Mercurial > hg > graal-compiler
diff src/cpu/x86/vm/assembler_x86_64.hpp @ 0:a61af66fc99e jdk7-b24
Initial load
author | duke |
---|---|
date | Sat, 01 Dec 2007 00:00:00 +0000 |
parents | |
children | 3d62cb85208d |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/src/cpu/x86/vm/assembler_x86_64.hpp Sat Dec 01 00:00:00 2007 +0000 @@ -0,0 +1,1447 @@ +/* + * Copyright 2003-2007 Sun Microsystems, Inc. All Rights Reserved. + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. + * + * This code is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 only, as + * published by the Free Software Foundation. + * + * This code is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License + * version 2 for more details (a copy is included in the LICENSE file that + * accompanied this code). + * + * You should have received a copy of the GNU General Public License version + * 2 along with this work; if not, write to the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. + * + * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara, + * CA 95054 USA or visit www.sun.com if you need additional information or + * have any questions. + * + */ + +class BiasedLockingCounters; + +// Contains all the definitions needed for amd64 assembly code generation. + +#ifdef _LP64 +// Calling convention +class Argument VALUE_OBJ_CLASS_SPEC { + public: + enum { +#ifdef _WIN64 + n_int_register_parameters_c = 4, // rcx, rdx, r8, r9 (c_rarg0, c_rarg1, ...) + n_float_register_parameters_c = 4, // xmm0 - xmm3 (c_farg0, c_farg1, ... ) +#else + n_int_register_parameters_c = 6, // rdi, rsi, rdx, rcx, r8, r9 (c_rarg0, c_rarg1, ...) + n_float_register_parameters_c = 8, // xmm0 - xmm7 (c_farg0, c_farg1, ... ) +#endif + n_int_register_parameters_j = 6, // j_rarg0, j_rarg1, ... + n_float_register_parameters_j = 8 // j_farg0, j_farg1, ... + }; +}; + + +// Symbolically name the register arguments used by the c calling convention. +// Windows is different from linux/solaris. So much for standards... + +#ifdef _WIN64 + +REGISTER_DECLARATION(Register, c_rarg0, rcx); +REGISTER_DECLARATION(Register, c_rarg1, rdx); +REGISTER_DECLARATION(Register, c_rarg2, r8); +REGISTER_DECLARATION(Register, c_rarg3, r9); + +REGISTER_DECLARATION(XMMRegister, c_farg0, xmm0); +REGISTER_DECLARATION(XMMRegister, c_farg1, xmm1); +REGISTER_DECLARATION(XMMRegister, c_farg2, xmm2); +REGISTER_DECLARATION(XMMRegister, c_farg3, xmm3); + +#else + +REGISTER_DECLARATION(Register, c_rarg0, rdi); +REGISTER_DECLARATION(Register, c_rarg1, rsi); +REGISTER_DECLARATION(Register, c_rarg2, rdx); +REGISTER_DECLARATION(Register, c_rarg3, rcx); +REGISTER_DECLARATION(Register, c_rarg4, r8); +REGISTER_DECLARATION(Register, c_rarg5, r9); + +REGISTER_DECLARATION(XMMRegister, c_farg0, xmm0); +REGISTER_DECLARATION(XMMRegister, c_farg1, xmm1); +REGISTER_DECLARATION(XMMRegister, c_farg2, xmm2); +REGISTER_DECLARATION(XMMRegister, c_farg3, xmm3); +REGISTER_DECLARATION(XMMRegister, c_farg4, xmm4); +REGISTER_DECLARATION(XMMRegister, c_farg5, xmm5); +REGISTER_DECLARATION(XMMRegister, c_farg6, xmm6); +REGISTER_DECLARATION(XMMRegister, c_farg7, xmm7); + +#endif + +// Symbolically name the register arguments used by the Java calling convention. +// We have control over the convention for java so we can do what we please. +// What pleases us is to offset the java calling convention so that when +// we call a suitable jni method the arguments are lined up and we don't +// have to do little shuffling. A suitable jni method is non-static and a +// small number of arguments (two fewer args on windows) +// +// |-------------------------------------------------------| +// | c_rarg0 c_rarg1 c_rarg2 c_rarg3 c_rarg4 c_rarg5 | +// |-------------------------------------------------------| +// | rcx rdx r8 r9 rdi* rsi* | windows (* not a c_rarg) +// | rdi rsi rdx rcx r8 r9 | solaris/linux +// |-------------------------------------------------------| +// | j_rarg5 j_rarg0 j_rarg1 j_rarg2 j_rarg3 j_rarg4 | +// |-------------------------------------------------------| + +REGISTER_DECLARATION(Register, j_rarg0, c_rarg1); +REGISTER_DECLARATION(Register, j_rarg1, c_rarg2); +REGISTER_DECLARATION(Register, j_rarg2, c_rarg3); +// Windows runs out of register args here +#ifdef _WIN64 +REGISTER_DECLARATION(Register, j_rarg3, rdi); +REGISTER_DECLARATION(Register, j_rarg4, rsi); +#else +REGISTER_DECLARATION(Register, j_rarg3, c_rarg4); +REGISTER_DECLARATION(Register, j_rarg4, c_rarg5); +#endif /* _WIN64 */ +REGISTER_DECLARATION(Register, j_rarg5, c_rarg0); + +REGISTER_DECLARATION(XMMRegister, j_farg0, xmm0); +REGISTER_DECLARATION(XMMRegister, j_farg1, xmm1); +REGISTER_DECLARATION(XMMRegister, j_farg2, xmm2); +REGISTER_DECLARATION(XMMRegister, j_farg3, xmm3); +REGISTER_DECLARATION(XMMRegister, j_farg4, xmm4); +REGISTER_DECLARATION(XMMRegister, j_farg5, xmm5); +REGISTER_DECLARATION(XMMRegister, j_farg6, xmm6); +REGISTER_DECLARATION(XMMRegister, j_farg7, xmm7); + +REGISTER_DECLARATION(Register, rscratch1, r10); // volatile +REGISTER_DECLARATION(Register, rscratch2, r11); // volatile + +REGISTER_DECLARATION(Register, r15_thread, r15); // callee-saved + +#endif // _LP64 + +// Address is an abstraction used to represent a memory location +// using any of the amd64 addressing modes with one object. +// +// Note: A register location is represented via a Register, not +// via an address for efficiency & simplicity reasons. + +class ArrayAddress; + +class Address VALUE_OBJ_CLASS_SPEC { + public: + enum ScaleFactor { + no_scale = -1, + times_1 = 0, + times_2 = 1, + times_4 = 2, + times_8 = 3 + }; + + private: + Register _base; + Register _index; + ScaleFactor _scale; + int _disp; + RelocationHolder _rspec; + + // Easily misused constructors make them private + Address(int disp, address loc, relocInfo::relocType rtype); + Address(int disp, address loc, RelocationHolder spec); + + public: + // creation + Address() + : _base(noreg), + _index(noreg), + _scale(no_scale), + _disp(0) { + } + + // No default displacement otherwise Register can be implicitly + // converted to 0(Register) which is quite a different animal. + + Address(Register base, int disp) + : _base(base), + _index(noreg), + _scale(no_scale), + _disp(disp) { + } + + Address(Register base, Register index, ScaleFactor scale, int disp = 0) + : _base (base), + _index(index), + _scale(scale), + _disp (disp) { + assert(!index->is_valid() == (scale == Address::no_scale), + "inconsistent address"); + } + + // The following two overloads are used in connection with the + // ByteSize type (see sizes.hpp). They simplify the use of + // ByteSize'd arguments in assembly code. Note that their equivalent + // for the optimized build are the member functions with int disp + // argument since ByteSize is mapped to an int type in that case. + // + // Note: DO NOT introduce similar overloaded functions for WordSize + // arguments as in the optimized mode, both ByteSize and WordSize + // are mapped to the same type and thus the compiler cannot make a + // distinction anymore (=> compiler errors). + +#ifdef ASSERT + Address(Register base, ByteSize disp) + : _base(base), + _index(noreg), + _scale(no_scale), + _disp(in_bytes(disp)) { + } + + Address(Register base, Register index, ScaleFactor scale, ByteSize disp) + : _base(base), + _index(index), + _scale(scale), + _disp(in_bytes(disp)) { + assert(!index->is_valid() == (scale == Address::no_scale), + "inconsistent address"); + } +#endif // ASSERT + + // accessors + bool uses(Register reg) const { + return _base == reg || _index == reg; + } + + // Convert the raw encoding form into the form expected by the constructor for + // Address. An index of 4 (rsp) corresponds to having no index, so convert + // that to noreg for the Address constructor. + static Address make_raw(int base, int index, int scale, int disp); + + static Address make_array(ArrayAddress); + + private: + bool base_needs_rex() const { + return _base != noreg && _base->encoding() >= 8; + } + + bool index_needs_rex() const { + return _index != noreg &&_index->encoding() >= 8; + } + + relocInfo::relocType reloc() const { return _rspec.type(); } + + friend class Assembler; + friend class MacroAssembler; + friend class LIR_Assembler; // base/index/scale/disp +}; + +// +// AddressLiteral has been split out from Address because operands of this type +// need to be treated specially on 32bit vs. 64bit platforms. By splitting it out +// the few instructions that need to deal with address literals are unique and the +// MacroAssembler does not have to implement every instruction in the Assembler +// in order to search for address literals that may need special handling depending +// on the instruction and the platform. As small step on the way to merging i486/amd64 +// directories. +// +class AddressLiteral VALUE_OBJ_CLASS_SPEC { + friend class ArrayAddress; + RelocationHolder _rspec; + // Typically we use AddressLiterals we want to use their rval + // However in some situations we want the lval (effect address) of the item. + // We provide a special factory for making those lvals. + bool _is_lval; + + // If the target is far we'll need to load the ea of this to + // a register to reach it. Otherwise if near we can do rip + // relative addressing. + + address _target; + + protected: + // creation + AddressLiteral() + : _is_lval(false), + _target(NULL) + {} + + public: + + + AddressLiteral(address target, relocInfo::relocType rtype); + + AddressLiteral(address target, RelocationHolder const& rspec) + : _rspec(rspec), + _is_lval(false), + _target(target) + {} + + AddressLiteral addr() { + AddressLiteral ret = *this; + ret._is_lval = true; + return ret; + } + + + private: + + address target() { return _target; } + bool is_lval() { return _is_lval; } + + relocInfo::relocType reloc() const { return _rspec.type(); } + const RelocationHolder& rspec() const { return _rspec; } + + friend class Assembler; + friend class MacroAssembler; + friend class Address; + friend class LIR_Assembler; +}; + +// Convience classes +class RuntimeAddress: public AddressLiteral { + + public: + + RuntimeAddress(address target) : AddressLiteral(target, relocInfo::runtime_call_type) {} + +}; + +class OopAddress: public AddressLiteral { + + public: + + OopAddress(address target) : AddressLiteral(target, relocInfo::oop_type){} + +}; + +class ExternalAddress: public AddressLiteral { + + public: + + ExternalAddress(address target) : AddressLiteral(target, relocInfo::external_word_type){} + +}; + +class InternalAddress: public AddressLiteral { + + public: + + InternalAddress(address target) : AddressLiteral(target, relocInfo::internal_word_type) {} + +}; + +// x86 can do array addressing as a single operation since disp can be an absolute +// address but amd64 can't [e.g. array_base(rx, ry:width) ]. We create a class +// that expresses the concept but does extra magic on amd64 to get the final result + +class ArrayAddress VALUE_OBJ_CLASS_SPEC { + private: + + AddressLiteral _base; + Address _index; + + public: + + ArrayAddress() {}; + ArrayAddress(AddressLiteral base, Address index): _base(base), _index(index) {}; + AddressLiteral base() { return _base; } + Address index() { return _index; } + +}; + +// The amd64 Assembler: Pure assembler doing NO optimizations on +// the instruction level (e.g. mov rax, 0 is not translated into xor +// rax, rax!); i.e., what you write is what you get. The Assembler is +// generating code into a CodeBuffer. + +const int FPUStateSizeInWords = 512 / wordSize; + +class Assembler : public AbstractAssembler { + friend class AbstractAssembler; // for the non-virtual hack + friend class StubGenerator; + + + protected: +#ifdef ASSERT + void check_relocation(RelocationHolder const& rspec, int format); +#endif + + inline void emit_long64(jlong x); + + void emit_data(jint data, relocInfo::relocType rtype, int format /* = 1 */); + void emit_data(jint data, RelocationHolder const& rspec, int format /* = 1 */); + void emit_data64(jlong data, relocInfo::relocType rtype, int format = 0); + void emit_data64(jlong data, RelocationHolder const& rspec, int format = 0); + + // Helper functions for groups of instructions + void emit_arith_b(int op1, int op2, Register dst, int imm8); + + void emit_arith(int op1, int op2, Register dst, int imm32); + // only x86?? + void emit_arith(int op1, int op2, Register dst, jobject obj); + void emit_arith(int op1, int op2, Register dst, Register src); + + void emit_operand(Register reg, + Register base, Register index, Address::ScaleFactor scale, + int disp, + RelocationHolder const& rspec, + int rip_relative_correction = 0); + void emit_operand(Register reg, Address adr, + int rip_relative_correction = 0); + void emit_operand(XMMRegister reg, + Register base, Register index, Address::ScaleFactor scale, + int disp, + RelocationHolder const& rspec, + int rip_relative_correction = 0); + void emit_operand(XMMRegister reg, Address adr, + int rip_relative_correction = 0); + + // Immediate-to-memory forms + void emit_arith_operand(int op1, Register rm, Address adr, int imm32); + + void emit_farith(int b1, int b2, int i); + + bool reachable(AddressLiteral adr); + + // These are all easily abused and hence protected + + // Make these disappear in 64bit mode since they would never be correct +#ifndef _LP64 + void cmp_literal32(Register src1, int32_t imm32, RelocationHolder const& rspec); + void cmp_literal32(Address src1, int32_t imm32, RelocationHolder const& rspec); + + void mov_literal32(Register dst, int32_t imm32, RelocationHolder const& rspec); + void mov_literal32(Address dst, int32_t imm32, RelocationHolder const& rspec); + + void push_literal32(int32_t imm32, RelocationHolder const& rspec); +#endif // _LP64 + + + void mov_literal64(Register dst, intptr_t imm64, RelocationHolder const& rspec); + + // These are unique in that we are ensured by the caller that the 32bit + // relative in these instructions will always be able to reach the potentially + // 64bit address described by entry. Since they can take a 64bit address they + // don't have the 32 suffix like the other instructions in this class. + void jmp_literal(address entry, RelocationHolder const& rspec); + void call_literal(address entry, RelocationHolder const& rspec); + + public: + enum Condition { // The amd64 condition codes used for conditional jumps/moves. + zero = 0x4, + notZero = 0x5, + equal = 0x4, + notEqual = 0x5, + less = 0xc, + lessEqual = 0xe, + greater = 0xf, + greaterEqual = 0xd, + below = 0x2, + belowEqual = 0x6, + above = 0x7, + aboveEqual = 0x3, + overflow = 0x0, + noOverflow = 0x1, + carrySet = 0x2, + carryClear = 0x3, + negative = 0x8, + positive = 0x9, + parity = 0xa, + noParity = 0xb + }; + + enum Prefix { + // segment overrides + // XXX remove segment prefixes + CS_segment = 0x2e, + SS_segment = 0x36, + DS_segment = 0x3e, + ES_segment = 0x26, + FS_segment = 0x64, + GS_segment = 0x65, + + REX = 0x40, + + REX_B = 0x41, + REX_X = 0x42, + REX_XB = 0x43, + REX_R = 0x44, + REX_RB = 0x45, + REX_RX = 0x46, + REX_RXB = 0x47, + + REX_W = 0x48, + + REX_WB = 0x49, + REX_WX = 0x4A, + REX_WXB = 0x4B, + REX_WR = 0x4C, + REX_WRB = 0x4D, + REX_WRX = 0x4E, + REX_WRXB = 0x4F + }; + + enum WhichOperand { + // input to locate_operand, and format code for relocations + imm64_operand = 0, // embedded 64-bit immediate operand + disp32_operand = 1, // embedded 32-bit displacement + call32_operand = 2, // embedded 32-bit self-relative displacement + _WhichOperand_limit = 3 + }; + + public: + + // Creation + Assembler(CodeBuffer* code) + : AbstractAssembler(code) { + } + + // Decoding + static address locate_operand(address inst, WhichOperand which); + static address locate_next_instruction(address inst); + + // Utilities + + static bool is_simm(int64_t x, int nbits) { return -( CONST64(1) << (nbits-1) ) <= x && x < ( CONST64(1) << (nbits-1) ); } + static bool is_simm32 (int64_t x) { return x == (int64_t)(int32_t)x; } + + + // Stack + void pushaq(); + void popaq(); + + void pushfq(); + void popfq(); + + void pushq(int imm32); + + void pushq(Register src); + void pushq(Address src); + + void popq(Register dst); + void popq(Address dst); + + // Instruction prefixes + void prefix(Prefix p); + + int prefix_and_encode(int reg_enc, bool byteinst = false); + int prefixq_and_encode(int reg_enc); + + int prefix_and_encode(int dst_enc, int src_enc, bool byteinst = false); + int prefixq_and_encode(int dst_enc, int src_enc); + + void prefix(Register reg); + void prefix(Address adr); + void prefixq(Address adr); + + void prefix(Address adr, Register reg, bool byteinst = false); + void prefixq(Address adr, Register reg); + + void prefix(Address adr, XMMRegister reg); + + // Moves + void movb(Register dst, Address src); + void movb(Address dst, int imm8); + void movb(Address dst, Register src); + + void movw(Address dst, int imm16); + void movw(Register dst, Address src); + void movw(Address dst, Register src); + + void movl(Register dst, int imm32); + void movl(Register dst, Register src); + void movl(Register dst, Address src); + void movl(Address dst, int imm32); + void movl(Address dst, Register src); + + void movq(Register dst, Register src); + void movq(Register dst, Address src); + void movq(Address dst, Register src); + // These prevent using movq from converting a zero (like NULL) into Register + // by giving the compiler two choices it can't resolve + void movq(Address dst, void* dummy); + void movq(Register dst, void* dummy); + + void mov64(Register dst, intptr_t imm64); + void mov64(Address dst, intptr_t imm64); + + void movsbl(Register dst, Address src); + void movsbl(Register dst, Register src); + void movswl(Register dst, Address src); + void movswl(Register dst, Register src); + void movslq(Register dst, Address src); + void movslq(Register dst, Register src); + + void movzbl(Register dst, Address src); + void movzbl(Register dst, Register src); + void movzwl(Register dst, Address src); + void movzwl(Register dst, Register src); + + protected: // Avoid using the next instructions directly. + // New cpus require use of movsd and movss to avoid partial register stall + // when loading from memory. But for old Opteron use movlpd instead of movsd. + // The selection is done in MacroAssembler::movdbl() and movflt(). + void movss(XMMRegister dst, XMMRegister src); + void movss(XMMRegister dst, Address src); + void movss(Address dst, XMMRegister src); + void movsd(XMMRegister dst, XMMRegister src); + void movsd(Address dst, XMMRegister src); + void movsd(XMMRegister dst, Address src); + void movlpd(XMMRegister dst, Address src); + // New cpus require use of movaps and movapd to avoid partial register stall + // when moving between registers. + void movapd(XMMRegister dst, XMMRegister src); + void movaps(XMMRegister dst, XMMRegister src); + public: + + void movdl(XMMRegister dst, Register src); + void movdl(Register dst, XMMRegister src); + void movdq(XMMRegister dst, Register src); + void movdq(Register dst, XMMRegister src); + + void cmovl(Condition cc, Register dst, Register src); + void cmovl(Condition cc, Register dst, Address src); + void cmovq(Condition cc, Register dst, Register src); + void cmovq(Condition cc, Register dst, Address src); + + // Prefetches + private: + void prefetch_prefix(Address src); + public: + void prefetcht0(Address src); + void prefetcht1(Address src); + void prefetcht2(Address src); + void prefetchnta(Address src); + void prefetchw(Address src); + + // Arithmetics + void adcl(Register dst, int imm32); + void adcl(Register dst, Address src); + void adcl(Register dst, Register src); + void adcq(Register dst, int imm32); + void adcq(Register dst, Address src); + void adcq(Register dst, Register src); + + void addl(Address dst, int imm32); + void addl(Address dst, Register src); + void addl(Register dst, int imm32); + void addl(Register dst, Address src); + void addl(Register dst, Register src); + void addq(Address dst, int imm32); + void addq(Address dst, Register src); + void addq(Register dst, int imm32); + void addq(Register dst, Address src); + void addq(Register dst, Register src); + + void andl(Register dst, int imm32); + void andl(Register dst, Address src); + void andl(Register dst, Register src); + void andq(Register dst, int imm32); + void andq(Register dst, Address src); + void andq(Register dst, Register src); + + void cmpb(Address dst, int imm8); + void cmpl(Address dst, int imm32); + void cmpl(Register dst, int imm32); + void cmpl(Register dst, Register src); + void cmpl(Register dst, Address src); + void cmpq(Address dst, int imm32); + void cmpq(Address dst, Register src); + void cmpq(Register dst, int imm32); + void cmpq(Register dst, Register src); + void cmpq(Register dst, Address src); + + void ucomiss(XMMRegister dst, XMMRegister src); + void ucomisd(XMMRegister dst, XMMRegister src); + + protected: + // Don't use next inc() and dec() methods directly. INC & DEC instructions + // could cause a partial flag stall since they don't set CF flag. + // Use MacroAssembler::decrement() & MacroAssembler::increment() methods + // which call inc() & dec() or add() & sub() in accordance with + // the product flag UseIncDec value. + + void decl(Register dst); + void decl(Address dst); + void decq(Register dst); + void decq(Address dst); + + void incl(Register dst); + void incl(Address dst); + void incq(Register dst); + void incq(Address dst); + + public: + void idivl(Register src); + void idivq(Register src); + void cdql(); + void cdqq(); + + void imull(Register dst, Register src); + void imull(Register dst, Register src, int value); + void imulq(Register dst, Register src); + void imulq(Register dst, Register src, int value); + + void leal(Register dst, Address src); + void leaq(Register dst, Address src); + + void mull(Address src); + void mull(Register src); + + void negl(Register dst); + void negq(Register dst); + + void notl(Register dst); + void notq(Register dst); + + void orl(Address dst, int imm32); + void orl(Register dst, int imm32); + void orl(Register dst, Address src); + void orl(Register dst, Register src); + void orq(Address dst, int imm32); + void orq(Register dst, int imm32); + void orq(Register dst, Address src); + void orq(Register dst, Register src); + + void rcll(Register dst, int imm8); + void rclq(Register dst, int imm8); + + void sarl(Register dst, int imm8); + void sarl(Register dst); + void sarq(Register dst, int imm8); + void sarq(Register dst); + + void sbbl(Address dst, int imm32); + void sbbl(Register dst, int imm32); + void sbbl(Register dst, Address src); + void sbbl(Register dst, Register src); + void sbbq(Address dst, int imm32); + void sbbq(Register dst, int imm32); + void sbbq(Register dst, Address src); + void sbbq(Register dst, Register src); + + void shll(Register dst, int imm8); + void shll(Register dst); + void shlq(Register dst, int imm8); + void shlq(Register dst); + + void shrl(Register dst, int imm8); + void shrl(Register dst); + void shrq(Register dst, int imm8); + void shrq(Register dst); + + void subl(Address dst, int imm32); + void subl(Address dst, Register src); + void subl(Register dst, int imm32); + void subl(Register dst, Address src); + void subl(Register dst, Register src); + void subq(Address dst, int imm32); + void subq(Address dst, Register src); + void subq(Register dst, int imm32); + void subq(Register dst, Address src); + void subq(Register dst, Register src); + + void testb(Register dst, int imm8); + void testl(Register dst, int imm32); + void testl(Register dst, Register src); + void testq(Register dst, int imm32); + void testq(Register dst, Register src); + + void xaddl(Address dst, Register src); + void xaddq(Address dst, Register src); + + void xorl(Register dst, int imm32); + void xorl(Register dst, Address src); + void xorl(Register dst, Register src); + void xorq(Register dst, int imm32); + void xorq(Register dst, Address src); + void xorq(Register dst, Register src); + + // Miscellaneous + void bswapl(Register reg); + void bswapq(Register reg); + void lock(); + + void xchgl(Register reg, Address adr); + void xchgl(Register dst, Register src); + void xchgq(Register reg, Address adr); + void xchgq(Register dst, Register src); + + void cmpxchgl(Register reg, Address adr); + void cmpxchgq(Register reg, Address adr); + + void nop(int i = 1); + void addr_nop_4(); + void addr_nop_5(); + void addr_nop_7(); + void addr_nop_8(); + + void hlt(); + void ret(int imm16); + void smovl(); + void rep_movl(); + void rep_movq(); + void rep_set(); + void repne_scan(); + void setb(Condition cc, Register dst); + + void clflush(Address adr); + + enum Membar_mask_bits { + StoreStore = 1 << 3, + LoadStore = 1 << 2, + StoreLoad = 1 << 1, + LoadLoad = 1 << 0 + }; + + // Serializes memory. + void membar(Membar_mask_bits order_constraint) { + // We only have to handle StoreLoad and LoadLoad + if (order_constraint & StoreLoad) { + // MFENCE subsumes LFENCE + mfence(); + } /* [jk] not needed currently: else if (order_constraint & LoadLoad) { + lfence(); + } */ + } + + void lfence() { + emit_byte(0x0F); + emit_byte(0xAE); + emit_byte(0xE8); + } + + void mfence() { + emit_byte(0x0F); + emit_byte(0xAE); + emit_byte(0xF0); + } + + // Identify processor type and features + void cpuid() { + emit_byte(0x0F); + emit_byte(0xA2); + } + + void cld() { emit_byte(0xfc); + } + + void std() { emit_byte(0xfd); + } + + + // Calls + + void call(Label& L, relocInfo::relocType rtype); + void call(Register reg); + void call(Address adr); + + // Jumps + + void jmp(Register reg); + void jmp(Address adr); + + // Label operations & relative jumps (PPUM Appendix D) + // unconditional jump to L + void jmp(Label& L, relocInfo::relocType rtype = relocInfo::none); + + + // Unconditional 8-bit offset jump to L. + // WARNING: be very careful using this for forward jumps. If the label is + // not bound within an 8-bit offset of this instruction, a run-time error + // will occur. + void jmpb(Label& L); + + // jcc is the generic conditional branch generator to run- time + // routines, jcc is used for branches to labels. jcc takes a branch + // opcode (cc) and a label (L) and generates either a backward + // branch or a forward branch and links it to the label fixup + // chain. Usage: + // + // Label L; // unbound label + // jcc(cc, L); // forward branch to unbound label + // bind(L); // bind label to the current pc + // jcc(cc, L); // backward branch to bound label + // bind(L); // illegal: a label may be bound only once + // + // Note: The same Label can be used for forward and backward branches + // but it may be bound only once. + + void jcc(Condition cc, Label& L, + relocInfo::relocType rtype = relocInfo::none); + + // Conditional jump to a 8-bit offset to L. + // WARNING: be very careful using this for forward jumps. If the label is + // not bound within an 8-bit offset of this instruction, a run-time error + // will occur. + void jccb(Condition cc, Label& L); + + // Floating-point operations + + void fxsave(Address dst); + void fxrstor(Address src); + void ldmxcsr(Address src); + void stmxcsr(Address dst); + + void addss(XMMRegister dst, XMMRegister src); + void addss(XMMRegister dst, Address src); + void subss(XMMRegister dst, XMMRegister src); + void subss(XMMRegister dst, Address src); + void mulss(XMMRegister dst, XMMRegister src); + void mulss(XMMRegister dst, Address src); + void divss(XMMRegister dst, XMMRegister src); + void divss(XMMRegister dst, Address src); + void addsd(XMMRegister dst, XMMRegister src); + void addsd(XMMRegister dst, Address src); + void subsd(XMMRegister dst, XMMRegister src); + void subsd(XMMRegister dst, Address src); + void mulsd(XMMRegister dst, XMMRegister src); + void mulsd(XMMRegister dst, Address src); + void divsd(XMMRegister dst, XMMRegister src); + void divsd(XMMRegister dst, Address src); + + // We only need the double form + void sqrtsd(XMMRegister dst, XMMRegister src); + void sqrtsd(XMMRegister dst, Address src); + + void xorps(XMMRegister dst, XMMRegister src); + void xorps(XMMRegister dst, Address src); + void xorpd(XMMRegister dst, XMMRegister src); + void xorpd(XMMRegister dst, Address src); + + void cvtsi2ssl(XMMRegister dst, Register src); + void cvtsi2ssq(XMMRegister dst, Register src); + void cvtsi2sdl(XMMRegister dst, Register src); + void cvtsi2sdq(XMMRegister dst, Register src); + void cvttss2sil(Register dst, XMMRegister src); // truncates + void cvttss2siq(Register dst, XMMRegister src); // truncates + void cvttsd2sil(Register dst, XMMRegister src); // truncates + void cvttsd2siq(Register dst, XMMRegister src); // truncates + void cvtss2sd(XMMRegister dst, XMMRegister src); + void cvtsd2ss(XMMRegister dst, XMMRegister src); + + void pxor(XMMRegister dst, Address src); // Xor Packed Byte Integer Values + void pxor(XMMRegister dst, XMMRegister src); // Xor Packed Byte Integer Values + + void movdqa(XMMRegister dst, Address src); // Move Aligned Double Quadword + void movdqa(XMMRegister dst, XMMRegister src); + void movdqa(Address dst, XMMRegister src); + + void movq(XMMRegister dst, Address src); + void movq(Address dst, XMMRegister src); + + void pshufd(XMMRegister dst, XMMRegister src, int mode); // Shuffle Packed Doublewords + void pshufd(XMMRegister dst, Address src, int mode); + void pshuflw(XMMRegister dst, XMMRegister src, int mode); // Shuffle Packed Low Words + void pshuflw(XMMRegister dst, Address src, int mode); + + void psrlq(XMMRegister dst, int shift); // Shift Right Logical Quadword Immediate + + void punpcklbw(XMMRegister dst, XMMRegister src); // Interleave Low Bytes + void punpcklbw(XMMRegister dst, Address src); +}; + + +// MacroAssembler extends Assembler by frequently used macros. +// +// Instructions for which a 'better' code sequence exists depending +// on arguments should also go in here. + +class MacroAssembler : public Assembler { + friend class LIR_Assembler; + protected: + + Address as_Address(AddressLiteral adr); + Address as_Address(ArrayAddress adr); + + // Support for VM calls + // + // This is the base routine called by the different versions of + // call_VM_leaf. The interpreter may customize this version by + // overriding it for its purposes (e.g., to save/restore additional + // registers when doing a VM call). + + virtual void call_VM_leaf_base( + address entry_point, // the entry point + int number_of_arguments // the number of arguments to + // pop after the call + ); + + // This is the base routine called by the different versions of + // call_VM. The interpreter may customize this version by overriding + // it for its purposes (e.g., to save/restore additional registers + // when doing a VM call). + // + // If no java_thread register is specified (noreg) than rdi will be + // used instead. call_VM_base returns the register which contains + // the thread upon return. If a thread register has been specified, + // the return value will correspond to that register. If no + // last_java_sp is specified (noreg) than rsp will be used instead. + virtual void call_VM_base( // returns the register + // containing the thread upon + // return + Register oop_result, // where an oop-result ends up + // if any; use noreg otherwise + Register java_thread, // the thread if computed + // before ; use noreg otherwise + Register last_java_sp, // to set up last_Java_frame in + // stubs; use noreg otherwise + address entry_point, // the entry point + int number_of_arguments, // the number of arguments (w/o + // thread) to pop after the + // call + bool check_exceptions // whether to check for pending + // exceptions after return + ); + + // This routines should emit JVMTI PopFrame handling and ForceEarlyReturn code. + // The implementation is only non-empty for the InterpreterMacroAssembler, + // as only the interpreter handles PopFrame and ForceEarlyReturn requests. + virtual void check_and_handle_popframe(Register java_thread); + virtual void check_and_handle_earlyret(Register java_thread); + + void call_VM_helper(Register oop_result, + address entry_point, + int number_of_arguments, + bool check_exceptions = true); + + public: + MacroAssembler(CodeBuffer* code) : Assembler(code) {} + + // Support for NULL-checks + // + // Generates code that causes a NULL OS exception if the content of + // reg is NULL. If the accessed location is M[reg + offset] and the + // offset is known, provide the offset. No explicit code generation + // is needed if the offset is within a certain range (0 <= offset <= + // page_size). + void null_check(Register reg, int offset = -1); + static bool needs_explicit_null_check(int offset); + + // Required platform-specific helpers for Label::patch_instructions. + // They _shadow_ the declarations in AbstractAssembler, which are undefined. + void pd_patch_instruction(address branch, address target); +#ifndef PRODUCT + static void pd_print_patched_instruction(address branch); +#endif + + + // The following 4 methods return the offset of the appropriate move + // instruction. Note: these are 32 bit instructions + + // Support for fast byte/word loading with zero extension (depending + // on particular CPU) + int load_unsigned_byte(Register dst, Address src); + int load_unsigned_word(Register dst, Address src); + + // Support for fast byte/word loading with sign extension (depending + // on particular CPU) + int load_signed_byte(Register dst, Address src); + int load_signed_word(Register dst, Address src); + + // Support for inc/dec with optimal instruction selection depending + // on value + void incrementl(Register reg, int value = 1); + void decrementl(Register reg, int value = 1); + void incrementq(Register reg, int value = 1); + void decrementq(Register reg, int value = 1); + + void incrementl(Address dst, int value = 1); + void decrementl(Address dst, int value = 1); + void incrementq(Address dst, int value = 1); + void decrementq(Address dst, int value = 1); + + // Support optimal SSE move instructions. + void movflt(XMMRegister dst, XMMRegister src) { + if (UseXmmRegToRegMoveAll) { movaps(dst, src); return; } + else { movss (dst, src); return; } + } + + void movflt(XMMRegister dst, Address src) { movss(dst, src); } + + void movflt(XMMRegister dst, AddressLiteral src); + + void movflt(Address dst, XMMRegister src) { movss(dst, src); } + + void movdbl(XMMRegister dst, XMMRegister src) { + if (UseXmmRegToRegMoveAll) { movapd(dst, src); return; } + else { movsd (dst, src); return; } + } + + void movdbl(XMMRegister dst, AddressLiteral src); + + void movdbl(XMMRegister dst, Address src) { + if (UseXmmLoadAndClearUpper) { movsd (dst, src); return; } + else { movlpd(dst, src); return; } + } + + void movdbl(Address dst, XMMRegister src) { movsd(dst, src); } + + void incrementl(AddressLiteral dst); + void incrementl(ArrayAddress dst); + + // Alignment + void align(int modulus); + + // Misc + void fat_nop(); // 5 byte nop + + + // C++ bool manipulation + + void movbool(Register dst, Address src); + void movbool(Address dst, bool boolconst); + void movbool(Address dst, Register src); + void testbool(Register dst); + + // Stack frame creation/removal + void enter(); + void leave(); + + // Support for getting the JavaThread pointer (i.e.; a reference to + // thread-local information) The pointer will be loaded into the + // thread register. + void get_thread(Register thread); + + void int3(); + + // Support for VM calls + // + // It is imperative that all calls into the VM are handled via the + // call_VM macros. They make sure that the stack linkage is setup + // correctly. call_VM's correspond to ENTRY/ENTRY_X entry points + // while call_VM_leaf's correspond to LEAF entry points. + void call_VM(Register oop_result, + address entry_point, + bool check_exceptions = true); + void call_VM(Register oop_result, + address entry_point, + Register arg_1, + bool check_exceptions = true); + void call_VM(Register oop_result, + address entry_point, + Register arg_1, Register arg_2, + bool check_exceptions = true); + void call_VM(Register oop_result, + address entry_point, + Register arg_1, Register arg_2, Register arg_3, + bool check_exceptions = true); + + // Overloadings with last_Java_sp + void call_VM(Register oop_result, + Register last_java_sp, + address entry_point, + int number_of_arguments = 0, + bool check_exceptions = true); + void call_VM(Register oop_result, + Register last_java_sp, + address entry_point, + Register arg_1, bool + check_exceptions = true); + void call_VM(Register oop_result, + Register last_java_sp, + address entry_point, + Register arg_1, Register arg_2, + bool check_exceptions = true); + void call_VM(Register oop_result, + Register last_java_sp, + address entry_point, + Register arg_1, Register arg_2, Register arg_3, + bool check_exceptions = true); + + void call_VM_leaf(address entry_point, + int number_of_arguments = 0); + void call_VM_leaf(address entry_point, + Register arg_1); + void call_VM_leaf(address entry_point, + Register arg_1, Register arg_2); + void call_VM_leaf(address entry_point, + Register arg_1, Register arg_2, Register arg_3); + + // last Java Frame (fills frame anchor) + void set_last_Java_frame(Register last_java_sp, + Register last_java_fp, + address last_java_pc); + void reset_last_Java_frame(bool clear_fp, bool clear_pc); + + // Stores + void store_check(Register obj); // store check for + // obj - register is + // destroyed + // afterwards + void store_check(Register obj, Address dst); // same as above, dst + // is exact store + // location (reg. is + // destroyed) + + // split store_check(Register obj) to enhance instruction interleaving + void store_check_part_1(Register obj); + void store_check_part_2(Register obj); + + // C 'boolean' to Java boolean: x == 0 ? 0 : 1 + void c2bool(Register x); + + // Int division/reminder for Java + // (as idivl, but checks for special case as described in JVM spec.) + // returns idivl instruction offset for implicit exception handling + int corrected_idivl(Register reg); + // Long division/reminder for Java + // (as idivq, but checks for special case as described in JVM spec.) + // returns idivq instruction offset for implicit exception handling + int corrected_idivq(Register reg); + + // Push and pop integer/fpu/cpu state + void push_IU_state(); + void pop_IU_state(); + + void push_FPU_state(); + void pop_FPU_state(); + + void push_CPU_state(); + void pop_CPU_state(); + + // Sign extension + void sign_extend_short(Register reg); + void sign_extend_byte(Register reg); + + // Division by power of 2, rounding towards 0 + void division_with_shift(Register reg, int shift_value); + + // Round up to a power of two + void round_to_l(Register reg, int modulus); + void round_to_q(Register reg, int modulus); + + // allocation + void eden_allocate( + Register obj, // result: pointer to object after + // successful allocation + Register var_size_in_bytes, // object size in bytes if unknown at + // compile time; invalid otherwise + int con_size_in_bytes, // object size in bytes if known at + // compile time + Register t1, // temp register + Label& slow_case // continuation point if fast + // allocation fails + ); + void tlab_allocate( + Register obj, // result: pointer to object after + // successful allocation + Register var_size_in_bytes, // object size in bytes if unknown at + // compile time; invalid otherwise + int con_size_in_bytes, // object size in bytes if known at + // compile time + Register t1, // temp register + Register t2, // temp register + Label& slow_case // continuation point if fast + // allocation fails + ); + void tlab_refill(Label& retry_tlab, Label& try_eden, Label& slow_case); + + //---- + + // Debugging + + // only if +VerifyOops + void verify_oop(Register reg, const char* s = "broken oop"); + void verify_oop_addr(Address addr, const char * s = "broken oop addr"); + + // only if +VerifyFPU + void verify_FPU(int stack_depth, const char* s = "illegal FPU state") {} + + // prints msg, dumps registers and stops execution + void stop(const char* msg); + + // prints message and continues + void warn(const char* msg); + + static void debug(char* msg, int64_t pc, int64_t regs[]); + + void os_breakpoint(); + + void untested() + { + stop("untested"); + } + + void unimplemented(const char* what = "") + { + char* b = new char[1024]; + sprintf(b, "unimplemented: %s", what); + stop(b); + } + + void should_not_reach_here() + { + stop("should not reach here"); + } + + // Stack overflow checking + void bang_stack_with_offset(int offset) + { + // stack grows down, caller passes positive offset + assert(offset > 0, "must bang with negative offset"); + movl(Address(rsp, (-offset)), rax); + } + + // Writes to stack successive pages until offset reached to check for + // stack overflow + shadow pages. Also, clobbers tmp + void bang_stack_size(Register offset, Register tmp); + + // Support for serializing memory accesses between threads. + void serialize_memory(Register thread, Register tmp); + + void verify_tlab(); + + // Biased locking support + // lock_reg and obj_reg must be loaded up with the appropriate values. + // swap_reg must be rax and is killed. + // tmp_reg must be supplied and is killed. + // If swap_reg_contains_mark is true then the code assumes that the + // mark word of the object has already been loaded into swap_reg. + // Optional slow case is for implementations (interpreter and C1) which branch to + // slow case directly. Leaves condition codes set for C2's Fast_Lock node. + // Returns offset of first potentially-faulting instruction for null + // check info (currently consumed only by C1). If + // swap_reg_contains_mark is true then returns -1 as it is assumed + // the calling code has already passed any potential faults. + int biased_locking_enter(Register lock_reg, Register obj_reg, Register swap_reg, Register tmp_reg, + bool swap_reg_contains_mark, + Label& done, Label* slow_case = NULL, + BiasedLockingCounters* counters = NULL); + void biased_locking_exit (Register obj_reg, Register temp_reg, Label& done); + + Condition negate_condition(Condition cond); + + // Instructions that use AddressLiteral operands. These instruction can handle 32bit/64bit + // operands. In general the names are modified to avoid hiding the instruction in Assembler + // so that we don't need to implement all the varieties in the Assembler with trivial wrappers + // here in MacroAssembler. The major exception to this rule is call + + // Arithmetics + + void cmp8(AddressLiteral src1, int8_t imm32); + + void cmp32(AddressLiteral src1, int32_t src2); + // compare reg - mem, or reg - &mem + void cmp32(Register src1, AddressLiteral src2); + + void cmp32(Register src1, Address src2); + +#ifndef _LP64 + void cmpoop(Address dst, jobject obj); + void cmpoop(Register dst, jobject obj); +#endif // _LP64 + + // NOTE src2 must be the lval. This is NOT an mem-mem compare + void cmpptr(Address src1, AddressLiteral src2); + + void cmpptr(Register src1, AddressLiteral src); + + // will be cmpreg(?) + void cmp64(Register src1, AddressLiteral src); + + void cmpxchgptr(Register reg, Address adr); + void cmpxchgptr(Register reg, AddressLiteral adr); + + // Helper functions for statistics gathering. + // Conditionally (atomically, on MPs) increments passed counter address, preserving condition codes. + void cond_inc32(Condition cond, AddressLiteral counter_addr); + // Unconditional atomic increment. + void atomic_incl(AddressLiteral counter_addr); + + + void lea(Register dst, AddressLiteral src); + void lea(Register dst, Address src); + + + // Calls + void call(Label& L, relocInfo::relocType rtype); + void call(Register entry); + void call(AddressLiteral entry); + + // Jumps + + // 32bit can do a case table jump in one instruction but we no longer allow the base + // to be installed in the Address class + void jump(ArrayAddress entry); + + void jump(AddressLiteral entry); + void jump_cc(Condition cc, AddressLiteral dst); + + // Floating + + void ldmxcsr(Address src) { Assembler::ldmxcsr(src); } + void ldmxcsr(AddressLiteral src); + +private: + // these are private because users should be doing movflt/movdbl + + void movss(XMMRegister dst, XMMRegister src) { Assembler::movss(dst, src); } + void movss(Address dst, XMMRegister src) { Assembler::movss(dst, src); } + void movss(XMMRegister dst, Address src) { Assembler::movss(dst, src); } + void movss(XMMRegister dst, AddressLiteral src); + + void movlpd(XMMRegister dst, Address src) {Assembler::movlpd(dst, src); } + void movlpd(XMMRegister dst, AddressLiteral src); + +public: + + + void xorpd(XMMRegister dst, XMMRegister src) {Assembler::xorpd(dst, src); } + void xorpd(XMMRegister dst, Address src) {Assembler::xorpd(dst, src); } + void xorpd(XMMRegister dst, AddressLiteral src); + + void xorps(XMMRegister dst, XMMRegister src) {Assembler::xorps(dst, src); } + void xorps(XMMRegister dst, Address src) {Assembler::xorps(dst, src); } + void xorps(XMMRegister dst, AddressLiteral src); + + + // Data + + void movoop(Register dst, jobject obj); + void movoop(Address dst, jobject obj); + + void movptr(ArrayAddress dst, Register src); + void movptr(Register dst, AddressLiteral src); + + void movptr(Register dst, intptr_t src); + void movptr(Address dst, intptr_t src); + + void movptr(Register dst, ArrayAddress src); + + // to avoid hiding movl + void mov32(AddressLiteral dst, Register src); + void mov32(Register dst, AddressLiteral src); + + void pushoop(jobject obj); + + // Can push value or effective address + void pushptr(AddressLiteral src); + +}; + +/** + * class SkipIfEqual: + * + * Instantiating this class will result in assembly code being output that will + * jump around any code emitted between the creation of the instance and it's + * automatic destruction at the end of a scope block, depending on the value of + * the flag passed to the constructor, which will be checked at run-time. + */ +class SkipIfEqual { + private: + MacroAssembler* _masm; + Label _label; + + public: + SkipIfEqual(MacroAssembler*, const bool* flag_addr, bool value); + ~SkipIfEqual(); +}; + + +#ifdef ASSERT +inline bool AbstractAssembler::pd_check_instruction_mark() { return true; } +#endif