| /* | 
 |  * Copyright (C) 2008 Apple Inc. All rights reserved. | 
 |  * | 
 |  * Redistribution and use in source and binary forms, with or without | 
 |  * modification, are permitted provided that the following conditions | 
 |  * are met: | 
 |  * 1. Redistributions of source code must retain the above copyright | 
 |  *    notice, this list of conditions and the following disclaimer. | 
 |  * 2. 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. | 
 |  * | 
 |  * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``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 APPLE INC. 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 "config.h" | 
 |  | 
 | #if ENABLE(JIT) | 
 | #include "JIT.h" | 
 |  | 
 | #include "CodeBlock.h" | 
 | #include "JITInlineMethods.h" | 
 | #include "JITStubCall.h" | 
 | #include "JITStubs.h" | 
 | #include "JSArray.h" | 
 | #include "JSFunction.h" | 
 | #include "Interpreter.h" | 
 | #include "ResultType.h" | 
 | #include "SamplingTool.h" | 
 |  | 
 | #ifndef NDEBUG | 
 | #include <stdio.h> | 
 | #endif | 
 |  | 
 | using namespace std; | 
 |  | 
 | namespace JSC { | 
 |  | 
 | void JIT::emit_op_jless(Instruction* currentInstruction) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJump(op_jless, op1, op2, target, LessThan); | 
 | } | 
 |  | 
 | void JIT::emit_op_jlesseq(Instruction* currentInstruction) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJump(op_jlesseq, op1, op2, target, LessThanOrEqual); | 
 | } | 
 |  | 
 | void JIT::emit_op_jgreater(Instruction* currentInstruction) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJump(op_jgreater, op1, op2, target, GreaterThan); | 
 | } | 
 |  | 
 | void JIT::emit_op_jgreatereq(Instruction* currentInstruction) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJump(op_jgreatereq, op1, op2, target, GreaterThanOrEqual); | 
 | } | 
 |  | 
 | void JIT::emit_op_jnless(Instruction* currentInstruction) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJump(op_jnless, op1, op2, target, GreaterThanOrEqual); | 
 | } | 
 |  | 
 | void JIT::emit_op_jnlesseq(Instruction* currentInstruction) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJump(op_jnlesseq, op1, op2, target, GreaterThan); | 
 | } | 
 |  | 
 | void JIT::emit_op_jngreater(Instruction* currentInstruction) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJump(op_jngreater, op1, op2, target, LessThanOrEqual); | 
 | } | 
 |  | 
 | void JIT::emit_op_jngreatereq(Instruction* currentInstruction) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJump(op_jngreatereq, op1, op2, target, LessThan); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_jless(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJumpSlow(op1, op2, target, DoubleLessThan, cti_op_jless, false, iter); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_jlesseq(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJumpSlow(op1, op2, target, DoubleLessThanOrEqual, cti_op_jlesseq, false, iter); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_jgreater(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJumpSlow(op1, op2, target, DoubleGreaterThan, cti_op_jgreater, false, iter); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_jgreatereq(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJumpSlow(op1, op2, target, DoubleGreaterThanOrEqual, cti_op_jgreatereq, false, iter); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_jnless(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJumpSlow(op1, op2, target, DoubleGreaterThanOrEqualOrUnordered, cti_op_jless, true, iter); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_jnlesseq(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJumpSlow(op1, op2, target, DoubleGreaterThanOrUnordered, cti_op_jlesseq, true, iter); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_jngreater(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJumpSlow(op1, op2, target, DoubleLessThanOrEqualOrUnordered, cti_op_jgreater, true, iter); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_jngreatereq(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned op1 = currentInstruction[1].u.operand; | 
 |     unsigned op2 = currentInstruction[2].u.operand; | 
 |     unsigned target = currentInstruction[3].u.operand; | 
 |  | 
 |     emit_compareAndJumpSlow(op1, op2, target, DoubleLessThanOrUnordered, cti_op_jgreatereq, true, iter); | 
 | } | 
 |  | 
 | #if USE(JSVALUE64) | 
 |  | 
 | void JIT::emit_op_lshift(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |  | 
 |     emitGetVirtualRegisters(op1, regT0, op2, regT2); | 
 |     // FIXME: would we be better using 'emitJumpSlowCaseIfNotImmediateIntegers'? - we *probably* ought to be consistent. | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT2); | 
 |     emitFastArithImmToInt(regT0); | 
 |     emitFastArithImmToInt(regT2); | 
 |     lshift32(regT2, regT0); | 
 |     emitFastArithReTagImmediate(regT0, regT0); | 
 |     emitPutVirtualRegister(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_lshift(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |  | 
 |     UNUSED_PARAM(op1); | 
 |     UNUSED_PARAM(op2); | 
 |     linkSlowCase(iter); | 
 |     linkSlowCase(iter); | 
 |     JITStubCall stubCall(this, cti_op_lshift); | 
 |     stubCall.addArgument(regT0); | 
 |     stubCall.addArgument(regT2); | 
 |     stubCall.call(result); | 
 | } | 
 |  | 
 | void JIT::emit_op_rshift(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |  | 
 |     if (isOperandConstantImmediateInt(op2)) { | 
 |         // isOperandConstantImmediateInt(op2) => 1 SlowCase | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         // Mask with 0x1f as per ecma-262 11.7.2 step 7. | 
 |         rshift32(Imm32(getConstantOperandImmediateInt(op2) & 0x1f), regT0); | 
 |     } else { | 
 |         emitGetVirtualRegisters(op1, regT0, op2, regT2); | 
 |         if (supportsFloatingPointTruncate()) { | 
 |             Jump lhsIsInt = emitJumpIfImmediateInteger(regT0); | 
 |             // supportsFloatingPoint() && USE(JSVALUE64) => 3 SlowCases | 
 |             addSlowCase(emitJumpIfNotImmediateNumber(regT0)); | 
 |             addPtr(tagTypeNumberRegister, regT0); | 
 |             movePtrToDouble(regT0, fpRegT0); | 
 |             addSlowCase(branchTruncateDoubleToInt32(fpRegT0, regT0)); | 
 |             lhsIsInt.link(this); | 
 |             emitJumpSlowCaseIfNotImmediateInteger(regT2); | 
 |         } else { | 
 |             // !supportsFloatingPoint() => 2 SlowCases | 
 |             emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |             emitJumpSlowCaseIfNotImmediateInteger(regT2); | 
 |         } | 
 |         emitFastArithImmToInt(regT2); | 
 |         rshift32(regT2, regT0); | 
 |     } | 
 |     emitFastArithIntToImmNoCheck(regT0, regT0); | 
 |     emitPutVirtualRegister(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_rshift(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |  | 
 |     JITStubCall stubCall(this, cti_op_rshift); | 
 |  | 
 |     if (isOperandConstantImmediateInt(op2)) { | 
 |         linkSlowCase(iter); | 
 |         stubCall.addArgument(regT0); | 
 |         stubCall.addArgument(op2, regT2); | 
 |     } else { | 
 |         if (supportsFloatingPointTruncate()) { | 
 |             linkSlowCase(iter); | 
 |             linkSlowCase(iter); | 
 |             linkSlowCase(iter); | 
 |             // We're reloading op1 to regT0 as we can no longer guarantee that | 
 |             // we have not munged the operand.  It may have already been shifted | 
 |             // correctly, but it still will not have been tagged. | 
 |             stubCall.addArgument(op1, regT0); | 
 |             stubCall.addArgument(regT2); | 
 |         } else { | 
 |             linkSlowCase(iter); | 
 |             linkSlowCase(iter); | 
 |             stubCall.addArgument(regT0); | 
 |             stubCall.addArgument(regT2); | 
 |         } | 
 |     } | 
 |  | 
 |     stubCall.call(result); | 
 | } | 
 |  | 
 | void JIT::emit_op_urshift(Instruction* currentInstruction) | 
 | { | 
 |     unsigned dst = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |  | 
 |     // Slow case of urshift makes assumptions about what registers hold the | 
 |     // shift arguments, so any changes must be updated there as well. | 
 |     if (isOperandConstantImmediateInt(op2)) { | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         emitFastArithImmToInt(regT0); | 
 |         int shift = getConstantOperand(op2).asInt32(); | 
 |         if (shift) | 
 |             urshift32(Imm32(shift & 0x1f), regT0); | 
 |         // unsigned shift < 0 or shift = k*2^32 may result in (essentially) | 
 |         // a toUint conversion, which can result in a value we can represent | 
 |         // as an immediate int. | 
 |         if (shift < 0 || !(shift & 31)) | 
 |             addSlowCase(branch32(LessThan, regT0, TrustedImm32(0))); | 
 |         emitFastArithReTagImmediate(regT0, regT0); | 
 |         emitPutVirtualRegister(dst, regT0); | 
 |         return; | 
 |     } | 
 |     emitGetVirtualRegisters(op1, regT0, op2, regT1); | 
 |     if (!isOperandConstantImmediateInt(op1)) | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT1); | 
 |     emitFastArithImmToInt(regT0); | 
 |     emitFastArithImmToInt(regT1); | 
 |     urshift32(regT1, regT0); | 
 |     addSlowCase(branch32(LessThan, regT0, TrustedImm32(0))); | 
 |     emitFastArithReTagImmediate(regT0, regT0); | 
 |     emitPutVirtualRegister(dst, regT0); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_urshift(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned dst = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     if (isOperandConstantImmediateInt(op2)) { | 
 |         int shift = getConstantOperand(op2).asInt32(); | 
 |         // op1 = regT0 | 
 |         linkSlowCase(iter); // int32 check | 
 |         if (supportsFloatingPointTruncate()) { | 
 |             JumpList failures; | 
 |             failures.append(emitJumpIfNotImmediateNumber(regT0)); // op1 is not a double | 
 |             addPtr(tagTypeNumberRegister, regT0); | 
 |             movePtrToDouble(regT0, fpRegT0); | 
 |             failures.append(branchTruncateDoubleToInt32(fpRegT0, regT0)); | 
 |             if (shift) | 
 |                 urshift32(Imm32(shift & 0x1f), regT0); | 
 |             if (shift < 0 || !(shift & 31)) | 
 |                 failures.append(branch32(LessThan, regT0, TrustedImm32(0))); | 
 |             emitFastArithReTagImmediate(regT0, regT0); | 
 |             emitPutVirtualRegister(dst, regT0); | 
 |             emitJumpSlowToHot(jump(), OPCODE_LENGTH(op_rshift)); | 
 |             failures.link(this); | 
 |         } | 
 |         if (shift < 0 || !(shift & 31)) | 
 |             linkSlowCase(iter); // failed to box in hot path | 
 |     } else { | 
 |         // op1 = regT0 | 
 |         // op2 = regT1 | 
 |         if (!isOperandConstantImmediateInt(op1)) { | 
 |             linkSlowCase(iter); // int32 check -- op1 is not an int | 
 |             if (supportsFloatingPointTruncate()) { | 
 |                 JumpList failures; | 
 |                 failures.append(emitJumpIfNotImmediateNumber(regT0)); // op1 is not a double | 
 |                 addPtr(tagTypeNumberRegister, regT0); | 
 |                 movePtrToDouble(regT0, fpRegT0); | 
 |                 failures.append(branchTruncateDoubleToInt32(fpRegT0, regT0)); | 
 |                 failures.append(emitJumpIfNotImmediateInteger(regT1)); // op2 is not an int | 
 |                 emitFastArithImmToInt(regT1); | 
 |                 urshift32(regT1, regT0); | 
 |                 failures.append(branch32(LessThan, regT0, TrustedImm32(0))); | 
 |                 emitFastArithReTagImmediate(regT0, regT0); | 
 |                 emitPutVirtualRegister(dst, regT0); | 
 |                 emitJumpSlowToHot(jump(), OPCODE_LENGTH(op_rshift)); | 
 |                 failures.link(this); | 
 |             } | 
 |         } | 
 |          | 
 |         linkSlowCase(iter); // int32 check - op2 is not an int | 
 |         linkSlowCase(iter); // Can't represent unsigned result as an immediate | 
 |     } | 
 |      | 
 |     JITStubCall stubCall(this, cti_op_urshift); | 
 |     stubCall.addArgument(op1, regT0); | 
 |     stubCall.addArgument(op2, regT1); | 
 |     stubCall.call(dst); | 
 | } | 
 |  | 
 | void JIT::emit_compareAndJump(OpcodeID, unsigned op1, unsigned op2, unsigned target, RelationalCondition condition) | 
 | { | 
 |     // We generate inline code for the following cases in the fast path: | 
 |     // - int immediate to constant int immediate | 
 |     // - constant int immediate to int immediate | 
 |     // - int immediate to int immediate | 
 |  | 
 |     if (isOperandConstantImmediateChar(op1)) { | 
 |         emitGetVirtualRegister(op2, regT0); | 
 |         addSlowCase(emitJumpIfNotJSCell(regT0)); | 
 |         JumpList failures; | 
 |         emitLoadCharacterString(regT0, regT0, failures); | 
 |         addSlowCase(failures); | 
 |         addJump(branch32(commute(condition), regT0, Imm32(asString(getConstantOperand(op1))->tryGetValue()[0])), target); | 
 |         return; | 
 |     } | 
 |     if (isOperandConstantImmediateChar(op2)) { | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         addSlowCase(emitJumpIfNotJSCell(regT0)); | 
 |         JumpList failures; | 
 |         emitLoadCharacterString(regT0, regT0, failures); | 
 |         addSlowCase(failures); | 
 |         addJump(branch32(condition, regT0, Imm32(asString(getConstantOperand(op2))->tryGetValue()[0])), target); | 
 |         return; | 
 |     } | 
 |     if (isOperandConstantImmediateInt(op2)) { | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         int32_t op2imm = getConstantOperandImmediateInt(op2); | 
 |         addJump(branch32(condition, regT0, Imm32(op2imm)), target); | 
 |     } else if (isOperandConstantImmediateInt(op1)) { | 
 |         emitGetVirtualRegister(op2, regT1); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT1); | 
 |         int32_t op1imm = getConstantOperandImmediateInt(op1); | 
 |         addJump(branch32(commute(condition), regT1, Imm32(op1imm)), target); | 
 |     } else { | 
 |         emitGetVirtualRegisters(op1, regT0, op2, regT1); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT1); | 
 |  | 
 |         addJump(branch32(condition, regT0, regT1), target); | 
 |     } | 
 | } | 
 |  | 
 | void JIT::emit_compareAndJumpSlow(unsigned op1, unsigned op2, unsigned target, DoubleCondition condition, int (JIT_STUB *stub)(STUB_ARGS_DECLARATION), bool invert, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     COMPILE_ASSERT(OPCODE_LENGTH(op_jless) == OPCODE_LENGTH(op_jlesseq), OPCODE_LENGTH_op_jlesseq_equals_op_jless); | 
 |     COMPILE_ASSERT(OPCODE_LENGTH(op_jless) == OPCODE_LENGTH(op_jnless), OPCODE_LENGTH_op_jnless_equals_op_jless); | 
 |     COMPILE_ASSERT(OPCODE_LENGTH(op_jless) == OPCODE_LENGTH(op_jnlesseq), OPCODE_LENGTH_op_jnlesseq_equals_op_jless); | 
 |     COMPILE_ASSERT(OPCODE_LENGTH(op_jless) == OPCODE_LENGTH(op_jgreater), OPCODE_LENGTH_op_jgreater_equals_op_jless); | 
 |     COMPILE_ASSERT(OPCODE_LENGTH(op_jless) == OPCODE_LENGTH(op_jgreatereq), OPCODE_LENGTH_op_jgreatereq_equals_op_jless); | 
 |     COMPILE_ASSERT(OPCODE_LENGTH(op_jless) == OPCODE_LENGTH(op_jngreater), OPCODE_LENGTH_op_jngreater_equals_op_jless); | 
 |     COMPILE_ASSERT(OPCODE_LENGTH(op_jless) == OPCODE_LENGTH(op_jngreatereq), OPCODE_LENGTH_op_jngreatereq_equals_op_jless); | 
 |      | 
 |     // We generate inline code for the following cases in the slow path: | 
 |     // - floating-point number to constant int immediate | 
 |     // - constant int immediate to floating-point number | 
 |     // - floating-point number to floating-point number. | 
 |     if (isOperandConstantImmediateChar(op1) || isOperandConstantImmediateChar(op2)) { | 
 |         linkSlowCase(iter); | 
 |         linkSlowCase(iter); | 
 |         linkSlowCase(iter); | 
 |         linkSlowCase(iter); | 
 |         JITStubCall stubCall(this, stub); | 
 |         stubCall.addArgument(op1, regT0); | 
 |         stubCall.addArgument(op2, regT1); | 
 |         stubCall.call(); | 
 |         emitJumpSlowToHot(branchTest32(invert ? Zero : NonZero, regT0), target); | 
 |         return; | 
 |     } | 
 |  | 
 |     if (isOperandConstantImmediateInt(op2)) { | 
 |         linkSlowCase(iter); | 
 |  | 
 |         if (supportsFloatingPoint()) { | 
 |             Jump fail1 = emitJumpIfNotImmediateNumber(regT0); | 
 |             addPtr(tagTypeNumberRegister, regT0); | 
 |             movePtrToDouble(regT0, fpRegT0); | 
 |  | 
 |             int32_t op2imm = getConstantOperand(op2).asInt32(); | 
 |  | 
 |             move(Imm32(op2imm), regT1); | 
 |             convertInt32ToDouble(regT1, fpRegT1); | 
 |  | 
 |             emitJumpSlowToHot(branchDouble(condition, fpRegT0, fpRegT1), target); | 
 |  | 
 |             emitJumpSlowToHot(jump(), OPCODE_LENGTH(op_jless)); | 
 |  | 
 |             fail1.link(this); | 
 |         } | 
 |  | 
 |         JITStubCall stubCall(this, stub); | 
 |         stubCall.addArgument(regT0); | 
 |         stubCall.addArgument(op2, regT2); | 
 |         stubCall.call(); | 
 |         emitJumpSlowToHot(branchTest32(invert ? Zero : NonZero, regT0), target); | 
 |  | 
 |     } else if (isOperandConstantImmediateInt(op1)) { | 
 |         linkSlowCase(iter); | 
 |  | 
 |         if (supportsFloatingPoint()) { | 
 |             Jump fail1 = emitJumpIfNotImmediateNumber(regT1); | 
 |             addPtr(tagTypeNumberRegister, regT1); | 
 |             movePtrToDouble(regT1, fpRegT1); | 
 |  | 
 |             int32_t op1imm = getConstantOperand(op1).asInt32(); | 
 |  | 
 |             move(Imm32(op1imm), regT0); | 
 |             convertInt32ToDouble(regT0, fpRegT0); | 
 |  | 
 |             emitJumpSlowToHot(branchDouble(condition, fpRegT0, fpRegT1), target); | 
 |  | 
 |             emitJumpSlowToHot(jump(), OPCODE_LENGTH(op_jless)); | 
 |  | 
 |             fail1.link(this); | 
 |         } | 
 |  | 
 |         JITStubCall stubCall(this, stub); | 
 |         stubCall.addArgument(op1, regT2); | 
 |         stubCall.addArgument(regT1); | 
 |         stubCall.call(); | 
 |         emitJumpSlowToHot(branchTest32(invert ? Zero : NonZero, regT0), target); | 
 |     } else { | 
 |         linkSlowCase(iter); | 
 |  | 
 |         if (supportsFloatingPoint()) { | 
 |             Jump fail1 = emitJumpIfNotImmediateNumber(regT0); | 
 |             Jump fail2 = emitJumpIfNotImmediateNumber(regT1); | 
 |             Jump fail3 = emitJumpIfImmediateInteger(regT1); | 
 |             addPtr(tagTypeNumberRegister, regT0); | 
 |             addPtr(tagTypeNumberRegister, regT1); | 
 |             movePtrToDouble(regT0, fpRegT0); | 
 |             movePtrToDouble(regT1, fpRegT1); | 
 |  | 
 |             emitJumpSlowToHot(branchDouble(condition, fpRegT0, fpRegT1), target); | 
 |  | 
 |             emitJumpSlowToHot(jump(), OPCODE_LENGTH(op_jless)); | 
 |  | 
 |             fail1.link(this); | 
 |             fail2.link(this); | 
 |             fail3.link(this); | 
 |         } | 
 |  | 
 |         linkSlowCase(iter); | 
 |         JITStubCall stubCall(this, stub); | 
 |         stubCall.addArgument(regT0); | 
 |         stubCall.addArgument(regT1); | 
 |         stubCall.call(); | 
 |         emitJumpSlowToHot(branchTest32(invert ? Zero : NonZero, regT0), target); | 
 |     } | 
 | } | 
 |  | 
 | void JIT::emit_op_bitand(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |  | 
 |     if (isOperandConstantImmediateInt(op1)) { | 
 |         emitGetVirtualRegister(op2, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         int32_t imm = getConstantOperandImmediateInt(op1); | 
 |         andPtr(Imm32(imm), regT0); | 
 |         if (imm >= 0) | 
 |             emitFastArithIntToImmNoCheck(regT0, regT0); | 
 |     } else if (isOperandConstantImmediateInt(op2)) { | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         int32_t imm = getConstantOperandImmediateInt(op2); | 
 |         andPtr(Imm32(imm), regT0); | 
 |         if (imm >= 0) | 
 |             emitFastArithIntToImmNoCheck(regT0, regT0); | 
 |     } else { | 
 |         emitGetVirtualRegisters(op1, regT0, op2, regT1); | 
 |         andPtr(regT1, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |     } | 
 |     emitPutVirtualRegister(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_bitand(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |  | 
 |     linkSlowCase(iter); | 
 |     if (isOperandConstantImmediateInt(op1)) { | 
 |         JITStubCall stubCall(this, cti_op_bitand); | 
 |         stubCall.addArgument(op1, regT2); | 
 |         stubCall.addArgument(regT0); | 
 |         stubCall.call(result); | 
 |     } else if (isOperandConstantImmediateInt(op2)) { | 
 |         JITStubCall stubCall(this, cti_op_bitand); | 
 |         stubCall.addArgument(regT0); | 
 |         stubCall.addArgument(op2, regT2); | 
 |         stubCall.call(result); | 
 |     } else { | 
 |         JITStubCall stubCall(this, cti_op_bitand); | 
 |         stubCall.addArgument(op1, regT2); | 
 |         stubCall.addArgument(regT1); | 
 |         stubCall.call(result); | 
 |     } | 
 | } | 
 |  | 
 | void JIT::emit_op_post_inc(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned srcDst = currentInstruction[2].u.operand; | 
 |  | 
 |     emitGetVirtualRegister(srcDst, regT0); | 
 |     move(regT0, regT1); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |     addSlowCase(branchAdd32(Overflow, TrustedImm32(1), regT1)); | 
 |     emitFastArithIntToImmNoCheck(regT1, regT1); | 
 |     emitPutVirtualRegister(srcDst, regT1); | 
 |     emitPutVirtualRegister(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_post_inc(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned srcDst = currentInstruction[2].u.operand; | 
 |  | 
 |     linkSlowCase(iter); | 
 |     linkSlowCase(iter); | 
 |     JITStubCall stubCall(this, cti_op_post_inc); | 
 |     stubCall.addArgument(regT0); | 
 |     stubCall.addArgument(Imm32(srcDst)); | 
 |     stubCall.call(result); | 
 | } | 
 |  | 
 | void JIT::emit_op_post_dec(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned srcDst = currentInstruction[2].u.operand; | 
 |  | 
 |     emitGetVirtualRegister(srcDst, regT0); | 
 |     move(regT0, regT1); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |     addSlowCase(branchSub32(Zero, TrustedImm32(1), regT1)); | 
 |     emitFastArithIntToImmNoCheck(regT1, regT1); | 
 |     emitPutVirtualRegister(srcDst, regT1); | 
 |     emitPutVirtualRegister(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_post_dec(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned srcDst = currentInstruction[2].u.operand; | 
 |  | 
 |     linkSlowCase(iter); | 
 |     linkSlowCase(iter); | 
 |     JITStubCall stubCall(this, cti_op_post_dec); | 
 |     stubCall.addArgument(regT0); | 
 |     stubCall.addArgument(Imm32(srcDst)); | 
 |     stubCall.call(result); | 
 | } | 
 |  | 
 | void JIT::emit_op_pre_inc(Instruction* currentInstruction) | 
 | { | 
 |     unsigned srcDst = currentInstruction[1].u.operand; | 
 |  | 
 |     emitGetVirtualRegister(srcDst, regT0); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |     addSlowCase(branchAdd32(Overflow, TrustedImm32(1), regT0)); | 
 |     emitFastArithIntToImmNoCheck(regT0, regT0); | 
 |     emitPutVirtualRegister(srcDst); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_pre_inc(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned srcDst = currentInstruction[1].u.operand; | 
 |  | 
 |     Jump notImm = getSlowCase(iter); | 
 |     linkSlowCase(iter); | 
 |     emitGetVirtualRegister(srcDst, regT0); | 
 |     notImm.link(this); | 
 |     JITStubCall stubCall(this, cti_op_pre_inc); | 
 |     stubCall.addArgument(regT0); | 
 |     stubCall.call(srcDst); | 
 | } | 
 |  | 
 | void JIT::emit_op_pre_dec(Instruction* currentInstruction) | 
 | { | 
 |     unsigned srcDst = currentInstruction[1].u.operand; | 
 |  | 
 |     emitGetVirtualRegister(srcDst, regT0); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |     addSlowCase(branchSub32(Zero, TrustedImm32(1), regT0)); | 
 |     emitFastArithIntToImmNoCheck(regT0, regT0); | 
 |     emitPutVirtualRegister(srcDst); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_pre_dec(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned srcDst = currentInstruction[1].u.operand; | 
 |  | 
 |     Jump notImm = getSlowCase(iter); | 
 |     linkSlowCase(iter); | 
 |     emitGetVirtualRegister(srcDst, regT0); | 
 |     notImm.link(this); | 
 |     JITStubCall stubCall(this, cti_op_pre_dec); | 
 |     stubCall.addArgument(regT0); | 
 |     stubCall.call(srcDst); | 
 | } | 
 |  | 
 | /* ------------------------------ BEGIN: OP_MOD ------------------------------ */ | 
 |  | 
 | #if CPU(X86) || CPU(X86_64) || CPU(MIPS) | 
 |  | 
 | void JIT::emit_op_mod(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |  | 
 | #if CPU(X86) || CPU(X86_64) | 
 |     // Make sure registers are correct for x86 IDIV instructions. | 
 |     ASSERT(regT0 == X86Registers::eax); | 
 |     ASSERT(regT1 == X86Registers::edx); | 
 |     ASSERT(regT2 == X86Registers::ecx); | 
 | #endif | 
 |  | 
 |     emitGetVirtualRegisters(op1, regT0, op2, regT2); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT2); | 
 |  | 
 |     addSlowCase(branchPtr(Equal, regT2, TrustedImmPtr(JSValue::encode(jsNumber(0))))); | 
 |     m_assembler.cdq(); | 
 |     m_assembler.idivl_r(regT2); | 
 |     emitFastArithReTagImmediate(regT1, regT0); | 
 |     emitPutVirtualRegister(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_mod(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |  | 
 |     linkSlowCase(iter); | 
 |     linkSlowCase(iter); | 
 |     linkSlowCase(iter); | 
 |     JITStubCall stubCall(this, cti_op_mod); | 
 |     stubCall.addArgument(regT0); | 
 |     stubCall.addArgument(regT2); | 
 |     stubCall.call(result); | 
 | } | 
 |  | 
 | #else // CPU(X86) || CPU(X86_64) || CPU(MIPS) | 
 |  | 
 | void JIT::emit_op_mod(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |  | 
 |     JITStubCall stubCall(this, cti_op_mod); | 
 |     stubCall.addArgument(op1, regT2); | 
 |     stubCall.addArgument(op2, regT2); | 
 |     stubCall.call(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_mod(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 | #if ENABLE(JIT_USE_SOFT_MODULO) | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     linkSlowCase(iter); | 
 |     linkSlowCase(iter); | 
 |     linkSlowCase(iter); | 
 |     JITStubCall stubCall(this, cti_op_mod); | 
 |     stubCall.addArgument(op1, regT2); | 
 |     stubCall.addArgument(op2, regT2); | 
 |     stubCall.call(result); | 
 | #else | 
 |     ASSERT_NOT_REACHED(); | 
 | #endif | 
 | } | 
 |  | 
 | #endif // CPU(X86) || CPU(X86_64) | 
 |  | 
 | /* ------------------------------ END: OP_MOD ------------------------------ */ | 
 |  | 
 | /* ------------------------------ BEGIN: USE(JSVALUE64) (OP_ADD, OP_SUB, OP_MUL) ------------------------------ */ | 
 |  | 
 | void JIT::compileBinaryArithOp(OpcodeID opcodeID, unsigned, unsigned op1, unsigned op2, OperandTypes) | 
 | { | 
 |     emitGetVirtualRegisters(op1, regT0, op2, regT1); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |     emitJumpSlowCaseIfNotImmediateInteger(regT1); | 
 |     if (opcodeID == op_add) | 
 |         addSlowCase(branchAdd32(Overflow, regT1, regT0)); | 
 |     else if (opcodeID == op_sub) | 
 |         addSlowCase(branchSub32(Overflow, regT1, regT0)); | 
 |     else { | 
 |         ASSERT(opcodeID == op_mul); | 
 |         addSlowCase(branchMul32(Overflow, regT1, regT0)); | 
 |         addSlowCase(branchTest32(Zero, regT0)); | 
 |     } | 
 |     emitFastArithIntToImmNoCheck(regT0, regT0); | 
 | } | 
 |  | 
 | void JIT::compileBinaryArithOpSlowCase(OpcodeID opcodeID, Vector<SlowCaseEntry>::iterator& iter, unsigned result, unsigned op1, unsigned op2, OperandTypes types, bool op1HasImmediateIntFastCase, bool op2HasImmediateIntFastCase) | 
 | { | 
 |     // We assume that subtracting TagTypeNumber is equivalent to adding DoubleEncodeOffset. | 
 |     COMPILE_ASSERT(((TagTypeNumber + DoubleEncodeOffset) == 0), TagTypeNumber_PLUS_DoubleEncodeOffset_EQUALS_0); | 
 |  | 
 |     Jump notImm1; | 
 |     Jump notImm2; | 
 |     if (op1HasImmediateIntFastCase) { | 
 |         notImm2 = getSlowCase(iter); | 
 |     } else if (op2HasImmediateIntFastCase) { | 
 |         notImm1 = getSlowCase(iter); | 
 |     } else { | 
 |         notImm1 = getSlowCase(iter); | 
 |         notImm2 = getSlowCase(iter); | 
 |     } | 
 |  | 
 |     linkSlowCase(iter); // Integer overflow case - we could handle this in JIT code, but this is likely rare. | 
 |     if (opcodeID == op_mul && !op1HasImmediateIntFastCase && !op2HasImmediateIntFastCase) // op_mul has an extra slow case to handle 0 * negative number. | 
 |         linkSlowCase(iter); | 
 |     emitGetVirtualRegister(op1, regT0); | 
 |  | 
 |     Label stubFunctionCall(this); | 
 |     JITStubCall stubCall(this, opcodeID == op_add ? cti_op_add : opcodeID == op_sub ? cti_op_sub : cti_op_mul); | 
 |     if (op1HasImmediateIntFastCase || op2HasImmediateIntFastCase) { | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         emitGetVirtualRegister(op2, regT1); | 
 |     } | 
 |     stubCall.addArgument(regT0); | 
 |     stubCall.addArgument(regT1); | 
 |     stubCall.call(result); | 
 |     Jump end = jump(); | 
 |  | 
 |     if (op1HasImmediateIntFastCase) { | 
 |         notImm2.link(this); | 
 |         if (!types.second().definitelyIsNumber()) | 
 |             emitJumpIfNotImmediateNumber(regT0).linkTo(stubFunctionCall, this); | 
 |         emitGetVirtualRegister(op1, regT1); | 
 |         convertInt32ToDouble(regT1, fpRegT1); | 
 |         addPtr(tagTypeNumberRegister, regT0); | 
 |         movePtrToDouble(regT0, fpRegT2); | 
 |     } else if (op2HasImmediateIntFastCase) { | 
 |         notImm1.link(this); | 
 |         if (!types.first().definitelyIsNumber()) | 
 |             emitJumpIfNotImmediateNumber(regT0).linkTo(stubFunctionCall, this); | 
 |         emitGetVirtualRegister(op2, regT1); | 
 |         convertInt32ToDouble(regT1, fpRegT1); | 
 |         addPtr(tagTypeNumberRegister, regT0); | 
 |         movePtrToDouble(regT0, fpRegT2); | 
 |     } else { | 
 |         // if we get here, eax is not an int32, edx not yet checked. | 
 |         notImm1.link(this); | 
 |         if (!types.first().definitelyIsNumber()) | 
 |             emitJumpIfNotImmediateNumber(regT0).linkTo(stubFunctionCall, this); | 
 |         if (!types.second().definitelyIsNumber()) | 
 |             emitJumpIfNotImmediateNumber(regT1).linkTo(stubFunctionCall, this); | 
 |         addPtr(tagTypeNumberRegister, regT0); | 
 |         movePtrToDouble(regT0, fpRegT1); | 
 |         Jump op2isDouble = emitJumpIfNotImmediateInteger(regT1); | 
 |         convertInt32ToDouble(regT1, fpRegT2); | 
 |         Jump op2wasInteger = jump(); | 
 |  | 
 |         // if we get here, eax IS an int32, edx is not. | 
 |         notImm2.link(this); | 
 |         if (!types.second().definitelyIsNumber()) | 
 |             emitJumpIfNotImmediateNumber(regT1).linkTo(stubFunctionCall, this); | 
 |         convertInt32ToDouble(regT0, fpRegT1); | 
 |         op2isDouble.link(this); | 
 |         addPtr(tagTypeNumberRegister, regT1); | 
 |         movePtrToDouble(regT1, fpRegT2); | 
 |         op2wasInteger.link(this); | 
 |     } | 
 |  | 
 |     if (opcodeID == op_add) | 
 |         addDouble(fpRegT2, fpRegT1); | 
 |     else if (opcodeID == op_sub) | 
 |         subDouble(fpRegT2, fpRegT1); | 
 |     else if (opcodeID == op_mul) | 
 |         mulDouble(fpRegT2, fpRegT1); | 
 |     else { | 
 |         ASSERT(opcodeID == op_div); | 
 |         divDouble(fpRegT2, fpRegT1); | 
 |     } | 
 |     moveDoubleToPtr(fpRegT1, regT0); | 
 |     subPtr(tagTypeNumberRegister, regT0); | 
 |     emitPutVirtualRegister(result, regT0); | 
 |  | 
 |     end.link(this); | 
 | } | 
 |  | 
 | void JIT::emit_op_add(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     OperandTypes types = OperandTypes::fromInt(currentInstruction[4].u.operand); | 
 |  | 
 |     if (!types.first().mightBeNumber() || !types.second().mightBeNumber()) { | 
 |         JITStubCall stubCall(this, cti_op_add); | 
 |         stubCall.addArgument(op1, regT2); | 
 |         stubCall.addArgument(op2, regT2); | 
 |         stubCall.call(result); | 
 |         return; | 
 |     } | 
 |  | 
 |     if (isOperandConstantImmediateInt(op1)) { | 
 |         emitGetVirtualRegister(op2, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         addSlowCase(branchAdd32(Overflow, Imm32(getConstantOperandImmediateInt(op1)), regT0)); | 
 |         emitFastArithIntToImmNoCheck(regT0, regT0); | 
 |     } else if (isOperandConstantImmediateInt(op2)) { | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         addSlowCase(branchAdd32(Overflow, Imm32(getConstantOperandImmediateInt(op2)), regT0)); | 
 |         emitFastArithIntToImmNoCheck(regT0, regT0); | 
 |     } else | 
 |         compileBinaryArithOp(op_add, result, op1, op2, types); | 
 |  | 
 |     emitPutVirtualRegister(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_add(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     OperandTypes types = OperandTypes::fromInt(currentInstruction[4].u.operand); | 
 |  | 
 |     if (!types.first().mightBeNumber() || !types.second().mightBeNumber()) | 
 |         return; | 
 |  | 
 |     bool op1HasImmediateIntFastCase = isOperandConstantImmediateInt(op1); | 
 |     bool op2HasImmediateIntFastCase = !op1HasImmediateIntFastCase && isOperandConstantImmediateInt(op2); | 
 |     compileBinaryArithOpSlowCase(op_add, iter, result, op1, op2, types, op1HasImmediateIntFastCase, op2HasImmediateIntFastCase); | 
 | } | 
 |  | 
 | void JIT::emit_op_mul(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     OperandTypes types = OperandTypes::fromInt(currentInstruction[4].u.operand); | 
 |  | 
 |     // For now, only plant a fast int case if the constant operand is greater than zero. | 
 |     int32_t value; | 
 |     if (isOperandConstantImmediateInt(op1) && ((value = getConstantOperandImmediateInt(op1)) > 0)) { | 
 |         emitGetVirtualRegister(op2, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         addSlowCase(branchMul32(Overflow, Imm32(value), regT0, regT0)); | 
 |         emitFastArithReTagImmediate(regT0, regT0); | 
 |     } else if (isOperandConstantImmediateInt(op2) && ((value = getConstantOperandImmediateInt(op2)) > 0)) { | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         emitJumpSlowCaseIfNotImmediateInteger(regT0); | 
 |         addSlowCase(branchMul32(Overflow, Imm32(value), regT0, regT0)); | 
 |         emitFastArithReTagImmediate(regT0, regT0); | 
 |     } else | 
 |         compileBinaryArithOp(op_mul, result, op1, op2, types); | 
 |  | 
 |     emitPutVirtualRegister(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_mul(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     OperandTypes types = OperandTypes::fromInt(currentInstruction[4].u.operand); | 
 |  | 
 |     bool op1HasImmediateIntFastCase = isOperandConstantImmediateInt(op1) && getConstantOperandImmediateInt(op1) > 0; | 
 |     bool op2HasImmediateIntFastCase = !op1HasImmediateIntFastCase && isOperandConstantImmediateInt(op2) && getConstantOperandImmediateInt(op2) > 0; | 
 |     compileBinaryArithOpSlowCase(op_mul, iter, result, op1, op2, types, op1HasImmediateIntFastCase, op2HasImmediateIntFastCase); | 
 | } | 
 |  | 
 | void JIT::emit_op_div(Instruction* currentInstruction) | 
 | { | 
 |     unsigned dst = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     OperandTypes types = OperandTypes::fromInt(currentInstruction[4].u.operand); | 
 |  | 
 |     if (isOperandConstantImmediateDouble(op1)) { | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         addPtr(tagTypeNumberRegister, regT0); | 
 |         movePtrToDouble(regT0, fpRegT0); | 
 |     } else if (isOperandConstantImmediateInt(op1)) { | 
 |         emitLoadInt32ToDouble(op1, fpRegT0); | 
 |     } else { | 
 |         emitGetVirtualRegister(op1, regT0); | 
 |         if (!types.first().definitelyIsNumber()) | 
 |             emitJumpSlowCaseIfNotImmediateNumber(regT0); | 
 |         Jump notInt = emitJumpIfNotImmediateInteger(regT0); | 
 |         convertInt32ToDouble(regT0, fpRegT0); | 
 |         Jump skipDoubleLoad = jump(); | 
 |         notInt.link(this); | 
 |         addPtr(tagTypeNumberRegister, regT0); | 
 |         movePtrToDouble(regT0, fpRegT0); | 
 |         skipDoubleLoad.link(this); | 
 |     } | 
 |  | 
 |     if (isOperandConstantImmediateDouble(op2)) { | 
 |         emitGetVirtualRegister(op2, regT1); | 
 |         addPtr(tagTypeNumberRegister, regT1); | 
 |         movePtrToDouble(regT1, fpRegT1); | 
 |     } else if (isOperandConstantImmediateInt(op2)) { | 
 |         emitLoadInt32ToDouble(op2, fpRegT1); | 
 |     } else { | 
 |         emitGetVirtualRegister(op2, regT1); | 
 |         if (!types.second().definitelyIsNumber()) | 
 |             emitJumpSlowCaseIfNotImmediateNumber(regT1); | 
 |         Jump notInt = emitJumpIfNotImmediateInteger(regT1); | 
 |         convertInt32ToDouble(regT1, fpRegT1); | 
 |         Jump skipDoubleLoad = jump(); | 
 |         notInt.link(this); | 
 |         addPtr(tagTypeNumberRegister, regT1); | 
 |         movePtrToDouble(regT1, fpRegT1); | 
 |         skipDoubleLoad.link(this); | 
 |     } | 
 |     divDouble(fpRegT1, fpRegT0); | 
 |  | 
 |     // Double result. | 
 |     moveDoubleToPtr(fpRegT0, regT0); | 
 |     subPtr(tagTypeNumberRegister, regT0); | 
 |  | 
 |     emitPutVirtualRegister(dst, regT0); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_div(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     OperandTypes types = OperandTypes::fromInt(currentInstruction[4].u.operand); | 
 |     if (types.first().definitelyIsNumber() && types.second().definitelyIsNumber()) { | 
 | #ifndef NDEBUG | 
 |         breakpoint(); | 
 | #endif | 
 |         return; | 
 |     } | 
 |     if (!isOperandConstantImmediateDouble(op1) && !isOperandConstantImmediateInt(op1)) { | 
 |         if (!types.first().definitelyIsNumber()) | 
 |             linkSlowCase(iter); | 
 |     } | 
 |     if (!isOperandConstantImmediateDouble(op2) && !isOperandConstantImmediateInt(op2)) { | 
 |         if (!types.second().definitelyIsNumber()) | 
 |             linkSlowCase(iter); | 
 |     } | 
 |     // There is an extra slow case for (op1 * -N) or (-N * op2), to check for 0 since this should produce a result of -0. | 
 |     JITStubCall stubCall(this, cti_op_div); | 
 |     stubCall.addArgument(op1, regT2); | 
 |     stubCall.addArgument(op2, regT2); | 
 |     stubCall.call(result); | 
 | } | 
 |  | 
 | void JIT::emit_op_sub(Instruction* currentInstruction) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     OperandTypes types = OperandTypes::fromInt(currentInstruction[4].u.operand); | 
 |  | 
 |     compileBinaryArithOp(op_sub, result, op1, op2, types); | 
 |     emitPutVirtualRegister(result); | 
 | } | 
 |  | 
 | void JIT::emitSlow_op_sub(Instruction* currentInstruction, Vector<SlowCaseEntry>::iterator& iter) | 
 | { | 
 |     unsigned result = currentInstruction[1].u.operand; | 
 |     unsigned op1 = currentInstruction[2].u.operand; | 
 |     unsigned op2 = currentInstruction[3].u.operand; | 
 |     OperandTypes types = OperandTypes::fromInt(currentInstruction[4].u.operand); | 
 |  | 
 |     compileBinaryArithOpSlowCase(op_sub, iter, result, op1, op2, types, false, false); | 
 | } | 
 |  | 
 | /* ------------------------------ END: OP_ADD, OP_SUB, OP_MUL ------------------------------ */ | 
 |  | 
 | #endif // USE(JSVALUE64) | 
 |  | 
 | } // namespace JSC | 
 |  | 
 | #endif // ENABLE(JIT) |