blob: 165c51dd2756e603bf3e99ac58984a3da920f406 [file] [log] [blame]
// Copyright 2009 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"
#if defined(V8_TARGET_ARCH_X64)
#include "bootstrapper.h"
#include "codegen-inl.h"
#include "assembler-x64.h"
#include "macro-assembler-x64.h"
#include "serialize.h"
#include "debug.h"
#include "heap.h"
namespace v8 {
namespace internal {
MacroAssembler::MacroAssembler(void* buffer, int size)
: Assembler(buffer, size),
generating_stub_(false),
allow_stub_calls_(true),
code_object_(Heap::undefined_value()) {
}
void MacroAssembler::LoadRoot(Register destination, Heap::RootListIndex index) {
movq(destination, Operand(kRootRegister, index << kPointerSizeLog2));
}
void MacroAssembler::StoreRoot(Register source, Heap::RootListIndex index) {
movq(Operand(kRootRegister, index << kPointerSizeLog2), source);
}
void MacroAssembler::PushRoot(Heap::RootListIndex index) {
push(Operand(kRootRegister, index << kPointerSizeLog2));
}
void MacroAssembler::CompareRoot(Register with, Heap::RootListIndex index) {
cmpq(with, Operand(kRootRegister, index << kPointerSizeLog2));
}
void MacroAssembler::CompareRoot(Operand with, Heap::RootListIndex index) {
LoadRoot(kScratchRegister, index);
cmpq(with, kScratchRegister);
}
void MacroAssembler::StackLimitCheck(Label* on_stack_overflow) {
CompareRoot(rsp, Heap::kStackLimitRootIndex);
j(below, on_stack_overflow);
}
void MacroAssembler::RecordWriteHelper(Register object,
Register addr,
Register scratch) {
if (FLAG_debug_code) {
// Check that the object is not in new space.
Label not_in_new_space;
InNewSpace(object, scratch, not_equal, &not_in_new_space);
Abort("new-space object passed to RecordWriteHelper");
bind(&not_in_new_space);
}
// Compute the page start address from the heap object pointer, and reuse
// the 'object' register for it.
and_(object, Immediate(~Page::kPageAlignmentMask));
// Compute number of region covering addr. See Page::GetRegionNumberForAddress
// method for more details.
shrl(addr, Immediate(Page::kRegionSizeLog2));
andl(addr, Immediate(Page::kPageAlignmentMask >> Page::kRegionSizeLog2));
// Set dirty mark for region.
bts(Operand(object, Page::kDirtyFlagOffset), addr);
}
void MacroAssembler::RecordWrite(Register object,
int offset,
Register value,
Register index) {
// The compiled code assumes that record write doesn't change the
// context register, so we check that none of the clobbered
// registers are rsi.
ASSERT(!object.is(rsi) && !value.is(rsi) && !index.is(rsi));
// First, check if a write barrier is even needed. The tests below
// catch stores of Smis and stores into young gen.
Label done;
JumpIfSmi(value, &done);
RecordWriteNonSmi(object, offset, value, index);
bind(&done);
// Clobber all input registers when running with the debug-code flag
// turned on to provoke errors. This clobbering repeats the
// clobbering done inside RecordWriteNonSmi but it's necessary to
// avoid having the fast case for smis leave the registers
// unchanged.
if (FLAG_debug_code) {
movq(object, BitCast<int64_t>(kZapValue), RelocInfo::NONE);
movq(value, BitCast<int64_t>(kZapValue), RelocInfo::NONE);
movq(index, BitCast<int64_t>(kZapValue), RelocInfo::NONE);
}
}
void MacroAssembler::RecordWrite(Register object,
Register address,
Register value) {
// The compiled code assumes that record write doesn't change the
// context register, so we check that none of the clobbered
// registers are esi.
ASSERT(!object.is(rsi) && !value.is(rsi) && !address.is(rsi));
// First, check if a write barrier is even needed. The tests below
// catch stores of Smis and stores into young gen.
Label done;
JumpIfSmi(value, &done);
InNewSpace(object, value, equal, &done);
RecordWriteHelper(object, address, value);
bind(&done);
// Clobber all input registers when running with the debug-code flag
// turned on to provoke errors.
if (FLAG_debug_code) {
movq(object, BitCast<int64_t>(kZapValue), RelocInfo::NONE);
movq(address, BitCast<int64_t>(kZapValue), RelocInfo::NONE);
movq(value, BitCast<int64_t>(kZapValue), RelocInfo::NONE);
}
}
void MacroAssembler::RecordWriteNonSmi(Register object,
int offset,
Register scratch,
Register index) {
Label done;
if (FLAG_debug_code) {
Label okay;
JumpIfNotSmi(object, &okay);
Abort("MacroAssembler::RecordWriteNonSmi cannot deal with smis");
bind(&okay);
if (offset == 0) {
// index must be int32.
Register tmp = index.is(rax) ? rbx : rax;
push(tmp);
movl(tmp, index);
cmpq(tmp, index);
Check(equal, "Index register for RecordWrite must be untagged int32.");
pop(tmp);
}
}
// Test that the object address is not in the new space. We cannot
// update page dirty marks for new space pages.
InNewSpace(object, scratch, equal, &done);
// The offset is relative to a tagged or untagged HeapObject pointer,
// so either offset or offset + kHeapObjectTag must be a
// multiple of kPointerSize.
ASSERT(IsAligned(offset, kPointerSize) ||
IsAligned(offset + kHeapObjectTag, kPointerSize));
Register dst = index;
if (offset != 0) {
lea(dst, Operand(object, offset));
} else {
// array access: calculate the destination address in the same manner as
// KeyedStoreIC::GenerateGeneric.
lea(dst, FieldOperand(object,
index,
times_pointer_size,
FixedArray::kHeaderSize));
}
RecordWriteHelper(object, dst, scratch);
bind(&done);
// Clobber all input registers when running with the debug-code flag
// turned on to provoke errors.
if (FLAG_debug_code) {
movq(object, BitCast<int64_t>(kZapValue), RelocInfo::NONE);
movq(scratch, BitCast<int64_t>(kZapValue), RelocInfo::NONE);
movq(index, BitCast<int64_t>(kZapValue), RelocInfo::NONE);
}
}
void MacroAssembler::InNewSpace(Register object,
Register scratch,
Condition cc,
Label* branch) {
if (Serializer::enabled()) {
// Can't do arithmetic on external references if it might get serialized.
// The mask isn't really an address. We load it as an external reference in
// case the size of the new space is different between the snapshot maker
// and the running system.
if (scratch.is(object)) {
movq(kScratchRegister, ExternalReference::new_space_mask());
and_(scratch, kScratchRegister);
} else {
movq(scratch, ExternalReference::new_space_mask());
and_(scratch, object);
}
movq(kScratchRegister, ExternalReference::new_space_start());
cmpq(scratch, kScratchRegister);
j(cc, branch);
} else {
ASSERT(is_int32(static_cast<int64_t>(Heap::NewSpaceMask())));
intptr_t new_space_start =
reinterpret_cast<intptr_t>(Heap::NewSpaceStart());
movq(kScratchRegister, -new_space_start, RelocInfo::NONE);
if (scratch.is(object)) {
addq(scratch, kScratchRegister);
} else {
lea(scratch, Operand(object, kScratchRegister, times_1, 0));
}
and_(scratch, Immediate(static_cast<int32_t>(Heap::NewSpaceMask())));
j(cc, branch);
}
}
void MacroAssembler::Assert(Condition cc, const char* msg) {
if (FLAG_debug_code) Check(cc, msg);
}
void MacroAssembler::AssertFastElements(Register elements) {
if (FLAG_debug_code) {
Label ok;
CompareRoot(FieldOperand(elements, HeapObject::kMapOffset),
Heap::kFixedArrayMapRootIndex);
j(equal, &ok);
CompareRoot(FieldOperand(elements, HeapObject::kMapOffset),
Heap::kFixedCOWArrayMapRootIndex);
j(equal, &ok);
Abort("JSObject with fast elements map has slow elements");
bind(&ok);
}
}
void MacroAssembler::Check(Condition cc, const char* msg) {
Label L;
j(cc, &L);
Abort(msg);
// will not return here
bind(&L);
}
void MacroAssembler::CheckStackAlignment() {
int frame_alignment = OS::ActivationFrameAlignment();
int frame_alignment_mask = frame_alignment - 1;
if (frame_alignment > kPointerSize) {
ASSERT(IsPowerOf2(frame_alignment));
Label alignment_as_expected;
testq(rsp, Immediate(frame_alignment_mask));
j(zero, &alignment_as_expected);
// Abort if stack is not aligned.
int3();
bind(&alignment_as_expected);
}
}
void MacroAssembler::NegativeZeroTest(Register result,
Register op,
Label* then_label) {
Label ok;
testl(result, result);
j(not_zero, &ok);
testl(op, op);
j(sign, then_label);
bind(&ok);
}
void MacroAssembler::Abort(const char* msg) {
// We want to pass the msg string like a smi to avoid GC
// problems, however msg is not guaranteed to be aligned
// properly. Instead, we pass an aligned pointer that is
// a proper v8 smi, but also pass the alignment difference
// from the real pointer as a smi.
intptr_t p1 = reinterpret_cast<intptr_t>(msg);
intptr_t p0 = (p1 & ~kSmiTagMask) + kSmiTag;
// Note: p0 might not be a valid Smi *value*, but it has a valid Smi tag.
ASSERT(reinterpret_cast<Object*>(p0)->IsSmi());
#ifdef DEBUG
if (msg != NULL) {
RecordComment("Abort message: ");
RecordComment(msg);
}
#endif
// Disable stub call restrictions to always allow calls to abort.
set_allow_stub_calls(true);
push(rax);
movq(kScratchRegister, p0, RelocInfo::NONE);
push(kScratchRegister);
movq(kScratchRegister,
reinterpret_cast<intptr_t>(Smi::FromInt(static_cast<int>(p1 - p0))),
RelocInfo::NONE);
push(kScratchRegister);
CallRuntime(Runtime::kAbort, 2);
// will not return here
int3();
}
void MacroAssembler::CallStub(CodeStub* stub) {
ASSERT(allow_stub_calls()); // calls are not allowed in some stubs
Call(stub->GetCode(), RelocInfo::CODE_TARGET);
}
Object* MacroAssembler::TryCallStub(CodeStub* stub) {
ASSERT(allow_stub_calls()); // Calls are not allowed in some stubs.
Object* result = stub->TryGetCode();
if (!result->IsFailure()) {
call(Handle<Code>(Code::cast(result)), RelocInfo::CODE_TARGET);
}
return result;
}
void MacroAssembler::TailCallStub(CodeStub* stub) {
ASSERT(allow_stub_calls()); // calls are not allowed in some stubs
Jump(stub->GetCode(), RelocInfo::CODE_TARGET);
}
Object* MacroAssembler::TryTailCallStub(CodeStub* stub) {
ASSERT(allow_stub_calls()); // Calls are not allowed in some stubs.
Object* result = stub->TryGetCode();
if (!result->IsFailure()) {
jmp(Handle<Code>(Code::cast(result)), RelocInfo::CODE_TARGET);
}
return result;
}
void MacroAssembler::StubReturn(int argc) {
ASSERT(argc >= 1 && generating_stub());
ret((argc - 1) * kPointerSize);
}
void MacroAssembler::IllegalOperation(int num_arguments) {
if (num_arguments > 0) {
addq(rsp, Immediate(num_arguments * kPointerSize));
}
LoadRoot(rax, Heap::kUndefinedValueRootIndex);
}
void MacroAssembler::IndexFromHash(Register hash, Register index) {
// The assert checks that the constants for the maximum number of digits
// for an array index cached in the hash field and the number of bits
// reserved for it does not conflict.
ASSERT(TenToThe(String::kMaxCachedArrayIndexLength) <
(1 << String::kArrayIndexValueBits));
// We want the smi-tagged index in key. Even if we subsequently go to
// the slow case, converting the key to a smi is always valid.
// key: string key
// hash: key's hash field, including its array index value.
and_(hash, Immediate(String::kArrayIndexValueMask));
shr(hash, Immediate(String::kHashShift));
// Here we actually clobber the key which will be used if calling into
// runtime later. However as the new key is the numeric value of a string key
// there is no difference in using either key.
Integer32ToSmi(index, hash);
}
void MacroAssembler::CallRuntime(Runtime::FunctionId id, int num_arguments) {
CallRuntime(Runtime::FunctionForId(id), num_arguments);
}
Object* MacroAssembler::TryCallRuntime(Runtime::FunctionId id,
int num_arguments) {
return TryCallRuntime(Runtime::FunctionForId(id), num_arguments);
}
void MacroAssembler::CallRuntime(Runtime::Function* f, int num_arguments) {
// If the expected number of arguments of the runtime function is
// constant, we check that the actual number of arguments match the
// expectation.
if (f->nargs >= 0 && f->nargs != num_arguments) {
IllegalOperation(num_arguments);
return;
}
// TODO(1236192): Most runtime routines don't need the number of
// arguments passed in because it is constant. At some point we
// should remove this need and make the runtime routine entry code
// smarter.
Set(rax, num_arguments);
movq(rbx, ExternalReference(f));
CEntryStub ces(f->result_size);
CallStub(&ces);
}
Object* MacroAssembler::TryCallRuntime(Runtime::Function* f,
int num_arguments) {
if (f->nargs >= 0 && f->nargs != num_arguments) {
IllegalOperation(num_arguments);
// Since we did not call the stub, there was no allocation failure.
// Return some non-failure object.
return Heap::undefined_value();
}
// TODO(1236192): Most runtime routines don't need the number of
// arguments passed in because it is constant. At some point we
// should remove this need and make the runtime routine entry code
// smarter.
Set(rax, num_arguments);
movq(rbx, ExternalReference(f));
CEntryStub ces(f->result_size);
return TryCallStub(&ces);
}
void MacroAssembler::CallExternalReference(const ExternalReference& ext,
int num_arguments) {
Set(rax, num_arguments);
movq(rbx, ext);
CEntryStub stub(1);
CallStub(&stub);
}
void MacroAssembler::TailCallExternalReference(const ExternalReference& ext,
int num_arguments,
int result_size) {
// ----------- S t a t e -------------
// -- rsp[0] : return address
// -- rsp[8] : argument num_arguments - 1
// ...
// -- rsp[8 * num_arguments] : argument 0 (receiver)
// -----------------------------------
// TODO(1236192): Most runtime routines don't need the number of
// arguments passed in because it is constant. At some point we
// should remove this need and make the runtime routine entry code
// smarter.
Set(rax, num_arguments);
JumpToExternalReference(ext, result_size);
}
void MacroAssembler::TailCallRuntime(Runtime::FunctionId fid,
int num_arguments,
int result_size) {
TailCallExternalReference(ExternalReference(fid), num_arguments, result_size);
}
static int Offset(ExternalReference ref0, ExternalReference ref1) {
int64_t offset = (ref0.address() - ref1.address());
// Check that fits into int.
ASSERT(static_cast<int>(offset) == offset);
return static_cast<int>(offset);
}
void MacroAssembler::PushHandleScope(Register scratch) {
ExternalReference extensions_address =
ExternalReference::handle_scope_extensions_address();
const int kExtensionsOffset = 0;
const int kNextOffset = Offset(
ExternalReference::handle_scope_next_address(),
extensions_address);
const int kLimitOffset = Offset(
ExternalReference::handle_scope_limit_address(),
extensions_address);
// Push the number of extensions, smi-tagged so the gc will ignore it.
movq(kScratchRegister, extensions_address);
movq(scratch, Operand(kScratchRegister, kExtensionsOffset));
movq(Operand(kScratchRegister, kExtensionsOffset), Immediate(0));
Integer32ToSmi(scratch, scratch);
push(scratch);
// Push next and limit pointers which will be wordsize aligned and
// hence automatically smi tagged.
push(Operand(kScratchRegister, kNextOffset));
push(Operand(kScratchRegister, kLimitOffset));
}
Object* MacroAssembler::PopHandleScopeHelper(Register saved,
Register scratch,
bool gc_allowed) {
ExternalReference extensions_address =
ExternalReference::handle_scope_extensions_address();
const int kExtensionsOffset = 0;
const int kNextOffset = Offset(
ExternalReference::handle_scope_next_address(),
extensions_address);
const int kLimitOffset = Offset(
ExternalReference::handle_scope_limit_address(),
extensions_address);
Object* result = NULL;
Label write_back;
movq(kScratchRegister, extensions_address);
cmpq(Operand(kScratchRegister, kExtensionsOffset), Immediate(0));
j(equal, &write_back);
push(saved);
if (gc_allowed) {
CallRuntime(Runtime::kDeleteHandleScopeExtensions, 0);
} else {
result = TryCallRuntime(Runtime::kDeleteHandleScopeExtensions, 0);
if (result->IsFailure()) return result;
}
pop(saved);
movq(kScratchRegister, extensions_address);
bind(&write_back);
pop(Operand(kScratchRegister, kLimitOffset));
pop(Operand(kScratchRegister, kNextOffset));
pop(scratch);
SmiToInteger32(scratch, scratch);
movq(Operand(kScratchRegister, kExtensionsOffset), scratch);
return result;
}
void MacroAssembler::PopHandleScope(Register saved, Register scratch) {
PopHandleScopeHelper(saved, scratch, true);
}
Object* MacroAssembler::TryPopHandleScope(Register saved, Register scratch) {
return PopHandleScopeHelper(saved, scratch, false);
}
void MacroAssembler::JumpToExternalReference(const ExternalReference& ext,
int result_size) {
// Set the entry point and jump to the C entry runtime stub.
movq(rbx, ext);
CEntryStub ces(result_size);
jmp(ces.GetCode(), RelocInfo::CODE_TARGET);
}
void MacroAssembler::InvokeBuiltin(Builtins::JavaScript id, InvokeFlag flag) {
// Calls are not allowed in some stubs.
ASSERT(flag == JUMP_FUNCTION || allow_stub_calls());
// Rely on the assertion to check that the number of provided
// arguments match the expected number of arguments. Fake a
// parameter count to avoid emitting code to do the check.
ParameterCount expected(0);
GetBuiltinEntry(rdx, id);
InvokeCode(rdx, expected, expected, flag);
}
void MacroAssembler::GetBuiltinFunction(Register target,
Builtins::JavaScript id) {
// Load the builtins object into target register.
movq(target, Operand(rsi, Context::SlotOffset(Context::GLOBAL_INDEX)));
movq(target, FieldOperand(target, GlobalObject::kBuiltinsOffset));
movq(target, FieldOperand(target,
JSBuiltinsObject::OffsetOfFunctionWithId(id)));
}
void MacroAssembler::GetBuiltinEntry(Register target, Builtins::JavaScript id) {
ASSERT(!target.is(rdi));
// Load the JavaScript builtin function from the builtins object.
GetBuiltinFunction(rdi, id);
movq(target, FieldOperand(rdi, JSFunction::kCodeEntryOffset));
}
void MacroAssembler::Set(Register dst, int64_t x) {
if (x == 0) {
xorl(dst, dst);
} else if (is_int32(x)) {
movq(dst, Immediate(static_cast<int32_t>(x)));
} else if (is_uint32(x)) {
movl(dst, Immediate(static_cast<uint32_t>(x)));
} else {
movq(dst, x, RelocInfo::NONE);
}
}
void MacroAssembler::Set(const Operand& dst, int64_t x) {
if (is_int32(x)) {
movq(dst, Immediate(static_cast<int32_t>(x)));
} else {
movq(kScratchRegister, x, RelocInfo::NONE);
movq(dst, kScratchRegister);
}
}
// ----------------------------------------------------------------------------
// Smi tagging, untagging and tag detection.
static int kSmiShift = kSmiTagSize + kSmiShiftSize;
Register MacroAssembler::GetSmiConstant(Smi* source) {
int value = source->value();
if (value == 0) {
xorl(kScratchRegister, kScratchRegister);
return kScratchRegister;
}
if (value == 1) {
return kSmiConstantRegister;
}
LoadSmiConstant(kScratchRegister, source);
return kScratchRegister;
}
void MacroAssembler::LoadSmiConstant(Register dst, Smi* source) {
if (FLAG_debug_code) {
movq(dst,
reinterpret_cast<uint64_t>(Smi::FromInt(kSmiConstantRegisterValue)),
RelocInfo::NONE);
cmpq(dst, kSmiConstantRegister);
if (allow_stub_calls()) {
Assert(equal, "Uninitialized kSmiConstantRegister");
} else {
Label ok;
j(equal, &ok);
int3();
bind(&ok);
}
}
if (source->value() == 0) {
xorl(dst, dst);
return;
}
int value = source->value();
bool negative = value < 0;
unsigned int uvalue = negative ? -value : value;
switch (uvalue) {
case 9:
lea(dst, Operand(kSmiConstantRegister, kSmiConstantRegister, times_8, 0));
break;
case 8:
xorl(dst, dst);
lea(dst, Operand(dst, kSmiConstantRegister, times_8, 0));
break;
case 4:
xorl(dst, dst);
lea(dst, Operand(dst, kSmiConstantRegister, times_4, 0));
break;
case 5:
lea(dst, Operand(kSmiConstantRegister, kSmiConstantRegister, times_4, 0));
break;
case 3:
lea(dst, Operand(kSmiConstantRegister, kSmiConstantRegister, times_2, 0));
break;
case 2:
lea(dst, Operand(kSmiConstantRegister, kSmiConstantRegister, times_1, 0));
break;
case 1:
movq(dst, kSmiConstantRegister);
break;
case 0:
UNREACHABLE();
return;
default:
movq(dst, reinterpret_cast<uint64_t>(source), RelocInfo::NONE);
return;
}
if (negative) {
neg(dst);
}
}
void MacroAssembler::Integer32ToSmi(Register dst, Register src) {
ASSERT_EQ(0, kSmiTag);
if (!dst.is(src)) {
movl(dst, src);
}
shl(dst, Immediate(kSmiShift));
}
void MacroAssembler::Integer32ToSmi(Register dst,
Register src,
Label* on_overflow) {
ASSERT_EQ(0, kSmiTag);
// 32-bit integer always fits in a long smi.
if (!dst.is(src)) {
movl(dst, src);
}
shl(dst, Immediate(kSmiShift));
}
void MacroAssembler::Integer32ToSmiField(const Operand& dst, Register src) {
if (FLAG_debug_code) {
testb(dst, Immediate(0x01));
Label ok;
j(zero, &ok);
if (allow_stub_calls()) {
Abort("Integer32ToSmiField writing to non-smi location");
} else {
int3();
}
bind(&ok);
}
ASSERT(kSmiShift % kBitsPerByte == 0);
movl(Operand(dst, kSmiShift / kBitsPerByte), src);
}
void MacroAssembler::Integer64PlusConstantToSmi(Register dst,
Register src,
int constant) {
if (dst.is(src)) {
addq(dst, Immediate(constant));
} else {
lea(dst, Operand(src, constant));
}
shl(dst, Immediate(kSmiShift));
}
void MacroAssembler::SmiToInteger32(Register dst, Register src) {
ASSERT_EQ(0, kSmiTag);
if (!dst.is(src)) {
movq(dst, src);
}
shr(dst, Immediate(kSmiShift));
}
void MacroAssembler::SmiToInteger32(Register dst, const Operand& src) {
movl(dst, Operand(src, kSmiShift / kBitsPerByte));
}
void MacroAssembler::SmiToInteger64(Register dst, Register src) {
ASSERT_EQ(0, kSmiTag);
if (!dst.is(src)) {
movq(dst, src);
}
sar(dst, Immediate(kSmiShift));
}
void MacroAssembler::SmiToInteger64(Register dst, const Operand& src) {
movsxlq(dst, Operand(src, kSmiShift / kBitsPerByte));
}
void MacroAssembler::SmiTest(Register src) {
testq(src, src);
}
void MacroAssembler::SmiCompare(Register dst, Register src) {
cmpq(dst, src);
}
void MacroAssembler::SmiCompare(Register dst, Smi* src) {
ASSERT(!dst.is(kScratchRegister));
if (src->value() == 0) {
testq(dst, dst);
} else {
Register constant_reg = GetSmiConstant(src);
cmpq(dst, constant_reg);
}
}
void MacroAssembler::SmiCompare(Register dst, const Operand& src) {
cmpq(dst, src);
}
void MacroAssembler::SmiCompare(const Operand& dst, Register src) {
cmpq(dst, src);
}
void MacroAssembler::SmiCompare(const Operand& dst, Smi* src) {
cmpl(Operand(dst, kSmiShift / kBitsPerByte), Immediate(src->value()));
}
void MacroAssembler::SmiCompareInteger32(const Operand& dst, Register src) {
cmpl(Operand(dst, kSmiShift / kBitsPerByte), src);
}
void MacroAssembler::PositiveSmiTimesPowerOfTwoToInteger64(Register dst,
Register src,
int power) {
ASSERT(power >= 0);
ASSERT(power < 64);
if (power == 0) {
SmiToInteger64(dst, src);
return;
}
if (!dst.is(src)) {
movq(dst, src);
}
if (power < kSmiShift) {
sar(dst, Immediate(kSmiShift - power));
} else if (power > kSmiShift) {
shl(dst, Immediate(power - kSmiShift));
}
}
void MacroAssembler::PositiveSmiDivPowerOfTwoToInteger32(Register dst,
Register src,
int power) {
ASSERT((0 <= power) && (power < 32));
if (dst.is(src)) {
shr(dst, Immediate(power + kSmiShift));
} else {
UNIMPLEMENTED(); // Not used.
}
}
Condition MacroAssembler::CheckSmi(Register src) {
ASSERT_EQ(0, kSmiTag);
testb(src, Immediate(kSmiTagMask));
return zero;
}
Condition MacroAssembler::CheckPositiveSmi(Register src) {
ASSERT_EQ(0, kSmiTag);
// Make mask 0x8000000000000001 and test that both bits are zero.
movq(kScratchRegister, src);
rol(kScratchRegister, Immediate(1));
testb(kScratchRegister, Immediate(3));
return zero;
}
Condition MacroAssembler::CheckBothSmi(Register first, Register second) {
if (first.is(second)) {
return CheckSmi(first);
}
ASSERT(kSmiTag == 0 && kHeapObjectTag == 1 && kHeapObjectTagMask == 3);
leal(kScratchRegister, Operand(first, second, times_1, 0));
testb(kScratchRegister, Immediate(0x03));
return zero;
}
Condition MacroAssembler::CheckBothPositiveSmi(Register first,
Register second) {
if (first.is(second)) {
return CheckPositiveSmi(first);
}
movq(kScratchRegister, first);
or_(kScratchRegister, second);
rol(kScratchRegister, Immediate(1));
testl(kScratchRegister, Immediate(0x03));
return zero;
}
Condition MacroAssembler::CheckEitherSmi(Register first,
Register second,
Register scratch) {
if (first.is(second)) {
return CheckSmi(first);
}
if (scratch.is(second)) {
andl(scratch, first);
} else {
if (!scratch.is(first)) {
movl(scratch, first);
}
andl(scratch, second);
}
testb(scratch, Immediate(kSmiTagMask));
return zero;
}
Condition MacroAssembler::CheckIsMinSmi(Register src) {
ASSERT(!src.is(kScratchRegister));
// If we overflow by subtracting one, it's the minimal smi value.
cmpq(src, kSmiConstantRegister);
return overflow;
}
Condition MacroAssembler::CheckInteger32ValidSmiValue(Register src) {
// A 32-bit integer value can always be converted to a smi.
return always;
}
Condition MacroAssembler::CheckUInteger32ValidSmiValue(Register src) {
// An unsigned 32-bit integer value is valid as long as the high bit
// is not set.
testl(src, src);
return positive;
}
void MacroAssembler::SmiNeg(Register dst, Register src, Label* on_smi_result) {
if (dst.is(src)) {
ASSERT(!dst.is(kScratchRegister));
movq(kScratchRegister, src);
neg(dst); // Low 32 bits are retained as zero by negation.
// Test if result is zero or Smi::kMinValue.
cmpq(dst, kScratchRegister);
j(not_equal, on_smi_result);
movq(src, kScratchRegister);
} else {
movq(dst, src);
neg(dst);
cmpq(dst, src);
// If the result is zero or Smi::kMinValue, negation failed to create a smi.
j(not_equal, on_smi_result);
}
}
void MacroAssembler::SmiAdd(Register dst,
Register src1,
Register src2,
Label* on_not_smi_result) {
ASSERT(!dst.is(src2));
if (on_not_smi_result == NULL) {
// No overflow checking. Use only when it's known that
// overflowing is impossible.
if (dst.is(src1)) {
addq(dst, src2);
} else {
movq(dst, src1);
addq(dst, src2);
}
Assert(no_overflow, "Smi addition overflow");
} else if (dst.is(src1)) {
movq(kScratchRegister, src1);
addq(kScratchRegister, src2);
j(overflow, on_not_smi_result);
movq(dst, kScratchRegister);
} else {
movq(dst, src1);
addq(dst, src2);
j(overflow, on_not_smi_result);
}
}
void MacroAssembler::SmiSub(Register dst,
Register src1,
Register src2,
Label* on_not_smi_result) {
ASSERT(!dst.is(src2));
if (on_not_smi_result == NULL) {
// No overflow checking. Use only when it's known that
// overflowing is impossible (e.g., subtracting two positive smis).
if (dst.is(src1)) {
subq(dst, src2);
} else {
movq(dst, src1);
subq(dst, src2);
}
Assert(no_overflow, "Smi subtraction overflow");
} else if (dst.is(src1)) {
cmpq(dst, src2);
j(overflow, on_not_smi_result);
subq(dst, src2);
} else {
movq(dst, src1);
subq(dst, src2);
j(overflow, on_not_smi_result);
}
}
void MacroAssembler::SmiSub(Register dst,
Register src1,
const Operand& src2,
Label* on_not_smi_result) {
if (on_not_smi_result == NULL) {
// No overflow checking. Use only when it's known that
// overflowing is impossible (e.g., subtracting two positive smis).
if (dst.is(src1)) {
subq(dst, src2);
} else {
movq(dst, src1);
subq(dst, src2);
}
Assert(no_overflow, "Smi subtraction overflow");
} else if (dst.is(src1)) {
movq(kScratchRegister, src2);
cmpq(src1, kScratchRegister);
j(overflow, on_not_smi_result);
subq(src1, kScratchRegister);
} else {
movq(dst, src1);
subq(dst, src2);
j(overflow, on_not_smi_result);
}
}
void MacroAssembler::SmiMul(Register dst,
Register src1,
Register src2,
Label* on_not_smi_result) {
ASSERT(!dst.is(src2));
ASSERT(!dst.is(kScratchRegister));
ASSERT(!src1.is(kScratchRegister));
ASSERT(!src2.is(kScratchRegister));
if (dst.is(src1)) {
Label failure, zero_correct_result;
movq(kScratchRegister, src1); // Create backup for later testing.
SmiToInteger64(dst, src1);
imul(dst, src2);
j(overflow, &failure);
// Check for negative zero result. If product is zero, and one
// argument is negative, go to slow case.
Label correct_result;
testq(dst, dst);
j(not_zero, &correct_result);
movq(dst, kScratchRegister);
xor_(dst, src2);
j(positive, &zero_correct_result); // Result was positive zero.
bind(&failure); // Reused failure exit, restores src1.
movq(src1, kScratchRegister);
jmp(on_not_smi_result);
bind(&zero_correct_result);
xor_(dst, dst);
bind(&correct_result);
} else {
SmiToInteger64(dst, src1);
imul(dst, src2);
j(overflow, on_not_smi_result);
// Check for negative zero result. If product is zero, and one
// argument is negative, go to slow case.
Label correct_result;
testq(dst, dst);
j(not_zero, &correct_result);
// One of src1 and src2 is zero, the check whether the other is
// negative.
movq(kScratchRegister, src1);
xor_(kScratchRegister, src2);
j(negative, on_not_smi_result);
bind(&correct_result);
}
}
void MacroAssembler::SmiTryAddConstant(Register dst,
Register src,
Smi* constant,
Label* on_not_smi_result) {
// Does not assume that src is a smi.
ASSERT_EQ(static_cast<int>(1), static_cast<int>(kSmiTagMask));
ASSERT_EQ(0, kSmiTag);
ASSERT(!dst.is(kScratchRegister));
ASSERT(!src.is(kScratchRegister));
JumpIfNotSmi(src, on_not_smi_result);
Register tmp = (dst.is(src) ? kScratchRegister : dst);
LoadSmiConstant(tmp, constant);
addq(tmp, src);
j(overflow, on_not_smi_result);
if (dst.is(src)) {
movq(dst, tmp);
}
}
void MacroAssembler::SmiAddConstant(Register dst, Register src, Smi* constant) {
if (constant->value() == 0) {
if (!dst.is(src)) {
movq(dst, src);
}
return;
} else if (dst.is(src)) {
ASSERT(!dst.is(kScratchRegister));
switch (constant->value()) {
case 1:
addq(dst, kSmiConstantRegister);
return;
case 2:
lea(dst, Operand(src, kSmiConstantRegister, times_2, 0));
return;
case 4:
lea(dst, Operand(src, kSmiConstantRegister, times_4, 0));
return;
case 8:
lea(dst, Operand(src, kSmiConstantRegister, times_8, 0));
return;
default:
Register constant_reg = GetSmiConstant(constant);
addq(dst, constant_reg);
return;
}
} else {
switch (constant->value()) {
case 1:
lea(dst, Operand(src, kSmiConstantRegister, times_1, 0));
return;
case 2:
lea(dst, Operand(src, kSmiConstantRegister, times_2, 0));
return;
case 4:
lea(dst, Operand(src, kSmiConstantRegister, times_4, 0));
return;
case 8:
lea(dst, Operand(src, kSmiConstantRegister, times_8, 0));
return;
default:
LoadSmiConstant(dst, constant);
addq(dst, src);
return;
}
}
}
void MacroAssembler::SmiAddConstant(const Operand& dst, Smi* constant) {
if (constant->value() != 0) {
addl(Operand(dst, kSmiShift / kBitsPerByte), Immediate(constant->value()));
}
}
void MacroAssembler::SmiAddConstant(Register dst,
Register src,
Smi* constant,
Label* on_not_smi_result) {
if (constant->value() == 0) {
if (!dst.is(src)) {
movq(dst, src);
}
} else if (dst.is(src)) {
ASSERT(!dst.is(kScratchRegister));
LoadSmiConstant(kScratchRegister, constant);
addq(kScratchRegister, src);
j(overflow, on_not_smi_result);
movq(dst, kScratchRegister);
} else {
LoadSmiConstant(dst, constant);
addq(dst, src);
j(overflow, on_not_smi_result);
}
}
void MacroAssembler::SmiSubConstant(Register dst, Register src, Smi* constant) {
if (constant->value() == 0) {
if (!dst.is(src)) {
movq(dst, src);
}
} else if (dst.is(src)) {
ASSERT(!dst.is(kScratchRegister));
Register constant_reg = GetSmiConstant(constant);
subq(dst, constant_reg);
} else {
if (constant->value() == Smi::kMinValue) {
LoadSmiConstant(dst, constant);
// Adding and subtracting the min-value gives the same result, it only
// differs on the overflow bit, which we don't check here.
addq(dst, src);
} else {
// Subtract by adding the negation.
LoadSmiConstant(dst, Smi::FromInt(-constant->value()));
addq(dst, src);
}
}
}
void MacroAssembler::SmiSubConstant(Register dst,
Register src,
Smi* constant,
Label* on_not_smi_result) {
if (constant->value() == 0) {
if (!dst.is(src)) {
movq(dst, src);
}
} else if (dst.is(src)) {
ASSERT(!dst.is(kScratchRegister));
if (constant->value() == Smi::kMinValue) {
// Subtracting min-value from any non-negative value will overflow.
// We test the non-negativeness before doing the subtraction.
testq(src, src);
j(not_sign, on_not_smi_result);
LoadSmiConstant(kScratchRegister, constant);
subq(dst, kScratchRegister);
} else {
// Subtract by adding the negation.
LoadSmiConstant(kScratchRegister, Smi::FromInt(-constant->value()));
addq(kScratchRegister, dst);
j(overflow, on_not_smi_result);
movq(dst, kScratchRegister);
}
} else {
if (constant->value() == Smi::kMinValue) {
// Subtracting min-value from any non-negative value will overflow.
// We test the non-negativeness before doing the subtraction.
testq(src, src);
j(not_sign, on_not_smi_result);
LoadSmiConstant(dst, constant);
// Adding and subtracting the min-value gives the same result, it only
// differs on the overflow bit, which we don't check here.
addq(dst, src);
} else {
// Subtract by adding the negation.
LoadSmiConstant(dst, Smi::FromInt(-(constant->value())));
addq(dst, src);
j(overflow, on_not_smi_result);
}
}
}
void MacroAssembler::SmiDiv(Register dst,
Register src1,
Register src2,
Label* on_not_smi_result) {
ASSERT(!src1.is(kScratchRegister));
ASSERT(!src2.is(kScratchRegister));
ASSERT(!dst.is(kScratchRegister));
ASSERT(!src2.is(rax));
ASSERT(!src2.is(rdx));
ASSERT(!src1.is(rdx));
// Check for 0 divisor (result is +/-Infinity).
Label positive_divisor;
testq(src2, src2);
j(zero, on_not_smi_result);
if (src1.is(rax)) {
movq(kScratchRegister, src1);
}
SmiToInteger32(rax, src1);
// We need to rule out dividing Smi::kMinValue by -1, since that would
// overflow in idiv and raise an exception.
// We combine this with negative zero test (negative zero only happens
// when dividing zero by a negative number).
// We overshoot a little and go to slow case if we divide min-value
// by any negative value, not just -1.
Label safe_div;
testl(rax, Immediate(0x7fffffff));
j(not_zero, &safe_div);
testq(src2, src2);
if (src1.is(rax)) {
j(positive, &safe_div);
movq(src1, kScratchRegister);
jmp(on_not_smi_result);
} else {
j(negative, on_not_smi_result);
}
bind(&safe_div);
SmiToInteger32(src2, src2);
// Sign extend src1 into edx:eax.
cdq();
idivl(src2);
Integer32ToSmi(src2, src2);
// Check that the remainder is zero.
testl(rdx, rdx);
if (src1.is(rax)) {
Label smi_result;
j(zero, &smi_result);
movq(src1, kScratchRegister);
jmp(on_not_smi_result);
bind(&smi_result);
} else {
j(not_zero, on_not_smi_result);
}
if (!dst.is(src1) && src1.is(rax)) {
movq(src1, kScratchRegister);
}
Integer32ToSmi(dst, rax);
}
void MacroAssembler::SmiMod(Register dst,
Register src1,
Register src2,
Label* on_not_smi_result) {
ASSERT(!dst.is(kScratchRegister));
ASSERT(!src1.is(kScratchRegister));
ASSERT(!src2.is(kScratchRegister));
ASSERT(!src2.is(rax));
ASSERT(!src2.is(rdx));
ASSERT(!src1.is(rdx));
ASSERT(!src1.is(src2));
testq(src2, src2);
j(zero, on_not_smi_result);
if (src1.is(rax)) {
movq(kScratchRegister, src1);
}
SmiToInteger32(rax, src1);
SmiToInteger32(src2, src2);
// Test for the edge case of dividing Smi::kMinValue by -1 (will overflow).
Label safe_div;
cmpl(rax, Immediate(Smi::kMinValue));
j(not_equal, &safe_div);
cmpl(src2, Immediate(-1));
j(not_equal, &safe_div);
// Retag inputs and go slow case.
Integer32ToSmi(src2, src2);
if (src1.is(rax)) {
movq(src1, kScratchRegister);
}
jmp(on_not_smi_result);
bind(&safe_div);
// Sign extend eax into edx:eax.
cdq();
idivl(src2);
// Restore smi tags on inputs.
Integer32ToSmi(src2, src2);
if (src1.is(rax)) {
movq(src1, kScratchRegister);
}
// Check for a negative zero result. If the result is zero, and the
// dividend is negative, go slow to return a floating point negative zero.
Label smi_result;
testl(rdx, rdx);
j(not_zero, &smi_result);
testq(src1, src1);
j(negative, on_not_smi_result);
bind(&smi_result);
Integer32ToSmi(dst, rdx);
}
void MacroAssembler::SmiNot(Register dst, Register src) {
ASSERT(!dst.is(kScratchRegister));
ASSERT(!src.is(kScratchRegister));
// Set tag and padding bits before negating, so that they are zero afterwards.
movl(kScratchRegister, Immediate(~0));
if (dst.is(src)) {
xor_(dst, kScratchRegister);
} else {
lea(dst, Operand(src, kScratchRegister, times_1, 0));
}
not_(dst);
}
void MacroAssembler::SmiAnd(Register dst, Register src1, Register src2) {
ASSERT(!dst.is(src2));
if (!dst.is(src1)) {
movq(dst, src1);
}
and_(dst, src2);
}
void MacroAssembler::SmiAndConstant(Register dst, Register src, Smi* constant) {
if (constant->value() == 0) {
xor_(dst, dst);
} else if (dst.is(src)) {
ASSERT(!dst.is(kScratchRegister));
Register constant_reg = GetSmiConstant(constant);
and_(dst, constant_reg);
} else {
LoadSmiConstant(dst, constant);
and_(dst, src);
}
}
void MacroAssembler::SmiOr(Register dst, Register src1, Register src2) {
if (!dst.is(src1)) {
movq(dst, src1);
}
or_(dst, src2);
}
void MacroAssembler::SmiOrConstant(Register dst, Register src, Smi* constant) {
if (dst.is(src)) {
ASSERT(!dst.is(kScratchRegister));
Register constant_reg = GetSmiConstant(constant);
or_(dst, constant_reg);
} else {
LoadSmiConstant(dst, constant);
or_(dst, src);
}
}
void MacroAssembler::SmiXor(Register dst, Register src1, Register src2) {
if (!dst.is(src1)) {
movq(dst, src1);
}
xor_(dst, src2);
}
void MacroAssembler::SmiXorConstant(Register dst, Register src, Smi* constant) {
if (dst.is(src)) {
ASSERT(!dst.is(kScratchRegister));
Register constant_reg = GetSmiConstant(constant);
xor_(dst, constant_reg);
} else {
LoadSmiConstant(dst, constant);
xor_(dst, src);
}
}
void MacroAssembler::SmiShiftArithmeticRightConstant(Register dst,
Register src,
int shift_value) {
ASSERT(is_uint5(shift_value));
if (shift_value > 0) {
if (dst.is(src)) {
sar(dst, Immediate(shift_value + kSmiShift));
shl(dst, Immediate(kSmiShift));
} else {
UNIMPLEMENTED(); // Not used.
}
}
}
void MacroAssembler::SmiShiftLogicalRightConstant(Register dst,
Register src,
int shift_value,
Label* on_not_smi_result) {
// Logic right shift interprets its result as an *unsigned* number.
if (dst.is(src)) {
UNIMPLEMENTED(); // Not used.
} else {
movq(dst, src);
if (shift_value == 0) {
testq(dst, dst);
j(negative, on_not_smi_result);
}
shr(dst, Immediate(shift_value + kSmiShift));
shl(dst, Immediate(kSmiShift));
}
}
void MacroAssembler::SmiShiftLeftConstant(Register dst,
Register src,
int shift_value) {
if (!dst.is(src)) {
movq(dst, src);
}
if (shift_value > 0) {
shl(dst, Immediate(shift_value));
}
}
void MacroAssembler::SmiShiftLeft(Register dst,
Register src1,
Register src2) {
ASSERT(!dst.is(rcx));
Label result_ok;
// Untag shift amount.
if (!dst.is(src1)) {
movq(dst, src1);
}
SmiToInteger32(rcx, src2);
// Shift amount specified by lower 5 bits, not six as the shl opcode.
and_(rcx, Immediate(0x1f));
shl_cl(dst);
}
void MacroAssembler::SmiShiftLogicalRight(Register dst,
Register src1,
Register src2,
Label* on_not_smi_result) {
ASSERT(!dst.is(kScratchRegister));
ASSERT(!src1.is(kScratchRegister));
ASSERT(!src2.is(kScratchRegister));
ASSERT(!dst.is(rcx));
Label result_ok;
if (src1.is(rcx) || src2.is(rcx)) {
movq(kScratchRegister, rcx);
}
if (!dst.is(src1)) {
movq(dst, src1);
}
SmiToInteger32(rcx, src2);
orl(rcx, Immediate(kSmiShift));
shr_cl(dst); // Shift is rcx modulo 0x1f + 32.
shl(dst, Immediate(kSmiShift));
testq(dst, dst);
if (src1.is(rcx) || src2.is(rcx)) {
Label positive_result;
j(positive, &positive_result);
if (src1.is(rcx)) {
movq(src1, kScratchRegister);
} else {
movq(src2, kScratchRegister);
}
jmp(on_not_smi_result);
bind(&positive_result);
} else {
j(negative, on_not_smi_result); // src2 was zero and src1 negative.
}
}
void MacroAssembler::SmiShiftArithmeticRight(Register dst,
Register src1,
Register src2) {
ASSERT(!dst.is(kScratchRegister));
ASSERT(!src1.is(kScratchRegister));
ASSERT(!src2.is(kScratchRegister));
ASSERT(!dst.is(rcx));
if (src1.is(rcx)) {
movq(kScratchRegister, src1);
} else if (src2.is(rcx)) {
movq(kScratchRegister, src2);
}
if (!dst.is(src1)) {
movq(dst, src1);
}
SmiToInteger32(rcx, src2);
orl(rcx, Immediate(kSmiShift));
sar_cl(dst); // Shift 32 + original rcx & 0x1f.
shl(dst, Immediate(kSmiShift));
if (src1.is(rcx)) {
movq(src1, kScratchRegister);
} else if (src2.is(rcx)) {
movq(src2, kScratchRegister);
}
}
void MacroAssembler::SelectNonSmi(Register dst,
Register src1,
Register src2,
Label* on_not_smis) {
ASSERT(!dst.is(kScratchRegister));
ASSERT(!src1.is(kScratchRegister));
ASSERT(!src2.is(kScratchRegister));
ASSERT(!dst.is(src1));
ASSERT(!dst.is(src2));
// Both operands must not be smis.
#ifdef DEBUG
if (allow_stub_calls()) { // Check contains a stub call.
Condition not_both_smis = NegateCondition(CheckBothSmi(src1, src2));
Check(not_both_smis, "Both registers were smis in SelectNonSmi.");
}
#endif
ASSERT_EQ(0, kSmiTag);
ASSERT_EQ(0, Smi::FromInt(0));
movl(kScratchRegister, Immediate(kSmiTagMask));
and_(kScratchRegister, src1);
testl(kScratchRegister, src2);
// If non-zero then both are smis.
j(not_zero, on_not_smis);
// Exactly one operand is a smi.
ASSERT_EQ(1, static_cast<int>(kSmiTagMask));
// kScratchRegister still holds src1 & kSmiTag, which is either zero or one.
subq(kScratchRegister, Immediate(1));
// If src1 is a smi, then scratch register all 1s, else it is all 0s.
movq(dst, src1);
xor_(dst, src2);
and_(dst, kScratchRegister);
// If src1 is a smi, dst holds src1 ^ src2, else it is zero.
xor_(dst, src1);
// If src1 is a smi, dst is src2, else it is src1, i.e., the non-smi.
}
SmiIndex MacroAssembler::SmiToIndex(Register dst,
Register src,
int shift) {
ASSERT(is_uint6(shift));
// There is a possible optimization if shift is in the range 60-63, but that
// will (and must) never happen.
if (!dst.is(src)) {
movq(dst, src);
}
if (shift < kSmiShift) {
sar(dst, Immediate(kSmiShift - shift));
} else {
shl(dst, Immediate(shift - kSmiShift));
}
return SmiIndex(dst, times_1);
}
SmiIndex MacroAssembler::SmiToNegativeIndex(Register dst,
Register src,
int shift) {
// Register src holds a positive smi.
ASSERT(is_uint6(shift));
if (!dst.is(src)) {
movq(dst, src);
}
neg(dst);
if (shift < kSmiShift) {
sar(dst, Immediate(kSmiShift - shift));
} else {
shl(dst, Immediate(shift - kSmiShift));
}
return SmiIndex(dst, times_1);
}
void MacroAssembler::JumpIfSmi(Register src, Label* on_smi) {
ASSERT_EQ(0, kSmiTag);
Condition smi = CheckSmi(src);
j(smi, on_smi);
}
void MacroAssembler::JumpIfNotSmi(Register src, Label* on_not_smi) {
Condition smi = CheckSmi(src);
j(NegateCondition(smi), on_not_smi);
}
void MacroAssembler::JumpIfNotPositiveSmi(Register src,
Label* on_not_positive_smi) {
Condition positive_smi = CheckPositiveSmi(src);
j(NegateCondition(positive_smi), on_not_positive_smi);
}
void MacroAssembler::JumpIfSmiEqualsConstant(Register src,
Smi* constant,
Label* on_equals) {
SmiCompare(src, constant);
j(equal, on_equals);
}
void MacroAssembler::JumpIfNotValidSmiValue(Register src, Label* on_invalid) {
Condition is_valid = CheckInteger32ValidSmiValue(src);
j(NegateCondition(is_valid), on_invalid);
}
void MacroAssembler::JumpIfUIntNotValidSmiValue(Register src,
Label* on_invalid) {
Condition is_valid = CheckUInteger32ValidSmiValue(src);
j(NegateCondition(is_valid), on_invalid);
}
void MacroAssembler::JumpIfNotBothSmi(Register src1, Register src2,
Label* on_not_both_smi) {
Condition both_smi = CheckBothSmi(src1, src2);
j(NegateCondition(both_smi), on_not_both_smi);
}
void MacroAssembler::JumpIfNotBothPositiveSmi(Register src1, Register src2,
Label* on_not_both_smi) {
Condition both_smi = CheckBothPositiveSmi(src1, src2);
j(NegateCondition(both_smi), on_not_both_smi);
}
void MacroAssembler::JumpIfNotBothSequentialAsciiStrings(Register first_object,
Register second_object,
Register scratch1,
Register scratch2,
Label* on_fail) {
// Check that both objects are not smis.
Condition either_smi = CheckEitherSmi(first_object, second_object);
j(either_smi, on_fail);
// Load instance type for both strings.
movq(scratch1, FieldOperand(first_object, HeapObject::kMapOffset));
movq(scratch2, FieldOperand(second_object, HeapObject::kMapOffset));
movzxbl(scratch1, FieldOperand(scratch1, Map::kInstanceTypeOffset));
movzxbl(scratch2, FieldOperand(scratch2, Map::kInstanceTypeOffset));
// Check that both are flat ascii strings.
ASSERT(kNotStringTag != 0);
const int kFlatAsciiStringMask =
kIsNotStringMask | kStringRepresentationMask | kStringEncodingMask;
const int kFlatAsciiStringTag = ASCII_STRING_TYPE;
andl(scratch1, Immediate(kFlatAsciiStringMask));
andl(scratch2, Immediate(kFlatAsciiStringMask));
// Interleave the bits to check both scratch1 and scratch2 in one test.
ASSERT_EQ(0, kFlatAsciiStringMask & (kFlatAsciiStringMask << 3));
lea(scratch1, Operand(scratch1, scratch2, times_8, 0));
cmpl(scratch1,
Immediate(kFlatAsciiStringTag + (kFlatAsciiStringTag << 3)));
j(not_equal, on_fail);
}
void MacroAssembler::JumpIfInstanceTypeIsNotSequentialAscii(
Register instance_type,
Register scratch,
Label *failure) {
if (!scratch.is(instance_type)) {
movl(scratch, instance_type);
}
const int kFlatAsciiStringMask =
kIsNotStringMask | kStringRepresentationMask | kStringEncodingMask;
andl(scratch, Immediate(kFlatAsciiStringMask));
cmpl(scratch, Immediate(kStringTag | kSeqStringTag | kAsciiStringTag));
j(not_equal, failure);
}
void MacroAssembler::JumpIfBothInstanceTypesAreNotSequentialAscii(
Register first_object_instance_type,
Register second_object_instance_type,
Register scratch1,
Register scratch2,
Label* on_fail) {
// Load instance type for both strings.
movq(scratch1, first_object_instance_type);
movq(scratch2, second_object_instance_type);
// Check that both are flat ascii strings.
ASSERT(kNotStringTag != 0);
const int kFlatAsciiStringMask =
kIsNotStringMask | kStringRepresentationMask | kStringEncodingMask;
const int kFlatAsciiStringTag = ASCII_STRING_TYPE;
andl(scratch1, Immediate(kFlatAsciiStringMask));
andl(scratch2, Immediate(kFlatAsciiStringMask));
// Interleave the bits to check both scratch1 and scratch2 in one test.
ASSERT_EQ(0, kFlatAsciiStringMask & (kFlatAsciiStringMask << 3));
lea(scratch1, Operand(scratch1, scratch2, times_8, 0));
cmpl(scratch1,
Immediate(kFlatAsciiStringTag + (kFlatAsciiStringTag << 3)));
j(not_equal, on_fail);
}
void MacroAssembler::Move(Register dst, Handle<Object> source) {
ASSERT(!source->IsFailure());
if (source->IsSmi()) {
Move(dst, Smi::cast(*source));
} else {
movq(dst, source, RelocInfo::EMBEDDED_OBJECT);
}
}
void MacroAssembler::Move(const Operand& dst, Handle<Object> source) {
ASSERT(!source->IsFailure());
if (source->IsSmi()) {
Move(dst, Smi::cast(*source));
} else {
movq(kScratchRegister, source, RelocInfo::EMBEDDED_OBJECT);
movq(dst, kScratchRegister);
}
}
void MacroAssembler::Cmp(Register dst, Handle<Object> source) {
if (source->IsSmi()) {
SmiCompare(dst, Smi::cast(*source));
} else {
Move(kScratchRegister, source);
cmpq(dst, kScratchRegister);
}
}
void MacroAssembler::Cmp(const Operand& dst, Handle<Object> source) {
if (source->IsSmi()) {
SmiCompare(dst, Smi::cast(*source));
} else {
ASSERT(source->IsHeapObject());
movq(kScratchRegister, source, RelocInfo::EMBEDDED_OBJECT);
cmpq(dst, kScratchRegister);
}
}
void MacroAssembler::Push(Handle<Object> source) {
if (source->IsSmi()) {
Push(Smi::cast(*source));
} else {
ASSERT(source->IsHeapObject());
movq(kScratchRegister, source, RelocInfo::EMBEDDED_OBJECT);
push(kScratchRegister);
}
}
void MacroAssembler::Push(Smi* source) {
intptr_t smi = reinterpret_cast<intptr_t>(source);
if (is_int32(smi)) {
push(Immediate(static_cast<int32_t>(smi)));
} else {
Register constant = GetSmiConstant(source);
push(constant);
}
}
void MacroAssembler::Drop(int stack_elements) {
if (stack_elements > 0) {
addq(rsp, Immediate(stack_elements * kPointerSize));
}
}
void MacroAssembler::Test(const Operand& src, Smi* source) {
testl(Operand(src, kIntSize), Immediate(source->value()));
}
void MacroAssembler::Jump(ExternalReference ext) {
movq(kScratchRegister, ext);
jmp(kScratchRegister);
}
void MacroAssembler::Jump(Address destination, RelocInfo::Mode rmode) {
movq(kScratchRegister, destination, rmode);
jmp(kScratchRegister);
}
void MacroAssembler::Jump(Handle<Code> code_object, RelocInfo::Mode rmode) {
// TODO(X64): Inline this
jmp(code_object, rmode);
}
void MacroAssembler::Call(ExternalReference ext) {
movq(kScratchRegister, ext);
call(kScratchRegister);
}
void MacroAssembler::Call(Address destination, RelocInfo::Mode rmode) {
movq(kScratchRegister, destination, rmode);
call(kScratchRegister);
}
void MacroAssembler::Call(Handle<Code> code_object, RelocInfo::Mode rmode) {
ASSERT(RelocInfo::IsCodeTarget(rmode));
WriteRecordedPositions();
call(code_object, rmode);
}
void MacroAssembler::PushTryHandler(CodeLocation try_location,
HandlerType type) {
// Adjust this code if not the case.
ASSERT(StackHandlerConstants::kSize == 4 * kPointerSize);
// The pc (return address) is already on TOS. This code pushes state,
// frame pointer and current handler. Check that they are expected
// next on the stack, in that order.
ASSERT_EQ(StackHandlerConstants::kStateOffset,
StackHandlerConstants::kPCOffset - kPointerSize);
ASSERT_EQ(StackHandlerConstants::kFPOffset,
StackHandlerConstants::kStateOffset - kPointerSize);
ASSERT_EQ(StackHandlerConstants::kNextOffset,
StackHandlerConstants::kFPOffset - kPointerSize);
if (try_location == IN_JAVASCRIPT) {
if (type == TRY_CATCH_HANDLER) {
push(Immediate(StackHandler::TRY_CATCH));
} else {
push(Immediate(StackHandler::TRY_FINALLY));
}
push(rbp);
} else {
ASSERT(try_location == IN_JS_ENTRY);
// The frame pointer does not point to a JS frame so we save NULL
// for rbp. We expect the code throwing an exception to check rbp
// before dereferencing it to restore the context.
push(Immediate(StackHandler::ENTRY));
push(Immediate(0)); // NULL frame pointer.
}
// Save the current handler.
movq(kScratchRegister, ExternalReference(Top::k_handler_address));
push(Operand(kScratchRegister, 0));
// Link this handler.
movq(Operand(kScratchRegister, 0), rsp);
}
void MacroAssembler::PopTryHandler() {
ASSERT_EQ(0, StackHandlerConstants::kNextOffset);
// Unlink this handler.
movq(kScratchRegister, ExternalReference(Top::k_handler_address));
pop(Operand(kScratchRegister, 0));
// Remove the remaining fields.
addq(rsp, Immediate(StackHandlerConstants::kSize - kPointerSize));
}
void MacroAssembler::Ret() {
ret(0);
}
void MacroAssembler::FCmp() {
fucomip();
fstp(0);
}
void MacroAssembler::CmpObjectType(Register heap_object,
InstanceType type,
Register map) {
movq(map, FieldOperand(heap_object, HeapObject::kMapOffset));
CmpInstanceType(map, type);
}
void MacroAssembler::CmpInstanceType(Register map, InstanceType type) {
cmpb(FieldOperand(map, Map::kInstanceTypeOffset),
Immediate(static_cast<int8_t>(type)));
}
void MacroAssembler::CheckMap(Register obj,
Handle<Map> map,
Label* fail,
bool is_heap_object) {
if (!is_heap_object) {
JumpIfSmi(obj, fail);
}
Cmp(FieldOperand(obj, HeapObject::kMapOffset), map);
j(not_equal, fail);
}
void MacroAssembler::AbortIfNotNumber(Register object) {
Label ok;
Condition is_smi = CheckSmi(object);
j(is_smi, &ok);
Cmp(FieldOperand(object, HeapObject::kMapOffset),
Factory::heap_number_map());
Assert(equal, "Operand not a number");
bind(&ok);
}
void MacroAssembler::AbortIfSmi(Register object) {
Label ok;
Condition is_smi = CheckSmi(object);
Assert(NegateCondition(is_smi), "Operand is a smi");
}
void MacroAssembler::AbortIfNotSmi(Register object) {
Label ok;
Condition is_smi = CheckSmi(object);
Assert(is_smi, "Operand is not a smi");
}
void MacroAssembler::AbortIfNotRootValue(Register src,
Heap::RootListIndex root_value_index,
const char* message) {
ASSERT(!src.is(kScratchRegister));
LoadRoot(kScratchRegister, root_value_index);
cmpq(src, kScratchRegister);
Check(equal, message);
}
Condition MacroAssembler::IsObjectStringType(Register heap_object,
Register map,
Register instance_type) {
movq(map, FieldOperand(heap_object, HeapObject::kMapOffset));
movzxbl(instance_type, FieldOperand(map, Map::kInstanceTypeOffset));
ASSERT(kNotStringTag != 0);
testb(instance_type, Immediate(kIsNotStringMask));
return zero;
}
void MacroAssembler::TryGetFunctionPrototype(Register function,
Register result,
Label* miss) {
// Check that the receiver isn't a smi.
testl(function, Immediate(kSmiTagMask));
j(zero, miss);
// Check that the function really is a function.
CmpObjectType(function, JS_FUNCTION_TYPE, result);
j(not_equal, miss);
// Make sure that the function has an instance prototype.
Label non_instance;
testb(FieldOperand(result, Map::kBitFieldOffset),
Immediate(1 << Map::kHasNonInstancePrototype));
j(not_zero, &non_instance);
// Get the prototype or initial map from the function.
movq(result,
FieldOperand(function, JSFunction::kPrototypeOrInitialMapOffset));
// If the prototype or initial map is the hole, don't return it and
// simply miss the cache instead. This will allow us to allocate a
// prototype object on-demand in the runtime system.
CompareRoot(result, Heap::kTheHoleValueRootIndex);
j(equal, miss);
// If the function does not have an initial map, we're done.
Label done;
CmpObjectType(result, MAP_TYPE, kScratchRegister);
j(not_equal, &done);
// Get the prototype from the initial map.
movq(result, FieldOperand(result, Map::kPrototypeOffset));
jmp(&done);
// Non-instance prototype: Fetch prototype from constructor field
// in initial map.
bind(&non_instance);
movq(result, FieldOperand(result, Map::kConstructorOffset));
// All done.
bind(&done);
}
void MacroAssembler::SetCounter(StatsCounter* counter, int value) {
if (FLAG_native_code_counters && counter->Enabled()) {
movq(kScratchRegister, ExternalReference(counter));
movl(Operand(kScratchRegister, 0), Immediate(value));
}
}
void MacroAssembler::IncrementCounter(StatsCounter* counter, int value) {
ASSERT(value > 0);
if (FLAG_native_code_counters && counter->Enabled()) {
movq(kScratchRegister, ExternalReference(counter));
Operand operand(kScratchRegister, 0);
if (value == 1) {
incl(operand);
} else {
addl(operand, Immediate(value));
}
}
}
void MacroAssembler::DecrementCounter(StatsCounter* counter, int value) {
ASSERT(value > 0);
if (FLAG_native_code_counters && counter->Enabled()) {
movq(kScratchRegister, ExternalReference(counter));
Operand operand(kScratchRegister, 0);
if (value == 1) {
decl(operand);
} else {
subl(operand, Immediate(value));
}
}
}
#ifdef ENABLE_DEBUGGER_SUPPORT
void MacroAssembler::DebugBreak() {
ASSERT(allow_stub_calls());
xor_(rax, rax); // no arguments
movq(rbx, ExternalReference(Runtime::kDebugBreak));
CEntryStub ces(1);
Call(ces.GetCode(), RelocInfo::DEBUG_BREAK);
}
#endif // ENABLE_DEBUGGER_SUPPORT
void MacroAssembler::InvokePrologue(const ParameterCount& expected,
const ParameterCount& actual,
Handle<Code> code_constant,
Register code_register,
Label* done,
InvokeFlag flag) {
bool definitely_matches = false;
Label invoke;
if (expected.is_immediate()) {
ASSERT(actual.is_immediate());
if (expected.immediate() == actual.immediate()) {
definitely_matches = true;
} else {
Set(rax, actual.immediate());
if (expected.immediate() ==
SharedFunctionInfo::kDontAdaptArgumentsSentinel) {
// Don't worry about adapting arguments for built-ins that
// don't want that done. Skip adaption code by making it look
// like we have a match between expected and actual number of
// arguments.
definitely_matches = true;
} else {
Set(rbx, expected.immediate());
}
}
} else {
if (actual.is_immediate()) {
// Expected is in register, actual is immediate. This is the
// case when we invoke function values without going through the
// IC mechanism.
cmpq(expected.reg(), Immediate(actual.immediate()));
j(equal, &invoke);
ASSERT(expected.reg().is(rbx));
Set(rax, actual.immediate());
} else if (!expected.reg().is(actual.reg())) {
// Both expected and actual are in (different) registers. This
// is the case when we invoke functions using call and apply.
cmpq(expected.reg(), actual.reg());
j(equal, &invoke);
ASSERT(actual.reg().is(rax));
ASSERT(expected.reg().is(rbx));
}
}
if (!definitely_matches) {
Handle<Code> adaptor =
Handle<Code>(Builtins::builtin(Builtins::ArgumentsAdaptorTrampoline));
if (!code_constant.is_null()) {
movq(rdx, code_constant, RelocInfo::EMBEDDED_OBJECT);
addq(rdx, Immediate(Code::kHeaderSize - kHeapObjectTag));
} else if (!code_register.is(rdx)) {
movq(rdx, code_register);
}
if (flag == CALL_FUNCTION) {
Call(adaptor, RelocInfo::CODE_TARGET);
jmp(done);
} else {
Jump(adaptor, RelocInfo::CODE_TARGET);
}
bind(&invoke);
}
}
void MacroAssembler::InvokeCode(Register code,
const ParameterCount& expected,
const ParameterCount& actual,
InvokeFlag flag) {
Label done;
InvokePrologue(expected, actual, Handle<Code>::null(), code, &done, flag);
if (flag == CALL_FUNCTION) {
call(code);
} else {
ASSERT(flag == JUMP_FUNCTION);
jmp(code);
}
bind(&done);
}
void MacroAssembler::InvokeCode(Handle<Code> code,
const ParameterCount& expected,
const ParameterCount& actual,
RelocInfo::Mode rmode,
InvokeFlag flag) {
Label done;
Register dummy = rax;
InvokePrologue(expected, actual, code, dummy, &done, flag);
if (flag == CALL_FUNCTION) {
Call(code, rmode);
} else {
ASSERT(flag == JUMP_FUNCTION);
Jump(code, rmode);
}
bind(&done);
}
void MacroAssembler::InvokeFunction(Register function,
const ParameterCount& actual,
InvokeFlag flag) {
ASSERT(function.is(rdi));
movq(rdx, FieldOperand(function, JSFunction::kSharedFunctionInfoOffset));
movq(rsi, FieldOperand(function, JSFunction::kContextOffset));
movsxlq(rbx,
FieldOperand(rdx, SharedFunctionInfo::kFormalParameterCountOffset));
// Advances rdx to the end of the Code object header, to the start of
// the executable code.
movq(rdx, FieldOperand(rdi, JSFunction::kCodeEntryOffset));
ParameterCount expected(rbx);
InvokeCode(rdx, expected, actual, flag);
}
void MacroAssembler::InvokeFunction(JSFunction* function,
const ParameterCount& actual,
InvokeFlag flag) {
ASSERT(function->is_compiled());
// Get the function and setup the context.
Move(rdi, Handle<JSFunction>(function));
movq(rsi, FieldOperand(rdi, JSFunction::kContextOffset));
// Invoke the cached code.
Handle<Code> code(function->code());
ParameterCount expected(function->shared()->formal_parameter_count());
InvokeCode(code, expected, actual, RelocInfo::CODE_TARGET, flag);
}
void MacroAssembler::EnterFrame(StackFrame::Type type) {
push(rbp);
movq(rbp, rsp);
push(rsi); // Context.
Push(Smi::FromInt(type));
movq(kScratchRegister, CodeObject(), RelocInfo::EMBEDDED_OBJECT);
push(kScratchRegister);
if (FLAG_debug_code) {
movq(kScratchRegister,
Factory::undefined_value(),
RelocInfo::EMBEDDED_OBJECT);
cmpq(Operand(rsp, 0), kScratchRegister);
Check(not_equal, "code object not properly patched");
}
}
void MacroAssembler::LeaveFrame(StackFrame::Type type) {
if (FLAG_debug_code) {
Move(kScratchRegister, Smi::FromInt(type));
cmpq(Operand(rbp, StandardFrameConstants::kMarkerOffset), kScratchRegister);
Check(equal, "stack frame types must match");
}
movq(rsp, rbp);
pop(rbp);
}
void MacroAssembler::EnterExitFramePrologue(bool save_rax) {
// Setup the frame structure on the stack.
// All constants are relative to the frame pointer of the exit frame.
ASSERT(ExitFrameConstants::kCallerSPDisplacement == +2 * kPointerSize);
ASSERT(ExitFrameConstants::kCallerPCOffset == +1 * kPointerSize);
ASSERT(ExitFrameConstants::kCallerFPOffset == 0 * kPointerSize);
push(rbp);
movq(rbp, rsp);
// Reserve room for entry stack pointer and push the code object.
ASSERT(ExitFrameConstants::kSPOffset == -1 * kPointerSize);
push(Immediate(0)); // Saved entry sp, patched before call.
movq(kScratchRegister, CodeObject(), RelocInfo::EMBEDDED_OBJECT);
push(kScratchRegister); // Accessed from EditFrame::code_slot.
// Save the frame pointer and the context in top.
ExternalReference c_entry_fp_address(Top::k_c_entry_fp_address);
ExternalReference context_address(Top::k_context_address);
if (save_rax) {
movq(r14, rax); // Backup rax before we use it.
}
movq(rax, rbp);
store_rax(c_entry_fp_address);
movq(rax, rsi);
store_rax(context_address);
}
void MacroAssembler::EnterExitFrameEpilogue(int result_size,
int argc) {
#ifdef _WIN64
// Reserve space on stack for result and argument structures, if necessary.
int result_stack_space = (result_size < 2) ? 0 : result_size * kPointerSize;
// Reserve space for the Arguments object. The Windows 64-bit ABI
// requires us to pass this structure as a pointer to its location on
// the stack. The structure contains 2 values.
int argument_stack_space = argc * kPointerSize;
// We also need backing space for 4 parameters, even though
// we only pass one or two parameter, and it is in a register.
int argument_mirror_space = 4 * kPointerSize;
int total_stack_space =
argument_mirror_space + argument_stack_space + result_stack_space;
subq(rsp, Immediate(total_stack_space));
#endif
// Get the required frame alignment for the OS.
static const int kFrameAlignment = OS::ActivationFrameAlignment();
if (kFrameAlignment > 0) {
ASSERT(IsPowerOf2(kFrameAlignment));
movq(kScratchRegister, Immediate(-kFrameAlignment));
and_(rsp, kScratchRegister);
}
// Patch the saved entry sp.
movq(Operand(rbp, ExitFrameConstants::kSPOffset), rsp);
}
void MacroAssembler::EnterExitFrame(int result_size) {
EnterExitFramePrologue(true);
// Setup argv in callee-saved register r12. It is reused in LeaveExitFrame,
// so it must be retained across the C-call.
int offset = StandardFrameConstants::kCallerSPOffset - kPointerSize;
lea(r12, Operand(rbp, r14, times_pointer_size, offset));
EnterExitFrameEpilogue(result_size, 2);
}
void MacroAssembler::EnterApiExitFrame(int stack_space,
int argc,
int result_size) {
EnterExitFramePrologue(false);
// Setup argv in callee-saved register r12. It is reused in LeaveExitFrame,
// so it must be retained across the C-call.
int offset = StandardFrameConstants::kCallerSPOffset - kPointerSize;
lea(r12, Operand(rbp, (stack_space * kPointerSize) + offset));
EnterExitFrameEpilogue(result_size, argc);
}
void MacroAssembler::LeaveExitFrame(int result_size) {
// Registers:
// r12 : argv
// Get the return address from the stack and restore the frame pointer.
movq(rcx, Operand(rbp, 1 * kPointerSize));
movq(rbp, Operand(rbp, 0 * kPointerSize));
// Pop everything up to and including the arguments and the receiver
// from the caller stack.
lea(rsp, Operand(r12, 1 * kPointerSize));
// Restore current context from top and clear it in debug mode.
ExternalReference context_address(Top::k_context_address);
movq(kScratchRegister, context_address);
movq(rsi, Operand(kScratchRegister, 0));
#ifdef DEBUG
movq(Operand(kScratchRegister, 0), Immediate(0));
#endif
// Push the return address to get ready to return.
push(rcx);
// Clear the top frame.
ExternalReference c_entry_fp_address(Top::k_c_entry_fp_address);
movq(kScratchRegister, c_entry_fp_address);
movq(Operand(kScratchRegister, 0), Immediate(0));
}
void MacroAssembler::CheckAccessGlobalProxy(Register holder_reg,
Register scratch,
Label* miss) {
Label same_contexts;
ASSERT(!holder_reg.is(scratch));
ASSERT(!scratch.is(kScratchRegister));
// Load current lexical context from the stack frame.
movq(scratch, Operand(rbp, StandardFrameConstants::kContextOffset));
// When generating debug code, make sure the lexical context is set.
if (FLAG_debug_code) {
cmpq(scratch, Immediate(0));
Check(not_equal, "we should not have an empty lexical context");
}
// Load the global context of the current context.
int offset = Context::kHeaderSize + Context::GLOBAL_INDEX * kPointerSize;
movq(scratch, FieldOperand(scratch, offset));
movq(scratch, FieldOperand(scratch, GlobalObject::kGlobalContextOffset));
// Check the context is a global context.
if (FLAG_debug_code) {
Cmp(FieldOperand(scratch, HeapObject::kMapOffset),
Factory::global_context_map());
Check(equal, "JSGlobalObject::global_context should be a global context.");
}
// Check if both contexts are the same.
cmpq(scratch, FieldOperand(holder_reg, JSGlobalProxy::kContextOffset));
j(equal, &same_contexts);
// Compare security tokens.
// Check that the security token in the calling global object is
// compatible with the security token in the receiving global
// object.
// Check the context is a global context.
if (FLAG_debug_code) {
// Preserve original value of holder_reg.
push(holder_reg);
movq(holder_reg, FieldOperand(holder_reg, JSGlobalProxy::kContextOffset));
CompareRoot(holder_reg, Heap::kNullValueRootIndex);
Check(not_equal, "JSGlobalProxy::context() should not be null.");
// Read the first word and compare to global_context_map(),
movq(holder_reg, FieldOperand(holder_reg, HeapObject::kMapOffset));
CompareRoot(holder_reg, Heap::kGlobalContextMapRootIndex);
Check(equal, "JSGlobalObject::global_context should be a global context.");
pop(holder_reg);
}
movq(kScratchRegister,
FieldOperand(holder_reg, JSGlobalProxy::kContextOffset));
int token_offset =
Context::kHeaderSize + Context::SECURITY_TOKEN_INDEX * kPointerSize;
movq(scratch, FieldOperand(scratch, token_offset));
cmpq(scratch, FieldOperand(kScratchRegister, token_offset));
j(not_equal, miss);
bind(&same_contexts);
}
void MacroAssembler::LoadAllocationTopHelper(Register result,
Register result_end,
Register scratch,
AllocationFlags flags) {
ExternalReference new_space_allocation_top =
ExternalReference::new_space_allocation_top_address();
// Just return if allocation top is already known.
if ((flags & RESULT_CONTAINS_TOP) != 0) {
// No use of scratch if allocation top is provided.
ASSERT(!scratch.is_valid());
#ifdef DEBUG
// Assert that result actually contains top on entry.
movq(kScratchRegister, new_space_allocation_top);
cmpq(result, Operand(kScratchRegister, 0));
Check(equal, "Unexpected allocation top");
#endif
return;
}
// Move address of new object to result. Use scratch register if available,
// and keep address in scratch until call to UpdateAllocationTopHelper.
if (scratch.is_valid()) {
ASSERT(!scratch.is(result_end));
movq(scratch, new_space_allocation_top);
movq(result, Operand(scratch, 0));
} else if (result.is(rax)) {
load_rax(new_space_allocation_top);
} else {
movq(kScratchRegister, new_space_allocation_top);
movq(result, Operand(kScratchRegister, 0));
}
}
void MacroAssembler::UpdateAllocationTopHelper(Register result_end,
Register scratch) {
if (FLAG_debug_code) {
testq(result_end, Immediate(kObjectAlignmentMask));
Check(zero, "Unaligned allocation in new space");
}
ExternalReference new_space_allocation_top =
ExternalReference::new_space_allocation_top_address();
// Update new top.
if (result_end.is(rax)) {
// rax can be stored directly to a memory location.
store_rax(new_space_allocation_top);
} else {
// Register required - use scratch provided if available.
if (scratch.is_valid()) {
movq(Operand(scratch, 0), result_end);
} else {
movq(kScratchRegister, new_space_allocation_top);
movq(Operand(kScratchRegister, 0), result_end);
}
}
}
void MacroAssembler::AllocateInNewSpace(int object_size,
Register result,
Register result_end,
Register scratch,
Label* gc_required,
AllocationFlags flags) {
ASSERT(!result.is(result_end));
// Load address of new object into result.
LoadAllocationTopHelper(result, result_end, scratch, flags);
// Calculate new top and bail out if new space is exhausted.
ExternalReference new_space_allocation_limit =
ExternalReference::new_space_allocation_limit_address();
Register top_reg = result_end.is_valid() ? result_end : result;
if (top_reg.is(result)) {
addq(top_reg, Immediate(object_size));
} else {
lea(top_reg, Operand(result, object_size));
}
movq(kScratchRegister, new_space_allocation_limit);
cmpq(top_reg, Operand(kScratchRegister, 0));
j(above, gc_required);
// Update allocation top.
UpdateAllocationTopHelper(top_reg, scratch);
if (top_reg.is(result)) {
if ((flags & TAG_OBJECT) != 0) {
subq(result, Immediate(object_size - kHeapObjectTag));
} else {
subq(result, Immediate(object_size));
}
} else if ((flags & TAG_OBJECT) != 0) {
// Tag the result if requested.
addq(result, Immediate(kHeapObjectTag));
}
}
void MacroAssembler::AllocateInNewSpace(int header_size,
ScaleFactor element_size,
Register element_count,
Register result,
Register result_end,
Register scratch,
Label* gc_required,
AllocationFlags flags) {
ASSERT(!result.is(result_end));
// Load address of new object into result.
LoadAllocationTopHelper(result, result_end, scratch, flags);
// Calculate new top and bail out if new space is exhausted.
ExternalReference new_space_allocation_limit =
ExternalReference::new_space_allocation_limit_address();
lea(result_end, Operand(result, element_count, element_size, header_size));
movq(kScratchRegister, new_space_allocation_limit);
cmpq(result_end, Operand(kScratchRegister, 0));
j(above, gc_required);
// Update allocation top.
UpdateAllocationTopHelper(result_end, scratch);
// Tag the result if requested.
if ((flags & TAG_OBJECT) != 0) {
addq(result, Immediate(kHeapObjectTag));
}
}
void MacroAssembler::AllocateInNewSpace(Register object_size,
Register result,
Register result_end,
Register scratch,
Label* gc_required,
AllocationFlags flags) {
// Load address of new object into result.
LoadAllocationTopHelper(result, result_end, scratch, flags);
// Calculate new top and bail out if new space is exhausted.
ExternalReference new_space_allocation_limit =
ExternalReference::new_space_allocation_limit_address();
if (!object_size.is(result_end)) {
movq(result_end, object_size);
}
addq(result_end, result);
movq(kScratchRegister, new_space_allocation_limit);
cmpq(result_end, Operand(kScratchRegister, 0));
j(above, gc_required);
// Update allocation top.
UpdateAllocationTopHelper(result_end, scratch);
// Tag the result if requested.
if ((flags & TAG_OBJECT) != 0) {
addq(result, Immediate(kHeapObjectTag));
}
}
void MacroAssembler::UndoAllocationInNewSpace(Register object) {
ExternalReference new_space_allocation_top =
ExternalReference::new_space_allocation_top_address();
// Make sure the object has no tag before resetting top.
and_(object, Immediate(~kHeapObjectTagMask));
movq(kScratchRegister, new_space_allocation_top);
#ifdef DEBUG
cmpq(object, Operand(kScratchRegister, 0));
Check(below, "Undo allocation of non allocated memory");
#endif
movq(Operand(kScratchRegister, 0), object);
}
void MacroAssembler::AllocateHeapNumber(Register result,
Register scratch,
Label* gc_required) {
// Allocate heap number in new space.
AllocateInNewSpace(HeapNumber::kSize,
result,
scratch,
no_reg,
gc_required,
TAG_OBJECT);
// Set the map.
LoadRoot(kScratchRegister, Heap::kHeapNumberMapRootIndex);
movq(FieldOperand(result, HeapObject::kMapOffset), kScratchRegister);
}
void MacroAssembler::AllocateTwoByteString(Register result,
Register length,
Register scratch1,
Register scratch2,
Register scratch3,
Label* gc_required) {
// Calculate the number of bytes needed for the characters in the string while
// observing object alignment.
const int kHeaderAlignment = SeqTwoByteString::kHeaderSize &
kObjectAlignmentMask;
ASSERT(kShortSize == 2);
// scratch1 = length * 2 + kObjectAlignmentMask.
lea(scratch1, Operand(length, length, times_1, kObjectAlignmentMask +
kHeaderAlignment));
and_(scratch1, Immediate(~kObjectAlignmentMask));
if (kHeaderAlignment > 0) {
subq(scratch1, Immediate(kHeaderAlignment));
}
// Allocate two byte string in new space.
AllocateInNewSpace(SeqTwoByteString::kHeaderSize,
times_1,
scratch1,
result,
scratch2,
scratch3,
gc_required,
TAG_OBJECT);
// Set the map, length and hash field.
LoadRoot(kScratchRegister, Heap::kStringMapRootIndex);
movq(FieldOperand(result, HeapObject::kMapOffset), kScratchRegister);
Integer32ToSmi(scratch1, length);
movq(FieldOperand(result, String::kLengthOffset), scratch1);
movq(FieldOperand(result, String::kHashFieldOffset),
Immediate(String::kEmptyHashField));
}
void MacroAssembler::AllocateAsciiString(Register result,
Register length,
Register scratch1,
Register scratch2,
Register scratch3,
Label* gc_required) {
// Calculate the number of bytes needed for the characters in the string while
// observing object alignment.
const int kHeaderAlignment = SeqAsciiString::kHeaderSize &
kObjectAlignmentMask;
movl(scratch1, length);
ASSERT(kCharSize == 1);
addq(scratch1, Immediate(kObjectAlignmentMask + kHeaderAlignment));
and_(scratch1, Immediate(~kObjectAlignmentMask));
if (kHeaderAlignment > 0) {
subq(scratch1, Immediate(kHeaderAlignment));
}
// Allocate ascii string in new space.
AllocateInNewSpace(SeqAsciiString::kHeaderSize,
times_1,
scratch1,
result,
scratch2,
scratch3,
gc_required,
TAG_OBJECT);
// Set the map, length and hash field.
LoadRoot(kScratchRegister, Heap::kAsciiStringMapRootIndex);
movq(FieldOperand(result, HeapObject::kMapOffset), kScratchRegister);
Integer32ToSmi(scratch1, length);
movq(FieldOperand(result, String::kLengthOffset), scratch1);
movq(FieldOperand(result, String::kHashFieldOffset),
Immediate(String::kEmptyHashField));
}
void MacroAssembler::AllocateConsString(Register result,
Register scratch1,
Register scratch2,
Label* gc_required) {
// Allocate heap number in new space.
AllocateInNewSpace(ConsString::kSize,
result,
scratch1,
scratch2,
gc_required,
TAG_OBJECT);
// Set the map. The other fields are left uninitialized.
LoadRoot(kScratchRegister, Heap::kConsStringMapRootIndex);
movq(FieldOperand(result, HeapObject::kMapOffset), kScratchRegister);
}
void MacroAssembler::AllocateAsciiConsString(Register result,
Register scratch1,
Register scratch2,
Label* gc_required) {
// Allocate heap number in new space.
AllocateInNewSpace(ConsString::kSize,
result,
scratch1,
scratch2,
gc_required,
TAG_OBJECT);
// Set the map. The other fields are left uninitialized.
LoadRoot(kScratchRegister, Heap::kConsAsciiStringMapRootIndex);
movq(FieldOperand(result, HeapObject::kMapOffset), kScratchRegister);
}
void MacroAssembler::LoadContext(Register dst, int context_chain_length) {
if (context_chain_length > 0) {
// Move up the chain of contexts to the context containing the slot.
movq(dst, Operand(rsi, Context::SlotOffset(Context::CLOSURE_INDEX)));
// Load the function context (which is the incoming, outer context).
movq(dst, FieldOperand(dst, JSFunction::kContextOffset));
for (int i = 1; i < context_chain_length; i++) {
movq(dst, Operand(dst, Context::SlotOffset(Context::CLOSURE_INDEX)));
movq(dst, FieldOperand(dst, JSFunction::kContextOffset));
}
// The context may be an intermediate context, not a function context.
movq(dst, Operand(dst, Context::SlotOffset(Context::FCONTEXT_INDEX)));
} else { // context is the current function context.
// The context may be an intermediate context, not a function context.
movq(dst, Operand(rsi, Context::SlotOffset(Context::FCONTEXT_INDEX)));
}
}
int MacroAssembler::ArgumentStackSlotsForCFunctionCall(int num_arguments) {
// On Windows 64 stack slots are reserved by the caller for all arguments
// including the ones passed in registers, and space is always allocated for
// the four register arguments even if the function takes fewer than four
// arguments.
// On AMD64 ABI (Linux/Mac) the first six arguments are passed in registers
// and the caller does not reserve stack slots for them.
ASSERT(num_arguments >= 0);
#ifdef _WIN64
static const int kMinimumStackSlots = 4;
if (num_arguments < kMinimumStackSlots) return kMinimumStackSlots;
return num_arguments;
#else
static const int kRegisterPassedArguments = 6;
if (num_arguments < kRegisterPassedArguments) return 0;
return num_arguments - kRegisterPassedArguments;
#endif
}
void MacroAssembler::PrepareCallCFunction(int num_arguments) {
int frame_alignment = OS::ActivationFrameAlignment();
ASSERT(frame_alignment != 0);
ASSERT(num_arguments >= 0);
// Make stack end at alignment and allocate space for arguments and old rsp.
movq(kScratchRegister, rsp);
ASSERT(IsPowerOf2(frame_alignment));
int argument_slots_on_stack =
ArgumentStackSlotsForCFunctionCall(num_arguments);
subq(rsp, Immediate((argument_slots_on_stack + 1) * kPointerSize));
and_(rsp, Immediate(-frame_alignment));
movq(Operand(rsp, argument_slots_on_stack * kPointerSize), kScratchRegister);
}
void MacroAssembler::CallCFunction(ExternalReference function,
int num_arguments) {
movq(rax, function);
CallCFunction(rax, num_arguments);
}
void MacroAssembler::CallCFunction(Register function, int num_arguments) {
// Check stack alignment.
if (FLAG_debug_code) {
CheckStackAlignment();
}
call(function);
ASSERT(OS::ActivationFrameAlignment() != 0);
ASSERT(num_arguments >= 0);
int argument_slots_on_stack =
ArgumentStackSlotsForCFunctionCall(num_arguments);
movq(rsp, Operand(rsp, argument_slots_on_stack * kPointerSize));
}
CodePatcher::CodePatcher(byte* address, int size)
: address_(address), size_(size), masm_(address, size + Assembler::kGap) {
// Create a new macro assembler pointing to the address of the code to patch.
// The size is adjusted with kGap on order for the assembler to generate size
// bytes of instructions without failing with buffer size constraints.
ASSERT(masm_.reloc_info_writer.pos() == address_ + size_ + Assembler::kGap);
}
CodePatcher::~CodePatcher() {
// Indicate that code has changed.
CPU::FlushICache(address_, size_);
// Check that the code was patched as expected.
ASSERT(masm_.pc_ == address_ + size_);
ASSERT(masm_.reloc_info_writer.pos() == address_ + size_ + Assembler::kGap);
}
} } // namespace v8::internal
#endif // V8_TARGET_ARCH_X64