->Next(); intptr_t input_value = input_->GetRegister(input_reg); if (FLAG_trace_deopt) { PrintF( " 0x%08" V8PRIxPTR ": [top + %d] <- 0x%08" V8PRIxPTR " ; %s\n", output_[frame_index]->GetTop() + output_offset, output_offset, input_value, converter.NameOfCPURegister(input_reg)); } output_[frame_index]->SetFrameSlot(output_offset, input_value); return; } case Translation::INT32_REGISTER: { int input_reg = iterator->Next(); intptr_t value = input_->GetRegister(input_reg); bool is_smi = Smi::IsValid(value); unsigned output_index = output_offset / kPointerSize; if (FLAG_trace_deopt) { PrintF( " 0x%08" V8PRIxPTR ": [top + %d] <- %" V8PRIdPTR " ; %s (%s)\n", output_[frame_index]->GetTop() + output_offset, output_offset, value, converter.NameOfCPURegister(input_reg), is_smi ? "smi" : "heap number"); } if (is_smi) { intptr_t tagged_value = reinterpret_cast(Smi::FromInt(static_cast(value))); output_[frame_index]->SetFrameSlot(output_offset, tagged_value); } else { // We save the untagged value on the side and store a GC-safe // temporary placeholder in the frame. AddInteger32Value(frame_index, output_index, static_cast(value)); output_[frame_index]->SetFrameSlot(output_offset, kPlaceholder); } return; } case Translation::DOUBLE_REGISTER: { int input_reg = iterator->Next(); double value = input_->GetDoubleRegister(input_reg); unsigned output_index = output_offset / kPointerSize; if (FLAG_trace_deopt) { PrintF(" 0x%08" V8PRIxPTR ": [top + %d] <- %e ; %s\n", output_[frame_index]->GetTop() + output_offset, output_offset, value, DoubleRegister::AllocationIndexToString(input_reg)); } // We save the untagged value on the side and store a GC-safe // temporary placeholder in the frame. AddDoubleValue(frame_index, output_index, value); output_[frame_index]->SetFrameSlot(output_offset, kPlaceholder); return; } case Translation::STACK_SLOT: { int input_slot_index = iterator->Next(); unsigned input_offset = input_->GetOffsetFromSlotIndex(this, input_slot_index); intptr_t input_value = input_->GetFrameSlot(input_offset); if (FLAG_trace_deopt) { PrintF(" 0x%08" V8PRIxPTR ": ", output_[frame_index]->GetTop() + output_offset); PrintF("[top + %d] <- 0x%08" V8PRIxPTR " ; [esp + %d]\n", output_offset, input_value, input_offset); } output_[frame_index]->SetFrameSlot(output_offset, input_value); return; } case Translation::INT32_STACK_SLOT: { int input_slot_index = iterator->Next(); unsigned input_offset = input_->GetOffsetFromSlotIndex(this, input_slot_index); intptr_t value = input_->GetFrameSlot(input_offset); bool is_smi = Smi::IsValid(value); unsigned output_index = output_offset / kPointerSize; if (FLAG_trace_deopt) { PrintF(" 0x%08" V8PRIxPTR ": ", output_[frame_index]->GetTop() + output_offset); PrintF("[top + %d] <- %" V8PRIdPTR " ; [esp + %d] (%s)\n", output_offset, value, input_offset, is_smi ? "smi" : "heap number"); } if (is_smi) { intptr_t tagged_value = reinterpret_cast(Smi::FromInt(static_cast(value))); output_[frame_index]->SetFrameSlot(output_offset, tagged_value); } else { // We save the untagged value on the side and store a GC-safe // temporary placeholder in the frame. AddInteger32Value(frame_index, output_index, static_cast(value)); output_[frame_index]->SetFrameSlot(output_offset, kPlaceholder); } return; } case Translation::DOUBLE_STACK_SLOT: { int input_slot_index = iterator->Next(); unsigned input_offset = input_->GetOffsetFromSlotIndex(this, input_slot_index); double value = input_->GetDoubleFrameSlot(input_offset); unsigned output_index = output_offset / kPointerSize; if (FLAG_trace_deopt) { PrintF(" 0x%08" V8PRIxPTR ": [top + %d] <- %e ; [esp + %d]\n", output_[frame_index]->GetTop() + output_offset, output_offset, value, input_offset); } // We save the untagged value on the side and store a GC-safe // temporary placeholder in the frame. AddDoubleValue(frame_index, output_index, value); output_[frame_index]->SetFrameSlot(output_offset, kPlaceholder); return; } case Translation::LITERAL: { Object* literal = ComputeLiteral(iterator->Next()); if (FLAG_trace_deopt) { PrintF(" 0x%08" V8PRIxPTR ": [top + %d] <- ", output_[frame_index]->GetTop() + output_offset, output_offset); literal->ShortPrint(); PrintF(" ; literal\n"); } intptr_t value = reinterpret_cast(literal); output_[frame_index]->SetFrameSlot(output_offset, value); return; } case Translation::ARGUMENTS_OBJECT: { // Use the hole value as a sentinel and fill in the arguments object // after the deoptimized frame is built. ASSERT(frame_index == 0); // Only supported for first frame. if (FLAG_trace_deopt) { PrintF(" 0x%08" V8PRIxPTR ": [top + %d] <- ", output_[frame_index]->GetTop() + output_offset, output_offset); Heap::the_hole_value()->ShortPrint(); PrintF(" ; arguments object\n"); } intptr_t value = reinterpret_cast(Heap::the_hole_value()); output_[frame_index]->SetFrameSlot(output_offset, value); return; } } } bool Deoptimizer::DoOsrTranslateCommand(TranslationIterator* iterator, int* input_offset) { disasm::NameConverter converter; FrameDescription* output = output_[0]; // The input values are all part of the unoptimized frame so they // are all tagged pointers. uintptr_t input_value = input_->GetFrameSlot(*input_offset); Object* input_object = reinterpret_cast(input_value); Translation::Opcode opcode = static_cast(iterator->Next()); bool duplicate = (opcode == Translation::DUPLICATE); if (duplicate) { opcode = static_cast(iterator->Next()); } switch (opcode) { case Translation::BEGIN: case Translation::FRAME: case Translation::DUPLICATE: UNREACHABLE(); // Malformed input. return false; case Translation::REGISTER: { int output_reg = iterator->Next(); if (FLAG_trace_osr) { PrintF(" %s <- 0x%08" V8PRIxPTR " ; [esp + %d]\n", converter.NameOfCPURegister(output_reg), input_value, *input_offset); } output->SetRegister(output_reg, input_value); break; } case Translation::INT32_REGISTER: { // Abort OSR if we don't have a number. if (!input_object->IsNumber()) return false; int output_reg = iterator->Next(); int int32_value = input_object->IsSmi() ? Smi::cast(input_object)->value() : FastD2I(input_object->Number()); // Abort the translation if the conversion lost information. if (!input_object->IsSmi() && FastI2D(int32_value) != input_object->Number()) { if (FLAG_trace_osr) { PrintF("**** %g could not be converted to int32 ****\n", input_object->Number()); } return false; } if (FLAG_trace_osr) { PrintF(" %s <- %d (int32) ; [esp + %d]\n", converter.NameOfCPURegister(output_reg), int32_value, *input_offset); } output->SetRegister(output_reg, int32_value); break; } case Translation::DOUBLE_REGISTER: { // Abort OSR if we don't have a number. if (!input_object->IsNumber()) return false; int output_reg = iterator->Next(); double double_value = input_object->Number(); if (FLAG_trace_osr) { PrintF(" %s <- %g (double) ; [esp + %d]\n", DoubleRegister::AllocationIndexToString(output_reg), double_value, *input_offset); } output->SetDoubleRegister(output_reg, double_value); break; } case Translation::STACK_SLOT: { int output_index = iterator->Next(); unsigned output_offset = output->GetOffsetFromSlotIndex(this, output_index); if (FLAG_trace_osr) { PrintF(" [esp + %d] <- 0x%08" V8PRIxPTR " ; [esp + %d]\n", output_offset, input_value, *input_offset); } output->SetFrameSlot(output_offset, input_value); break; } case Translation::INT32_STACK_SLOT: { // Abort OSR if we don't have a number. if (!input_object->IsNumber()) return false; int output_index = iterator->Next(); unsigned output_offset = output->GetOffsetFromSlotIndex(this, output_index); int int32_value = input_object->IsSmi() ? Smi::cast(input_object)->value() : DoubleToInt32(input_object->Number()); // Abort the translation if the conversion lost information. if (!input_object->IsSmi() && FastI2D(int32_value) != input_object->Number()) { if (FLAG_trace_osr) { PrintF("**** %g could not be converted to int32 ****\n", input_object->Number()); } return false; } if (FLAG_trace_osr) { PrintF(" [esp + %d] <- %d (int32) ; [esp + %d]\n", output_offset, int32_value, *input_offset); } output->SetFrameSlot(output_offset, int32_value); break; } case Translation::DOUBLE_STACK_SLOT: { static const int kLowerOffset = 0 * kPointerSize; static const int kUpperOffset = 1 * kPointerSize; // Abort OSR if we don't have a number. if (!input_object->IsNumber()) return false; int output_index = iterator->Next(); unsigned output_offset = output->GetOffsetFromSlotIndex(this, output_index); double double_value = input_object->Number(); uint64_t int_value = BitCast(double_value); int32_t lower = static_cast(int_value); int32_t upper = static_cast(int_value >> kBitsPerInt); if (FLAG_trace_osr) { PrintF(" [esp + %d] <- 0x%08x (upper bits of %g) ; [esp + %d]\n", output_offset + kUpperOffset, upper, double_value, *input_offset); PrintF(" [esp + %d] <- 0x%08x (lower bits of %g) ; [esp + %d]\n", output_offset + kLowerOffset, lower, double_value, *input_offset); } output->SetFrameSlot(output_offset + kLowerOffset, lower); output->SetFrameSlot(output_offset + kUpperOffset, upper); break; } case Translation::LITERAL: { // Just ignore non-materialized literals. iterator->Next(); break; } case Translation::ARGUMENTS_OBJECT: { // Optimized code assumes that the argument object has not been // materialized and so bypasses it when doing arguments access. // We should have bailed out before starting the frame // translation. UNREACHABLE(); return false; } } if (!duplicate) *input_offset -= kPointerSize; return true; } unsigned Deoptimizer::ComputeInputFrameSize() const { unsigned fixed_size = ComputeFixedSize(function_); // The fp-to-sp delta already takes the context and the function // into account so we have to avoid double counting them (-2). unsigned result = fixed_size + fp_to_sp_delta_ - (2 * kPointerSize); #ifdef DEBUG if (bailout_type_ == OSR) { // TODO(kasperl): It would be nice if we could verify that the // size matches with the stack height we can compute based on the // environment at the OSR entry. The code for that his built into // the DoComputeOsrOutputFrame function for now. } else { unsigned stack_slots = optimized_code_->stack_slots(); unsigned outgoing_size = ComputeOutgoingArgumentSize(); ASSERT(result == fixed_size + (stack_slots * kPointerSize) + outgoing_size); } #endif return result; } unsigned Deoptimizer::ComputeFixedSize(JSFunction* function) const { // The fixed part of the frame consists of the return address, frame // pointer, function, context, and all the incoming arguments. static const unsigned kFixedSlotSize = 4 * kPointerSize; return ComputeIncomingArgumentSize(function) + kFixedSlotSize; } unsigned Deoptimizer::ComputeIncomingArgumentSize(JSFunction* function) const { // The incoming arguments is the values for formal parameters and // the receiver. Every slot contains a pointer. unsigned arguments = function->shared()->formal_parameter_count() + 1; return arguments * kPointerSize; } unsigned Deoptimizer::ComputeOutgoingArgumentSize() const { DeoptimizationInputData* data = DeoptimizationInputData::cast( optimized_code_->deoptimization_data()); unsigned height = data->ArgumentsStackHeight(bailout_id_)->value(); return height * kPointerSize; } Object* Deoptimizer::ComputeLiteral(int index) const { DeoptimizationInputData* data = DeoptimizationInputData::cast( optimized_code_->deoptimization_data()); FixedArray* literals = data->LiteralArray(); return literals->get(index); } void Deoptimizer::AddInteger32Value(int frame_index, int slot_index, int32_t value) { ValueDescriptionInteger32 value_desc(slot_index, value); integer32_values_[frame_index].Add(value_desc); } void Deoptimizer::AddDoubleValue(int frame_index, int slot_index, double value) { ValueDescriptionDouble value_desc(slot_index, value); double_values_[frame_index].Add(value_desc); } LargeObjectChunk* Deoptimizer::CreateCode(BailoutType type) { // We cannot run this if the serializer is enabled because this will // cause us to emit relocation information for the external // references. This is fine because the deoptimizer's code section // isn't meant to be serialized at all. ASSERT(!Serializer::enabled()); bool old_debug_code = FLAG_debug_code; FLAG_debug_code = false; MacroAssembler masm(NULL, 16 * KB); GenerateDeoptimizationEntries(&masm, kNumberOfEntries, type); CodeDesc desc; masm.GetCode(&desc); ASSERT(desc.reloc_size == 0); LargeObjectChunk* chunk = LargeObjectChunk::New(desc.instr_size, EXECUTABLE); memcpy(chunk->GetStartAddress(), desc.buffer, desc.instr_size); CPU::FlushICache(chunk->GetStartAddress(), desc.instr_size); FLAG_debug_code = old_debug_code; return chunk; } Code* Deoptimizer::FindDeoptimizingCodeFromAddress(Address addr) { DeoptimizingCodeListNode* node = Deoptimizer::deoptimizing_code_list_; while (node != NULL) { if (node->code()->contains(addr)) return *node->code(); node = node->next(); } return NULL; } void Deoptimizer::RemoveDeoptimizingCode(Code* code) { ASSERT(deoptimizing_code_list_ != NULL); // Run through the code objects to find this one and remove it. DeoptimizingCodeListNode* prev = NULL; DeoptimizingCodeListNode* current = deoptimizing_code_list_; while (current != NULL) { if (*current->code() == code) { // Unlink from list. If prev is NULL we are looking at the first element. if (prev == NULL) { deoptimizing_code_list_ = current->next(); } else { prev->set_next(current->next()); } delete current; return; } // Move to next in list. prev = current; current = current->next(); } // Deoptimizing code is removed through weak callback. Each object is expected // to be removed once and only once. UNREACHABLE(); } FrameDescription::FrameDescription(uint32_t frame_size, JSFunction* function) : frame_size_(frame_size), function_(function), top_(kZapUint32), pc_(kZapUint32), fp_(kZapUint32) { // Zap all the registers. for (int r = 0; r < Register::kNumRegisters; r++) { SetRegister(r, kZapUint32); } // Zap all the slots. for (unsigned o = 0; o < frame_size; o += kPointerSize) { SetFrameSlot(o, kZapUint32); } } unsigned FrameDescription::GetOffsetFromSlotIndex(Deoptimizer* deoptimizer, int slot_index) { if (slot_index >= 0) { // Local or spill slots. Skip the fixed part of the frame // including all arguments. unsigned base = static_cast( GetFrameSize() - deoptimizer->ComputeFixedSize(GetFunction())); return base - ((slot_index + 1) * kPointerSize); } else { // Incoming parameter. unsigned base = static_cast(GetFrameSize() - deoptimizer->ComputeIncomingArgumentSize(GetFunction())); return base - ((slot_index + 1) * kPointerSize); } } void TranslationBuffer::Add(int32_t value) { // Encode the sign bit in the least significant bit. bool is_negative = (value < 0); uint32_t bits = ((is_negative ? -value : value) << 1) | static_cast(is_negative); // Encode the individual bytes using the least significant bit of // each byte to indicate whether or not more bytes follow. do { uint32_t next = bits >> 7; contents_.Add(((bits << 1) & 0xFF) | (next != 0)); bits = next; } while (bits != 0); } int32_t TranslationIterator::Next() { ASSERT(HasNext()); // Run through the bytes until we reach one with a least significant // bit of zero (marks the end). uint32_t bits = 0; for (int i = 0; true; i += 7) { uint8_t next = buffer_->get(index_++); bits |= (next >> 1) << i; if ((next & 1) == 0) break; } // The bits encode the sign in the least significant bit. bool is_negative = (bits & 1) == 1; int32_t result = bits >> 1; return is_negative ? -result : result; } Handle TranslationBuffer::CreateByteArray() { int length = contents_.length(); Handle result = Factory::NewByteArray(length, TENURED); memcpy(result->GetDataStartAddress(), contents_.ToVector().start(), length); return result; } void Translation::BeginFrame(int node_id, int literal_id, unsigned height) { buffer_->Add(FRAME); buffer_->Add(node_id); buffer_->Add(literal_id); buffer_->Add(height); } void Translation::StoreRegister(Register reg) { buffer_->Add(REGISTER); buffer_->Add(reg.code()); } void Translation::StoreInt32Register(Register reg) { buffer_->Add(INT32_REGISTER); buffer_->Add(reg.code()); } void Translation::StoreDoubleRegister(DoubleRegister reg) { buffer_->Add(DOUBLE_REGISTER); buffer_->Add(DoubleRegister::ToAllocationIndex(reg)); } void Translation::StoreStackSlot(int index) { buffer_->Add(STACK_SLOT); buffer_->Add(index); } void Translation::StoreInt32StackSlot(int index) { buffer_->Add(INT32_STACK_SLOT); buffer_->Add(index); } void Translation::StoreDoubleStackSlot(int index) { buffer_->Add(DOUBLE_STACK_SLOT); buffer_->Add(index); } void Translation::StoreLiteral(int literal_id) { buffer_->Add(LITERAL); buffer_->Add(literal_id); } void Translation::StoreArgumentsObject() { buffer_->Add(ARGUMENTS_OBJECT); } void Translation::MarkDuplicate() { buffer_->Add(DUPLICATE); } int Translation::NumberOfOperandsFor(Opcode opcode) { switch (opcode) { case ARGUMENTS_OBJECT: case DUPLICATE: return 0; case BEGIN: case REGISTER: case INT32_REGISTER: case DOUBLE_REGISTER: case STACK_SLOT: case INT32_STACK_SLOT: case DOUBLE_STACK_SLOT: case LITERAL: return 1; case FRAME: return 3; } UNREACHABLE(); return -1; } #ifdef DEBUG const char* Translation::StringFor(Opcode opcode) { switch (opcode) { case BEGIN: return "BEGIN"; case FRAME: return "FRAME"; case REGISTER: return "REGISTER"; case INT32_REGISTER: return "INT32_REGISTER"; case DOUBLE_REGISTER: return "DOUBLE_REGISTER"; case STACK_SLOT: return "STACK_SLOT"; case INT32_STACK_SLOT: return "INT32_STACK_SLOT"; case DOUBLE_STACK_SLOT: return "DOUBLE_STACK_SLOT"; case LITERAL: return "LITERAL"; case ARGUMENTS_OBJECT: return "ARGUMENTS_OBJECT"; case DUPLICATE: return "DUPLICATE"; } UNREACHABLE(); return ""; } #endif DeoptimizingCodeListNode::DeoptimizingCodeListNode(Code* code): next_(NULL) { // Globalize the code object and make it weak. code_ = Handle::cast((GlobalHandles::Create(code))); GlobalHandles::MakeWeak(reinterpret_cast(code_.location()), this, Deoptimizer::HandleWeakDeoptimizedCode); } DeoptimizingCodeListNode::~DeoptimizingCodeListNode() { GlobalHandles::Destroy(reinterpret_cast(code_.location())); } } } // namespace v8::internal