gecko/js/src/nanojit/Nativei386.h
Nicholas Nethercote 76b149adce Bug 557705 - nanojit: clean up codegen of i386 jumps. r=edwsmith.
--HG--
extra : convert_revision : ec93bc283385c4e19bfbfda1aa1ecf94aef932ec
2010-04-06 20:09:32 -07:00

443 lines
17 KiB
C++

/* -*- Mode: C++; c-basic-offset: 4; indent-tabs-mode: nil; tab-width: 4 -*- */
/* vi: set ts=4 sw=4 expandtab: (add to ~/.vimrc: set modeline modelines=5) */
/* ***** BEGIN LICENSE BLOCK *****
* Version: MPL 1.1/GPL 2.0/LGPL 2.1
*
* The contents of this file are subject to the Mozilla Public License Version
* 1.1 (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
* http://www.mozilla.org/MPL/
*
* Software distributed under the License is distributed on an "AS IS" basis,
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
* for the specific language governing rights and limitations under the
* License.
*
* The Original Code is [Open Source Virtual Machine].
*
* The Initial Developer of the Original Code is
* Adobe System Incorporated.
* Portions created by the Initial Developer are Copyright (C) 2004-2007
* the Initial Developer. All Rights Reserved.
*
* Contributor(s):
* Adobe AS3 Team
*
* Alternatively, the contents of this file may be used under the terms of
* either the GNU General Public License Version 2 or later (the "GPL"), or
* the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
* in which case the provisions of the GPL or the LGPL are applicable instead
* of those above. If you wish to allow use of your version of this file only
* under the terms of either the GPL or the LGPL, and not to allow others to
* use your version of this file under the terms of the MPL, indicate your
* decision by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL or the LGPL. If you do not delete
* the provisions above, a recipient may use your version of this file under
* the terms of any one of the MPL, the GPL or the LGPL.
*
* ***** END LICENSE BLOCK ***** */
#ifndef __nanojit_Nativei386__
#define __nanojit_Nativei386__
#ifdef PERFM
#define DOPROF
#include "../vprof/vprof.h"
#define count_instr() _nvprof("x86",1)
#define count_ret() _nvprof("x86-ret",1); count_instr();
#define count_push() _nvprof("x86-push",1); count_instr();
#define count_pop() _nvprof("x86-pop",1); count_instr();
#define count_st() _nvprof("x86-st",1); count_instr();
#define count_stq() _nvprof("x86-stq",1); count_instr();
#define count_ld() _nvprof("x86-ld",1); count_instr();
#define count_ldq() _nvprof("x86-ldq",1); count_instr();
#define count_call() _nvprof("x86-call",1); count_instr();
#define count_calli() _nvprof("x86-calli",1); count_instr();
#define count_prolog() _nvprof("x86-prolog",1); count_instr();
#define count_alu() _nvprof("x86-alu",1); count_instr();
#define count_mov() _nvprof("x86-mov",1); count_instr();
#define count_fpu() _nvprof("x86-fpu",1); count_instr();
#define count_jmp() _nvprof("x86-jmp",1); count_instr();
#define count_jcc() _nvprof("x86-jcc",1); count_instr();
#define count_fpuld() _nvprof("x86-ldq",1); _nvprof("x86-fpu",1); count_instr()
#define count_aluld() _nvprof("x86-ld",1); _nvprof("x86-alu",1); count_instr()
#define count_alust() _nvprof("x86-ld",1); _nvprof("x86-alu",1); _nvprof("x86-st",1); count_instr()
#define count_pushld() _nvprof("x86-ld",1); _nvprof("x86-push",1); count_instr()
#define count_imt() _nvprof("x86-imt",1) count_instr()
#else
#define count_instr()
#define count_ret()
#define count_push()
#define count_pop()
#define count_st()
#define count_stq()
#define count_ld()
#define count_ldq()
#define count_call()
#define count_calli()
#define count_prolog()
#define count_alu()
#define count_mov()
#define count_fpu()
#define count_jmp()
#define count_jcc()
#define count_fpuld()
#define count_aluld()
#define count_alust()
#define count_pushld()
#define count_imt()
#endif
namespace nanojit
{
const int NJ_MAX_REGISTERS = 24; // gpregs, x87 regs, xmm regs
#define NJ_MAX_STACK_ENTRY 4096
#define NJ_MAX_PARAMETERS 1
#define NJ_USES_QUAD_CONSTANTS 1
#define NJ_JTBL_SUPPORTED 1
#define NJ_EXPANDED_LOADSTORE_SUPPORTED 1
#define NJ_F2I_SUPPORTED 1
#define NJ_SOFTFLOAT_SUPPORTED 0
// Preserve a 16-byte stack alignment, to support the use of
// SSE instructions like MOVDQA (if not by Tamarin itself,
// then by the C functions it calls).
const int NJ_ALIGN_STACK = 16;
const int32_t LARGEST_UNDERRUN_PROT = 32; // largest value passed to underrunProtect
typedef uint8_t NIns;
// Bytes of icache to flush after patch
const size_t LARGEST_BRANCH_PATCH = 16 * sizeof(NIns);
// These are used as register numbers in various parts of the code
typedef enum
{
// general purpose 32bit regs
EAX = 0, // return value, scratch
ECX = 1, // this/arg0, scratch
EDX = 2, // arg1, return-msw, scratch
EBX = 3,
ESP = 4, // stack pointer
EBP = 5, // frame pointer
ESI = 6,
EDI = 7,
SP = ESP, // alias SP to ESP for convenience
FP = EBP, // alias FP to EBP for convenience
// SSE regs come before X87 so we prefer them
XMM0 = 8,
XMM1 = 9,
XMM2 = 10,
XMM3 = 11,
XMM4 = 12,
XMM5 = 13,
XMM6 = 14,
XMM7 = 15,
// X87 regs
FST0 = 16,
FirstReg = 0,
LastReg = 16,
deprecated_UnknownReg = 17, // XXX: remove eventually, see bug 538924
UnspecifiedReg = 17
}
Register;
typedef int RegisterMask;
static const int NumSavedRegs = 3;
static const RegisterMask SavedRegs = 1<<EBX | 1<<EDI | 1<<ESI;
static const RegisterMask GpRegs = SavedRegs | 1<<EAX | 1<<ECX | 1<<EDX;
static const RegisterMask XmmRegs = 1<<XMM0|1<<XMM1|1<<XMM2|1<<XMM3|1<<XMM4|1<<XMM5|1<<XMM6|1<<XMM7;
static const RegisterMask x87Regs = 1<<FST0;
static const RegisterMask FpRegs = x87Regs | XmmRegs;
static const RegisterMask ScratchRegs = 1<<EAX | 1<<ECX | 1<<EDX | FpRegs;
static const RegisterMask AllowableFlagRegs = 1<<EAX |1<<ECX | 1<<EDX | 1<<EBX;
#define _rmask_(r) (1<<(r))
#define _is_xmm_reg_(r) ((_rmask_(r)&XmmRegs)!=0)
#define _is_x87_reg_(r) ((_rmask_(r)&x87Regs)!=0)
#define _is_fp_reg_(r) ((_rmask_(r)&FpRegs)!=0)
#define _is_gp_reg_(r) ((_rmask_(r)&GpRegs)!=0)
verbose_only( extern const char* regNames[]; )
#define DECLARE_PLATFORM_STATS()
#define DECLARE_PLATFORM_REGALLOC()
#define JCC32 0x0f
#define JMP8 0xeb
#define JMP32 0xe9
#define DECLARE_PLATFORM_ASSEMBLER() \
const static Register argRegs[2], retRegs[2]; \
int32_t max_stk_args;\
void nativePageReset();\
void nativePageSetup();\
void underrunProtect(int);\
void asm_immi(Register r, int32_t val, bool canClobberCCs);\
void asm_stkarg(LInsp p, int32_t& stkd);\
void asm_farg(LInsp, int32_t& stkd);\
void asm_arg(ArgType ty, LInsp p, Register r, int32_t& stkd);\
void asm_pusharg(LInsp);\
void asm_fcmp(LIns *cond);\
NIns* asm_fbranch(bool, LIns*, NIns*);\
void asm_cmp(LIns *cond); \
void asm_div_mod(LIns *cond); \
void asm_load(int d, Register r); \
void asm_immf(Register r, uint64_t q, double d, bool canClobberCCs); \
void IMM8(int32_t i) { \
_nIns -= 1; \
*((int8_t*)_nIns) = (int8_t)(i); \
}; \
void IMM16(int32_t i) { \
_nIns -= 2; \
*((int16_t*)_nIns) = (int16_t)(i); \
}; \
void IMM32(int32_t i) { \
_nIns -= 4; \
*((int32_t*)_nIns) = (int32_t)(i); \
}; \
void MODRMs(int32_t r, int32_t d, Register b, int32_t l, int32_t i); \
void MODRMm(int32_t r, int32_t d, Register b); \
void MODRMSIB(Register reg, Register base, int32_t index, int32_t scale, int32_t disp); \
void MODRMdm(int32_t r, int32_t addr); \
void MODRM(int32_t d, int32_t s) { \
NanoAssert((unsigned(d))<8 && (unsigned(s))<8); \
*(--_nIns) = (uint8_t) ( 3<<6 | d<<3 | s ); \
}; \
void ALU0(int32_t o); \
void ALUm(int32_t c, int32_t r, int32_t d, Register b); \
void ALUdm(int32_t c, int32_t r, int32_t addr); \
void ALUsib(int32_t c, Register r, Register base, int32_t index, int32_t scale, int32_t disp); \
void ALUm16(int32_t c, int32_t r, int32_t d, Register b); \
void ALU2dm(int32_t c, int32_t r, int32_t addr); \
void ALU2m(int32_t c, int32_t r, int32_t d, Register b); \
void ALU2sib(int32_t c, Register r, Register base, int32_t index, int32_t scale, int32_t disp); \
void ALU(int32_t c, int32_t d, int32_t s) { \
underrunProtect(2); \
MODRM(d,s); \
*(--_nIns) = uint8_t(c); \
}; \
void ALUi(int32_t c, int32_t r, int32_t i); \
void ALUmi(int32_t c, int32_t d, Register b, int32_t i); \
void ALU2(int32_t c, int32_t d, int32_t s); \
void LAHF(); \
void SAHF(); \
void OR(Register l, Register r); \
void AND(Register l, Register r); \
void XOR(Register l, Register r); \
void ADD(Register l, Register r); \
void SUB(Register l, Register r); \
void MUL(Register l, Register r); \
void DIV(Register r); \
void NOT(Register r); \
void NEG(Register r); \
void SHR(Register r, Register s); \
void SAR(Register r, Register s); \
void SHL(Register r, Register s); \
void SHIFT(int32_t c, Register r, int32_t i); \
void SHLi(Register r, int32_t i); \
void SHRi(Register r, int32_t i); \
void SARi(Register r, int32_t i); \
void MOVZX8(Register d, Register s); \
void SUBi(Register r, int32_t i); \
void ADDi(Register r, int32_t i); \
void ANDi(Register r, int32_t i); \
void ORi(Register r, int32_t i); \
void XORi(Register r, int32_t i); \
void ADDmi(int32_t d, Register b, int32_t i); \
void TEST(Register d, Register s); \
void CMP(Register l, Register r); \
void CMPi(Register r, int32_t i); \
void MR(Register d, Register s) { \
count_mov(); \
ALU(0x8b,d,s); \
asm_output("mov %s,%s",gpn(d),gpn(s)); \
}; \
void LEA(Register r, int32_t d, Register b); \
void LEAmi4(Register r, int32_t d, int32_t i); \
void CDQ(); \
void INCLi(int32_t p); \
void SETE( Register r); \
void SETNP(Register r); \
void SETL( Register r); \
void SETLE(Register r); \
void SETG( Register r); \
void SETGE(Register r); \
void SETB( Register r); \
void SETBE(Register r); \
void SETA( Register r); \
void SETAE(Register r); \
void SETO( Register r); \
void MREQ(Register d, Register s); \
void MRNE(Register d, Register s); \
void MRL( Register d, Register s); \
void MRLE(Register d, Register s); \
void MRG( Register d, Register s); \
void MRGE(Register d, Register s); \
void MRB( Register d, Register s); \
void MRBE(Register d, Register s); \
void MRA( Register d, Register s); \
void MRAE(Register d, Register s); \
void MRNO(Register d, Register s); \
void LD(Register reg, int32_t disp, Register base); \
void LDdm(Register reg, int32_t addr); \
void LDsib(Register reg, int32_t disp, Register base, int32_t index, int32_t scale); \
void LD16S(Register r, int32_t d, Register b); \
void LD16Sdm(Register r, int32_t addr); \
void LD16Ssib(Register r, int32_t disp, Register base, int32_t index, int32_t scale); \
void LD16Z(Register r, int32_t d, Register b); \
void LD16Zdm(Register r, int32_t addr); \
void LD16Zsib(Register r, int32_t disp, Register base, int32_t index, int32_t scale); \
void LD8Z(Register r, int32_t d, Register b); \
void LD8Zdm(Register r, int32_t addr); \
void LD8Zsib(Register r, int32_t disp, Register base, int32_t ndex, int32_t scale); \
void LD8S(Register r, int32_t d, Register b); \
void LD8Sdm(Register r, int32_t addr); \
void LD8Ssib(Register r, int32_t disp, Register base, int32_t index, int32_t scale); \
void LDi(Register r, int32_t i); \
void ST8(Register base, int32_t disp, Register reg); \
void ST16(Register base, int32_t disp, Register reg); \
void ST(Register base, int32_t disp, Register reg); \
void ST8i(Register base, int32_t disp, int32_t imm); \
void ST16i(Register base, int32_t disp, int32_t imm); \
void STi(Register base, int32_t disp, int32_t imm); \
void RET(); \
void NOP(); \
void INT3(); \
void PUSHi(int32_t i); \
void PUSHi32(int32_t i); \
void PUSHr(Register r); \
void PUSHm(int32_t d, Register b); \
void POPr(Register r); \
void JCC(int32_t o, NIns* t, const char* n); \
void JMP_long(NIns* t); \
void JMP(NIns* t) { \
count_jmp(); \
underrunProtect(5); \
intptr_t tt = t ? (intptr_t)t - (intptr_t)_nIns : 0; \
if (t && isS8(tt)) { \
_nIns -= 2; \
_nIns[0] = JMP8; \
_nIns[1] = uint8_t(tt & 0xff); \
} else { \
IMM32(tt); \
*(--_nIns) = JMP32; \
} \
asm_output("jmp %p", t); \
}; \
void JMP_indirect(Register r); \
void JMP_indexed(Register x, int32_t ss, NIns** addr); \
void JE(NIns* t); \
void JNE(NIns* t); \
void JP(NIns* t); \
void JNP(NIns* t); \
void JB(NIns* t); \
void JNB(NIns* t); \
void JBE(NIns* t); \
void JNBE(NIns* t); \
void JA(NIns* t); \
void JNA(NIns* t); \
void JAE(NIns* t); \
void JNAE(NIns* t); \
void JL(NIns* t); \
void JNL(NIns* t); \
void JLE(NIns* t); \
void JNLE(NIns* t); \
void JG(NIns* t); \
void JNG(NIns* t); \
void JGE(NIns* t); \
void JNGE(NIns* t); \
void JO(NIns* t); \
void JNO(NIns* t); \
void SSE(int32_t c, int32_t d, int32_t s); \
void SSEm(int32_t c, int32_t r, int32_t d, Register b); \
void LDSDm(Register r, const double* addr); \
void SSE_LDSD(Register r, int32_t d, Register b); \
void SSE_LDQ( Register r, int32_t d, Register b); \
void SSE_LDSS(Register r, int32_t d, Register b); \
void SSE_STSD(int32_t d, Register b, Register r); \
void SSE_STQ( int32_t d, Register b, Register r); \
void SSE_STSS(int32_t d, Register b, Register r); \
void SSE_CVTSI2SD(Register xr, Register gr); \
void SSE_CVTSD2SI(Register gr, Register xr); \
void SSE_CVTSD2SS(Register xr, Register gr); \
void SSE_CVTSS2SD(Register xr, Register gr); \
void SSE_CVTDQ2PD(Register d, Register r); \
void SSE_MOVD(Register d, Register s); \
void SSE_MOVSD(Register rd, Register rs); \
void SSE_MOVDm(Register d, Register b, Register xrs); \
void SSE_ADDSD(Register rd, Register rs); \
void SSE_ADDSDm(Register r, const double* addr); \
void SSE_SUBSD(Register rd, Register rs); \
void SSE_MULSD(Register rd, Register rs); \
void SSE_DIVSD(Register rd, Register rs); \
void SSE_UCOMISD(Register rl, Register rr); \
void SSE_CVTSI2SDm(Register xr, Register d, Register b); \
void SSE_XORPD(Register r, const uint32_t* maskaddr); \
void SSE_XORPDr(Register rd, Register rs); \
void FPUc(int32_t o); \
void FPU(int32_t o, Register r) { \
underrunProtect(2); \
*(--_nIns) = uint8_t(((uint8_t)(o)&0xff) | (r&7)); \
*(--_nIns) = (uint8_t)((o>>8)&0xff); \
}; \
void FPUm(int32_t o, int32_t d, Register b); \
void FPUdm(int32_t o, const double* const m); \
void TEST_AH(int32_t i); \
void TEST_AX(int32_t i); \
void FNSTSW_AX(); \
void FCHS(); \
void FLD1(); \
void FLDZ(); \
void FFREE(Register r); \
void FST32(bool p, int32_t d, Register b); \
void FSTQ(bool p, int32_t d, Register b); \
void FSTPQ(int32_t d, Register b); \
void FCOM(bool p, int32_t d, Register b); \
void FCOMdm(bool p, const double* dm); \
void FLD32(int32_t d, Register b); \
void FLDQ(int32_t d, Register b); \
void FLDQdm(const double* dm); \
void FILDQ(int32_t d, Register b); \
void FILD(int32_t d, Register b); \
void FIST(bool p, int32_t d, Register b); \
void FADD( int32_t d, Register b); \
void FSUB( int32_t d, Register b); \
void FSUBR(int32_t d, Register b); \
void FMUL( int32_t d, Register b); \
void FDIV( int32_t d, Register b); \
void FDIVR(int32_t d, Register b); \
void FADDdm( const double *dm); \
void FSUBRdm(const double* dm); \
void FMULdm( const double* dm); \
void FDIVRdm(const double* dm); \
void FINCSTP(); \
void FSTP(Register r) { \
count_fpu(); \
FPU(0xddd8, r); \
asm_output("fstp %s",gpn(r)); fpu_pop(); \
}; \
void FCOMP(); \
void FCOMPP(); \
void FLDr(Register r); \
void EMMS(); \
void CALL(const CallInfo* ci); \
void CALLr(const CallInfo* ci, Register r);
}
#endif // __nanojit_Nativei386__