v8/src/codegen.cc

486 lines
16 KiB
C++
Raw Normal View History

// Copyright 2007-2008 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 "bootstrapper.h"
#include "codegen-inl.h"
#include "debug.h"
#include "prettyprinter.h"
#include "scopeinfo.h"
#include "runtime.h"
#include "stub-cache.h"
namespace v8 { namespace internal {
DeferredCode::DeferredCode(CodeGenerator* generator)
: masm_(generator->masm()),
generator_(generator),
Defer the writing of the source position data to the relocation information until a possible debug break location is reached. Currently this is call sites with calls to code objects and JS return. Source position information in the code therefore no longer refers to the "first" instruction generated for a given source position (which was not the case defered code anyway) but to the first break location after that source position was passed (again defered code always start with source position information). This doesn't make a difference for the debugger as it will always be stopped only at debug break locations. However, this makes the life of the peep-hole optimizer much easier as many oportunities for posh/pop eliminations where previosly blocked by relocation information already written to the code object. Two types of source positions are still collected. Statement positions indicate the position of the start of the statement leading to this code and (plain) positions indicate other places typically call sites to help indicate current position in backtraces. The two different types of positions are also used to distinguish between step next and step in. Runs all the tests (including debugger tests) as before. Moved the checking for the FLAG_debug_info to one place. I will do the same changes to the ARM codegenerator in a seperate changelist. Review URL: http://codereview.chromium.org/2957 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@335 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2008-09-18 08:51:43 +00:00
statement_position_(masm_->last_statement_position()),
position_(masm_->last_position()) {
generator->AddDeferred(this);
#ifdef DEBUG
comment_ = "";
#endif
}
void CodeGenerator::ProcessDeferred() {
while (!deferred_.is_empty()) {
DeferredCode* code = deferred_.RemoveLast();
MacroAssembler* masm = code->masm();
// Record position of deferred code stub.
if (code->statement_position() != RelocInfo::kNoPosition) {
Defer the writing of the source position data to the relocation information until a possible debug break location is reached. Currently this is call sites with calls to code objects and JS return. Source position information in the code therefore no longer refers to the "first" instruction generated for a given source position (which was not the case defered code anyway) but to the first break location after that source position was passed (again defered code always start with source position information). This doesn't make a difference for the debugger as it will always be stopped only at debug break locations. However, this makes the life of the peep-hole optimizer much easier as many oportunities for posh/pop eliminations where previosly blocked by relocation information already written to the code object. Two types of source positions are still collected. Statement positions indicate the position of the start of the statement leading to this code and (plain) positions indicate other places typically call sites to help indicate current position in backtraces. The two different types of positions are also used to distinguish between step next and step in. Runs all the tests (including debugger tests) as before. Moved the checking for the FLAG_debug_info to one place. I will do the same changes to the ARM codegenerator in a seperate changelist. Review URL: http://codereview.chromium.org/2957 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@335 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2008-09-18 08:51:43 +00:00
masm->RecordStatementPosition(code->statement_position());
}
if (code->position() != RelocInfo::kNoPosition) {
masm->RecordPosition(code->position());
}
// Bind labels and generate the code.
masm->bind(code->enter());
Comment cmnt(masm, code->comment());
code->Generate();
if (code->exit()->is_bound()) {
masm->jmp(code->exit()); // platform independent?
}
}
}
// Generate the code. Takes a function literal, generates code for it, assemble
// all the pieces into a Code object. This function is only to be called by
// the compiler.cc code.
Handle<Code> CodeGenerator::MakeCode(FunctionLiteral* flit,
Handle<Script> script,
bool is_eval) {
#ifdef ENABLE_DISASSEMBLER
bool print_code = FLAG_print_code && !Bootstrapper::IsActive();
#endif
#ifdef DEBUG
bool print_source = false;
bool print_ast = false;
const char* ftype;
if (Bootstrapper::IsActive()) {
print_source = FLAG_print_builtin_source;
print_ast = FLAG_print_builtin_ast;
print_code = FLAG_print_builtin_code;
ftype = "builtin";
} else {
print_source = FLAG_print_source;
print_ast = FLAG_print_ast;
ftype = "user-defined";
}
if (FLAG_trace_codegen || print_source || print_ast) {
PrintF("*** Generate code for %s function: ", ftype);
flit->name()->ShortPrint();
PrintF(" ***\n");
}
if (print_source) {
PrintF("--- Source from AST ---\n%s\n", PrettyPrinter().PrintProgram(flit));
}
if (print_ast) {
PrintF("--- AST ---\n%s\n", AstPrinter().PrintProgram(flit));
}
#endif // DEBUG
// Generate code.
const int initial_buffer_size = 4 * KB;
CodeGenerator cgen(initial_buffer_size, script, is_eval);
cgen.GenCode(flit);
if (cgen.HasStackOverflow()) {
ASSERT(!Top::has_pending_exception());
return Handle<Code>::null();
}
// Process any deferred code.
cgen.ProcessDeferred();
// Allocate and install the code.
CodeDesc desc;
cgen.masm()->GetCode(&desc);
ScopeInfo<> sinfo(flit->scope());
Code::Flags flags = Code::ComputeFlags(Code::FUNCTION);
Handle<Code> code = Factory::NewCode(desc, &sinfo, flags);
// Add unresolved entries in the code to the fixup list.
Bootstrapper::AddFixup(*code, cgen.masm());
#ifdef ENABLE_DISASSEMBLER
if (print_code) {
// Print the source code if available.
if (!script->IsUndefined() && !script->source()->IsUndefined()) {
PrintF("--- Raw source ---\n");
StringInputBuffer stream(String::cast(script->source()));
stream.Seek(flit->start_position());
// flit->end_position() points to the last character in the stream. We
// need to compensate by adding one to calculate the length.
int source_len = flit->end_position() - flit->start_position() + 1;
for (int i = 0; i < source_len; i++) {
if (stream.has_more()) PrintF("%c", stream.GetNext());
}
PrintF("\n\n");
}
PrintF("--- Code ---\n");
code->Disassemble();
}
#endif // ENABLE_DISASSEMBLER
if (!code.is_null()) {
Counters::total_compiled_code_size.Increment(code->instruction_size());
}
return code;
}
// Sets the function info on a function.
// The start_position points to the first '(' character after the function name
// in the full script source. When counting characters in the script source the
// the first character is number 0 (not 1).
void CodeGenerator::SetFunctionInfo(Handle<JSFunction> fun,
int length,
int function_token_position,
int start_position,
int end_position,
bool is_expression,
bool is_toplevel,
Handle<Script> script) {
fun->shared()->set_length(length);
fun->shared()->set_formal_parameter_count(length);
fun->shared()->set_script(*script);
fun->shared()->set_function_token_position(function_token_position);
fun->shared()->set_start_position(start_position);
fun->shared()->set_end_position(end_position);
fun->shared()->set_is_expression(is_expression);
fun->shared()->set_is_toplevel(is_toplevel);
}
static Handle<Code> ComputeLazyCompile(int argc) {
CALL_HEAP_FUNCTION(StubCache::ComputeLazyCompile(argc), Code);
}
Handle<JSFunction> CodeGenerator::BuildBoilerplate(FunctionLiteral* node) {
// Determine if the function can be lazily compiled. This is
// necessary to allow some of our builtin JS files to be lazily
// compiled. These builtins cannot be handled lazily by the parser,
// since we have to know if a function uses the special natives
// syntax, which is something the parser records.
bool allow_lazy = node->AllowsLazyCompilation();
// Generate code
Handle<Code> code;
if (FLAG_lazy && allow_lazy) {
code = ComputeLazyCompile(node->num_parameters());
} else {
code = MakeCode(node, script_, false);
// Check for stack-overflow exception.
if (code.is_null()) {
SetStackOverflow();
return Handle<JSFunction>::null();
}
// Function compilation complete.
LOG(CodeCreateEvent("Function", *code, *node->name()));
}
// Create a boilerplate function.
Handle<JSFunction> function =
Factory::NewFunctionBoilerplate(node->name(),
node->materialized_literal_count(),
node->contains_array_literal(),
code);
CodeGenerator::SetFunctionInfo(function, node->num_parameters(),
node->function_token_position(),
node->start_position(), node->end_position(),
node->is_expression(), false, script_);
// Notify debugger that a new function has been added.
Debugger::OnNewFunction(function);
// Set the expected number of properties for instances and return
// the resulting function.
SetExpectedNofPropertiesFromEstimate(function,
node->expected_property_count());
return function;
}
Handle<Code> CodeGenerator::ComputeCallInitialize(int argc) {
CALL_HEAP_FUNCTION(StubCache::ComputeCallInitialize(argc), Code);
}
Handle<Code> CodeGenerator::ComputeCallInitializeInLoop(int argc) {
// Force the creation of the corresponding stub outside loops,
// because it will be used when clearing the ICs later - when we
// don't know if we're inside a loop or not.
ComputeCallInitialize(argc);
CALL_HEAP_FUNCTION(StubCache::ComputeCallInitializeInLoop(argc), Code);
}
void CodeGenerator::ProcessDeclarations(ZoneList<Declaration*>* declarations) {
int length = declarations->length();
int globals = 0;
for (int i = 0; i < length; i++) {
Declaration* node = declarations->at(i);
Variable* var = node->proxy()->var();
Slot* slot = var->slot();
// If it was not possible to allocate the variable at compile
// time, we need to "declare" it at runtime to make sure it
// actually exists in the local context.
if ((slot != NULL && slot->type() == Slot::LOOKUP) || !var->is_global()) {
VisitDeclaration(node);
} else {
// Count global variables and functions for later processing
globals++;
}
}
// Return in case of no declared global functions or variables.
if (globals == 0) return;
// Compute array of global variable and function declarations.
Handle<FixedArray> array = Factory::NewFixedArray(2 * globals, TENURED);
for (int j = 0, i = 0; i < length; i++) {
Declaration* node = declarations->at(i);
Variable* var = node->proxy()->var();
Slot* slot = var->slot();
if ((slot != NULL && slot->type() == Slot::LOOKUP) || !var->is_global()) {
// Skip - already processed.
} else {
array->set(j++, *(var->name()));
if (node->fun() == NULL) {
if (var->mode() == Variable::CONST) {
// In case this is const property use the hole.
array->set_the_hole(j++);
} else {
array->set_undefined(j++);
}
} else {
Handle<JSFunction> function = BuildBoilerplate(node->fun());
// Check for stack-overflow exception.
if (HasStackOverflow()) return;
array->set(j++, *function);
}
}
}
// Invoke the platform-dependent code generator to do the actual
// declaration the global variables and functions.
DeclareGlobals(array);
}
struct InlineRuntimeLUT {
void (CodeGenerator::*method)(ZoneList<Expression*>*);
const char* name;
};
bool CodeGenerator::CheckForInlineRuntimeCall(CallRuntime* node) {
ZoneList<Expression*>* args = node->arguments();
// Special cases: These 'runtime calls' manipulate the current
// frame and are only used 1 or two places, so we generate them
// inline instead of generating calls to them. They are used
// for implementing Function.prototype.call() and
// Function.prototype.apply().
static const InlineRuntimeLUT kInlineRuntimeLUT[] = {
{&v8::internal::CodeGenerator::GenerateIsSmi,
"_IsSmi"},
{&v8::internal::CodeGenerator::GenerateIsNonNegativeSmi,
"_IsNonNegativeSmi"},
{&v8::internal::CodeGenerator::GenerateIsArray,
"_IsArray"},
{&v8::internal::CodeGenerator::GenerateArgumentsLength,
"_ArgumentsLength"},
{&v8::internal::CodeGenerator::GenerateArgumentsAccess,
"_Arguments"},
{&v8::internal::CodeGenerator::GenerateValueOf,
"_ValueOf"},
{&v8::internal::CodeGenerator::GenerateSetValueOf,
"_SetValueOf"},
{&v8::internal::CodeGenerator::GenerateFastCharCodeAt,
"_FastCharCodeAt"},
{&v8::internal::CodeGenerator::GenerateObjectEquals,
"_ObjectEquals"}
};
Handle<String> name = node->name();
StringShape shape(*name);
if (name->length(shape) > 0 && name->Get(shape, 0) == '_') {
for (unsigned i = 0;
i < sizeof(kInlineRuntimeLUT) / sizeof(InlineRuntimeLUT);
i++) {
const InlineRuntimeLUT* entry = kInlineRuntimeLUT + i;
if (name->IsEqualTo(CStrVector(entry->name))) {
((*this).*(entry->method))(args);
return true;
}
}
}
return false;
}
void CodeGenerator::GenerateFastCaseSwitchStatement(SwitchStatement* node,
int min_index,
int range,
int default_index) {
ZoneList<CaseClause*>* cases = node->cases();
int length = cases->length();
// Label pointer per number in range
SmartPointer<Label*> case_targets(NewArray<Label*>(range));
// Label per switch case
SmartPointer<Label> case_labels(NewArray<Label>(length));
Label* fail_label = default_index >= 0 ? &(case_labels[default_index])
: node->break_target();
// Populate array of label pointers for each number in the range.
// Initally put the failure label everywhere.
for (int i = 0; i < range; i++) {
case_targets[i] = fail_label;
}
// Overwrite with label of a case for the number value of that case.
// (In reverse order, so that if the same label occurs twice, the
// first one wins).
for (int i = length-1; i >= 0 ; i--) {
CaseClause* clause = cases->at(i);
if (!clause->is_default()) {
Object* label_value = *(clause->label()->AsLiteral()->handle());
int case_value = Smi::cast(label_value)->value();
case_targets[case_value - min_index] = &(case_labels[i]);
}
}
GenerateFastCaseSwitchJumpTable(node,
min_index,
range,
fail_label,
Vector<Label*>(*case_targets, range),
Vector<Label>(*case_labels, length));
}
void CodeGenerator::GenerateFastCaseSwitchCases(
SwitchStatement* node,
Vector<Label> case_labels) {
ZoneList<CaseClause*>* cases = node->cases();
int length = cases->length();
for (int i = 0; i < length; i++) {
Comment cmnt(masm(), "[ Case clause");
masm()->bind(&(case_labels[i]));
VisitStatements(cases->at(i)->statements());
}
masm()->bind(node->break_target());
}
bool CodeGenerator::TryGenerateFastCaseSwitchStatement(SwitchStatement* node) {
ZoneList<CaseClause*>* cases = node->cases();
int length = cases->length();
if (length < FastCaseSwitchMinCaseCount()) {
return false;
}
// Test whether fast-case should be used.
int default_index = -1;
int min_index = Smi::kMaxValue;
int max_index = Smi::kMinValue;
for (int i = 0; i < length; i++) {
CaseClause* clause = cases->at(i);
if (clause->is_default()) {
if (default_index >= 0) {
return false; // More than one default label:
// Defer to normal case for error.
}
default_index = i;
} else {
Expression* label = clause->label();
Literal* literal = label->AsLiteral();
if (literal == NULL) {
return false; // fail fast case
}
Object* value = *(literal->handle());
if (!value->IsSmi()) {
return false;
}
int smi = Smi::cast(value)->value();
if (smi < min_index) { min_index = smi; }
if (smi > max_index) { max_index = smi; }
}
}
// All labels are known to be Smis.
int range = max_index - min_index + 1; // |min..max| inclusive
if (range / FastCaseSwitchMaxOverheadFactor() > length) {
return false; // range of labels is too sparse
}
// Optimization accepted, generate code.
GenerateFastCaseSwitchStatement(node, min_index, range, default_index);
return true;
}
const char* RuntimeStub::GetName() {
return Runtime::FunctionForId(id_)->stub_name;
}
void RuntimeStub::Generate(MacroAssembler* masm) {
masm->TailCallRuntime(ExternalReference(id_), num_arguments_);
}
void ArgumentsAccessStub::Generate(MacroAssembler* masm) {
switch (type_) {
case READ_LENGTH: GenerateReadLength(masm); break;
case READ_ELEMENT: GenerateReadElement(masm); break;
case NEW_OBJECT: GenerateNewObject(masm); break;
}
}
} } // namespace v8::internal