Merge pull request #9446 from Dentomologist/convert_shifttype_to_enum_class
Arm64Emitter: Convert ShiftType to enum class
This commit is contained in:
commit
04ccd4cb80
|
@ -1263,7 +1263,7 @@ void ARM64XEmitter::ISB(BarrierType type)
|
|||
// Add/Subtract (extended register)
|
||||
void ARM64XEmitter::ADD(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
ADD(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0));
|
||||
ADD(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
|
||||
void ARM64XEmitter::ADD(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm, ArithOption Option)
|
||||
|
@ -1273,7 +1273,7 @@ void ARM64XEmitter::ADD(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm, ArithOption Optio
|
|||
|
||||
void ARM64XEmitter::ADDS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
EncodeArithmeticInst(0, true, Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0));
|
||||
EncodeArithmeticInst(0, true, Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
|
||||
void ARM64XEmitter::ADDS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm, ArithOption Option)
|
||||
|
@ -1283,7 +1283,7 @@ void ARM64XEmitter::ADDS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm, ArithOption Opti
|
|||
|
||||
void ARM64XEmitter::SUB(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
SUB(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0));
|
||||
SUB(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
|
||||
void ARM64XEmitter::SUB(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm, ArithOption Option)
|
||||
|
@ -1293,7 +1293,7 @@ void ARM64XEmitter::SUB(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm, ArithOption Optio
|
|||
|
||||
void ARM64XEmitter::SUBS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
EncodeArithmeticInst(1, true, Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0));
|
||||
EncodeArithmeticInst(1, true, Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
|
||||
void ARM64XEmitter::SUBS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm, ArithOption Option)
|
||||
|
@ -1303,7 +1303,7 @@ void ARM64XEmitter::SUBS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm, ArithOption Opti
|
|||
|
||||
void ARM64XEmitter::CMN(ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
CMN(Rn, Rm, ArithOption(Rn, ST_LSL, 0));
|
||||
CMN(Rn, Rm, ArithOption(Rn, ShiftType::LSL, 0));
|
||||
}
|
||||
|
||||
void ARM64XEmitter::CMN(ARM64Reg Rn, ARM64Reg Rm, ArithOption Option)
|
||||
|
@ -1313,7 +1313,7 @@ void ARM64XEmitter::CMN(ARM64Reg Rn, ARM64Reg Rm, ArithOption Option)
|
|||
|
||||
void ARM64XEmitter::CMP(ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
CMP(Rn, Rm, ArithOption(Rn, ST_LSL, 0));
|
||||
CMP(Rn, Rm, ArithOption(Rn, ShiftType::LSL, 0));
|
||||
}
|
||||
|
||||
void ARM64XEmitter::CMP(ARM64Reg Rn, ARM64Reg Rm, ArithOption Option)
|
||||
|
@ -1553,13 +1553,13 @@ void ARM64XEmitter::MOV(ARM64Reg Rd, ARM64Reg Rm, ArithOption Shift)
|
|||
void ARM64XEmitter::MOV(ARM64Reg Rd, ARM64Reg Rm)
|
||||
{
|
||||
if (IsGPR(Rd) && IsGPR(Rm))
|
||||
ORR(Rd, Is64Bit(Rd) ? ZR : WZR, Rm, ArithOption(Rm, ST_LSL, 0));
|
||||
ORR(Rd, Is64Bit(Rd) ? ZR : WZR, Rm, ArithOption(Rm, ShiftType::LSL, 0));
|
||||
else
|
||||
ASSERT_MSG(DYNA_REC, false, "Non-GPRs not supported in MOV");
|
||||
}
|
||||
void ARM64XEmitter::MVN(ARM64Reg Rd, ARM64Reg Rm)
|
||||
{
|
||||
ORN(Rd, Is64Bit(Rd) ? ZR : WZR, Rm, ArithOption(Rm, ST_LSL, 0));
|
||||
ORN(Rd, Is64Bit(Rd) ? ZR : WZR, Rm, ArithOption(Rm, ShiftType::LSL, 0));
|
||||
}
|
||||
void ARM64XEmitter::LSL(ARM64Reg Rd, ARM64Reg Rm, int shift)
|
||||
{
|
||||
|
@ -2016,7 +2016,7 @@ void ARM64XEmitter::MOVI2R(ARM64Reg Rd, u64 imm, bool optimize)
|
|||
// Max unsigned value (or if signed, -1)
|
||||
// Set to ~ZR
|
||||
ARM64Reg ZR = Is64Bit(Rd) ? SP : WSP;
|
||||
ORN(Rd, ZR, ZR, ArithOption(ZR, ST_LSL, 0));
|
||||
ORN(Rd, ZR, ZR, ArithOption(ZR, ShiftType::LSL, 0));
|
||||
return;
|
||||
}
|
||||
|
||||
|
|
|
@ -277,12 +277,16 @@ constexpr ARM64Reg EncodeRegToQuad(ARM64Reg reg)
|
|||
return static_cast<ARM64Reg>(reg | 0xC0);
|
||||
}
|
||||
|
||||
enum ShiftType
|
||||
enum class ShiftType
|
||||
{
|
||||
ST_LSL = 0,
|
||||
ST_LSR = 1,
|
||||
ST_ASR = 2,
|
||||
ST_ROR = 3,
|
||||
// Logical Shift Left
|
||||
LSL = 0,
|
||||
// Logical Shift Right
|
||||
LSR = 1,
|
||||
// Arithmetic Shift Right
|
||||
ASR = 2,
|
||||
// Rotate Right
|
||||
ROR = 3,
|
||||
};
|
||||
|
||||
enum class IndexType
|
||||
|
@ -437,7 +441,7 @@ public:
|
|||
m_width = WidthSpecifier::Width32Bit;
|
||||
m_extend = ExtendSpecifier::UXTW;
|
||||
}
|
||||
m_shifttype = ST_LSL;
|
||||
m_shifttype = ShiftType::LSL;
|
||||
}
|
||||
ArithOption(ARM64Reg Rd, ShiftType shift_type, u32 shift)
|
||||
{
|
||||
|
@ -466,7 +470,7 @@ public:
|
|||
case TypeSpecifier::ExtendedReg:
|
||||
return (static_cast<u32>(m_extend) << 13) | (m_shift << 10);
|
||||
case TypeSpecifier::ShiftedReg:
|
||||
return (m_shifttype << 22) | (m_shift << 10);
|
||||
return (static_cast<u32>(m_shifttype) << 22) | (m_shift << 10);
|
||||
default:
|
||||
DEBUG_ASSERT_MSG(DYNA_REC, false, "Invalid type in GetData");
|
||||
break;
|
||||
|
@ -699,14 +703,38 @@ public:
|
|||
void BICS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm, ArithOption Shift);
|
||||
|
||||
// Wrap the above for saner syntax
|
||||
void AND(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm) { AND(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0)); }
|
||||
void BIC(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm) { BIC(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0)); }
|
||||
void ORR(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm) { ORR(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0)); }
|
||||
void ORN(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm) { ORN(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0)); }
|
||||
void EOR(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm) { EOR(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0)); }
|
||||
void EON(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm) { EON(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0)); }
|
||||
void ANDS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm) { ANDS(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0)); }
|
||||
void BICS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm) { BICS(Rd, Rn, Rm, ArithOption(Rd, ST_LSL, 0)); }
|
||||
void AND(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
AND(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
void BIC(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
BIC(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
void ORR(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
ORR(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
void ORN(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
ORN(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
void EOR(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
EOR(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
void EON(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
EON(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
void ANDS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
ANDS(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
void BICS(ARM64Reg Rd, ARM64Reg Rn, ARM64Reg Rm)
|
||||
{
|
||||
BICS(Rd, Rn, Rm, ArithOption(Rd, ShiftType::LSL, 0));
|
||||
}
|
||||
// Convenience wrappers around ORR. These match the official convenience syntax.
|
||||
void MOV(ARM64Reg Rd, ARM64Reg Rm, ArithOption Shift);
|
||||
void MOV(ARM64Reg Rd, ARM64Reg Rm);
|
||||
|
|
|
@ -567,7 +567,7 @@ void JitArm64::rlwinmx(UGeckoInstruction inst)
|
|||
{
|
||||
ARM64Reg WA = gpr.GetReg();
|
||||
MOVI2R(WA, mask);
|
||||
AND(gpr.R(a), WA, gpr.R(s), ArithOption(gpr.R(s), ST_ROR, 32 - inst.SH));
|
||||
AND(gpr.R(a), WA, gpr.R(s), ArithOption(gpr.R(s), ShiftType::ROR, 32 - inst.SH));
|
||||
gpr.Unlock(WA);
|
||||
}
|
||||
|
||||
|
@ -592,7 +592,7 @@ void JitArm64::rlwnmx(UGeckoInstruction inst)
|
|||
{
|
||||
gpr.BindToRegister(a, a == s);
|
||||
ARM64Reg WA = gpr.GetReg();
|
||||
ArithOption Shift(gpr.R(s), ST_ROR, 32 - (gpr.GetImm(b) & 0x1f));
|
||||
ArithOption Shift(gpr.R(s), ShiftType::ROR, 32 - (gpr.GetImm(b) & 0x1f));
|
||||
MOVI2R(WA, mask);
|
||||
AND(gpr.R(a), WA, gpr.R(s), Shift);
|
||||
gpr.Unlock(WA);
|
||||
|
@ -656,7 +656,7 @@ void JitArm64::srawix(UGeckoInstruction inst)
|
|||
if (a != s)
|
||||
{
|
||||
ASR(RA, RS, amount);
|
||||
ANDS(dest, RA, RS, ArithOption(RS, ST_LSL, 32 - amount));
|
||||
ANDS(dest, RA, RS, ArithOption(RS, ShiftType::LSL, 32 - amount));
|
||||
}
|
||||
else
|
||||
{
|
||||
|
@ -1500,7 +1500,7 @@ void JitArm64::rlwimix(UGeckoInstruction inst)
|
|||
|
||||
MOVI2R(WA, mask);
|
||||
BIC(WB, gpr.R(a), WA);
|
||||
AND(WA, WA, gpr.R(s), ArithOption(gpr.R(s), ST_ROR, 32 - inst.SH));
|
||||
AND(WA, WA, gpr.R(s), ArithOption(gpr.R(s), ShiftType::ROR, 32 - inst.SH));
|
||||
ORR(gpr.R(a), WB, WA);
|
||||
|
||||
gpr.Unlock(WA, WB);
|
||||
|
|
|
@ -91,7 +91,7 @@ void JitArm64::mcrxr(UGeckoInstruction inst)
|
|||
LDRB(IndexType::Unsigned, WB, PPC_REG, PPCSTATE_OFF(xer_so_ov));
|
||||
|
||||
// [0 SO OV CA]
|
||||
ADD(WA, WA, WB, ArithOption(WB, ST_LSL, 2));
|
||||
ADD(WA, WA, WB, ArithOption(WB, ShiftType::LSL, 2));
|
||||
// [SO OV CA 0] << 3
|
||||
LSL(WA, WA, 4);
|
||||
|
||||
|
@ -136,7 +136,7 @@ void JitArm64::mfsrin(UGeckoInstruction inst)
|
|||
ARM64Reg RB = gpr.R(b);
|
||||
|
||||
UBFM(index, RB, 28, 31);
|
||||
ADD(index64, PPC_REG, index64, ArithOption(index64, ST_LSL, 2));
|
||||
ADD(index64, PPC_REG, index64, ArithOption(index64, ShiftType::LSL, 2));
|
||||
LDR(IndexType::Unsigned, gpr.R(d), index64, PPCSTATE_OFF(sr[0]));
|
||||
|
||||
gpr.Unlock(index);
|
||||
|
@ -155,7 +155,7 @@ void JitArm64::mtsrin(UGeckoInstruction inst)
|
|||
ARM64Reg RB = gpr.R(b);
|
||||
|
||||
UBFM(index, RB, 28, 31);
|
||||
ADD(index64, PPC_REG, index64, ArithOption(index64, ST_LSL, 2));
|
||||
ADD(index64, PPC_REG, index64, ArithOption(index64, ShiftType::LSL, 2));
|
||||
STR(IndexType::Unsigned, gpr.R(d), index64, PPCSTATE_OFF(sr[0]));
|
||||
|
||||
gpr.Unlock(index);
|
||||
|
@ -282,7 +282,7 @@ void JitArm64::mfspr(UGeckoInstruction inst)
|
|||
ADD(XB, XB, 1);
|
||||
UMULH(Xresult, Xresult, XB);
|
||||
|
||||
ADD(Xresult, XA, Xresult, ArithOption(Xresult, ST_LSR, 3));
|
||||
ADD(Xresult, XA, Xresult, ArithOption(Xresult, ShiftType::LSR, 3));
|
||||
STR(IndexType::Unsigned, Xresult, PPC_REG, PPCSTATE_OFF(spr[SPR_TL]));
|
||||
|
||||
if (CanMergeNextInstructions(1))
|
||||
|
@ -332,9 +332,9 @@ void JitArm64::mfspr(UGeckoInstruction inst)
|
|||
ARM64Reg WA = gpr.GetReg();
|
||||
LDRH(IndexType::Unsigned, RD, PPC_REG, PPCSTATE_OFF(xer_stringctrl));
|
||||
LDRB(IndexType::Unsigned, WA, PPC_REG, PPCSTATE_OFF(xer_ca));
|
||||
ORR(RD, RD, WA, ArithOption(WA, ST_LSL, XER_CA_SHIFT));
|
||||
ORR(RD, RD, WA, ArithOption(WA, ShiftType::LSL, XER_CA_SHIFT));
|
||||
LDRB(IndexType::Unsigned, WA, PPC_REG, PPCSTATE_OFF(xer_so_ov));
|
||||
ORR(RD, RD, WA, ArithOption(WA, ST_LSL, XER_OV_SHIFT));
|
||||
ORR(RD, RD, WA, ArithOption(WA, ShiftType::LSL, XER_OV_SHIFT));
|
||||
gpr.Unlock(WA);
|
||||
}
|
||||
break;
|
||||
|
@ -633,7 +633,7 @@ void JitArm64::mfcr(UGeckoInstruction inst)
|
|||
else
|
||||
{
|
||||
UBFX(XC, CR, 61, 1);
|
||||
ORR(XA, XC, XA, ArithOption(XA, ST_LSL, 4));
|
||||
ORR(XA, XC, XA, ArithOption(XA, ShiftType::LSL, 4));
|
||||
}
|
||||
|
||||
// EQ
|
||||
|
@ -648,7 +648,7 @@ void JitArm64::mfcr(UGeckoInstruction inst)
|
|||
|
||||
// LT
|
||||
UBFX(XC, CR, 62, 1);
|
||||
ORR(WA, WA, WC, ArithOption(WC, ST_LSL, 3));
|
||||
ORR(WA, WA, WC, ArithOption(WC, ShiftType::LSL, 3));
|
||||
}
|
||||
|
||||
gpr.Unlock(WC);
|
||||
|
|
|
@ -97,7 +97,7 @@ void JitArm64::GenerateAsm()
|
|||
ARM64Reg cache_base = X27;
|
||||
ARM64Reg block = X30;
|
||||
ORRI2R(pc_masked, WZR, JitBaseBlockCache::FAST_BLOCK_MAP_MASK << 3);
|
||||
AND(pc_masked, pc_masked, DISPATCHER_PC, ArithOption(DISPATCHER_PC, ST_LSL, 1));
|
||||
AND(pc_masked, pc_masked, DISPATCHER_PC, ArithOption(DISPATCHER_PC, ShiftType::LSL, 1));
|
||||
MOVP2R(cache_base, GetBlockCache()->GetFastBlockMap());
|
||||
LDR(block, cache_base, EncodeRegTo64(pc_masked));
|
||||
FixupBranch not_found = CBZ(block);
|
||||
|
@ -224,7 +224,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
float_emit.UCVTF(32, D0, D0);
|
||||
|
||||
MOVP2R(addr_reg, &m_dequantizeTableS);
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
RET(X30);
|
||||
|
@ -238,7 +238,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
float_emit.SCVTF(32, D0, D0);
|
||||
|
||||
MOVP2R(addr_reg, &m_dequantizeTableS);
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
RET(X30);
|
||||
|
@ -252,7 +252,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
float_emit.UCVTF(32, D0, D0);
|
||||
|
||||
MOVP2R(addr_reg, &m_dequantizeTableS);
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
RET(X30);
|
||||
|
@ -266,7 +266,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
float_emit.SCVTF(32, D0, D0);
|
||||
|
||||
MOVP2R(addr_reg, &m_dequantizeTableS);
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
RET(X30);
|
||||
|
@ -288,7 +288,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
float_emit.UCVTF(32, D0, D0);
|
||||
|
||||
MOVP2R(addr_reg, &m_dequantizeTableS);
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
RET(X30);
|
||||
|
@ -302,7 +302,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
float_emit.SCVTF(32, D0, D0);
|
||||
|
||||
MOVP2R(addr_reg, &m_dequantizeTableS);
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
RET(X30);
|
||||
|
@ -316,7 +316,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
float_emit.UCVTF(32, D0, D0);
|
||||
|
||||
MOVP2R(addr_reg, &m_dequantizeTableS);
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
RET(X30);
|
||||
|
@ -330,7 +330,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
float_emit.SCVTF(32, D0, D0);
|
||||
|
||||
MOVP2R(addr_reg, &m_dequantizeTableS);
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, addr_reg, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
RET(X30);
|
||||
|
@ -387,7 +387,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
{
|
||||
auto emit_quantize = [this, &float_emit, scale_reg]() {
|
||||
MOVP2R(X2, &m_quantizeTableS);
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
|
||||
|
@ -414,7 +414,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
{
|
||||
auto emit_quantize = [this, &float_emit, scale_reg]() {
|
||||
MOVP2R(X2, &m_quantizeTableS);
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
|
||||
|
@ -442,7 +442,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
{
|
||||
auto emit_quantize = [this, &float_emit, scale_reg]() {
|
||||
MOVP2R(X2, &m_quantizeTableS);
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
|
||||
|
@ -469,7 +469,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
{
|
||||
auto emit_quantize = [this, &float_emit, scale_reg]() {
|
||||
MOVP2R(X2, &m_quantizeTableS);
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1, 0);
|
||||
|
||||
|
@ -511,7 +511,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
{
|
||||
auto emit_quantize = [this, &float_emit, scale_reg]() {
|
||||
MOVP2R(X2, &m_quantizeTableS);
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1);
|
||||
|
||||
|
@ -537,7 +537,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
{
|
||||
auto emit_quantize = [this, &float_emit, scale_reg]() {
|
||||
MOVP2R(X2, &m_quantizeTableS);
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1);
|
||||
|
||||
|
@ -563,7 +563,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
{
|
||||
auto emit_quantize = [this, &float_emit, scale_reg]() {
|
||||
MOVP2R(X2, &m_quantizeTableS);
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1);
|
||||
|
||||
|
@ -589,7 +589,7 @@ void JitArm64::GenerateCommonAsm()
|
|||
{
|
||||
auto emit_quantize = [this, &float_emit, scale_reg]() {
|
||||
MOVP2R(X2, &m_quantizeTableS);
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ST_LSL, 3));
|
||||
ADD(scale_reg, X2, scale_reg, ArithOption(scale_reg, ShiftType::LSL, 3));
|
||||
float_emit.LDR(32, IndexType::Unsigned, D1, scale_reg, 0);
|
||||
float_emit.FMUL(32, D0, D0, D1);
|
||||
|
||||
|
|
|
@ -195,7 +195,7 @@ int VertexLoaderARM64::ReadVertex(u64 attribute, int format, int count_in, int c
|
|||
FixupBranch dont_store = B(CC_GT);
|
||||
MOVP2R(EncodeRegTo64(scratch2_reg), VertexLoaderManager::position_cache);
|
||||
ADD(EncodeRegTo64(scratch1_reg), EncodeRegTo64(scratch2_reg), EncodeRegTo64(count_reg),
|
||||
ArithOption(EncodeRegTo64(count_reg), ST_LSL, 4));
|
||||
ArithOption(EncodeRegTo64(count_reg), ShiftType::LSL, 4));
|
||||
m_float_emit.STUR(write_size, coords, EncodeRegTo64(scratch1_reg), -16);
|
||||
SetJumpTarget(dont_store);
|
||||
}
|
||||
|
@ -248,20 +248,20 @@ void VertexLoaderARM64::ReadColor(u64 attribute, int format, s32 offset)
|
|||
|
||||
// B
|
||||
AND(scratch2_reg, scratch3_reg, 32, 4);
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 3));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSR, 5));
|
||||
ORR(scratch1_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 16));
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 3));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSR, 5));
|
||||
ORR(scratch1_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 16));
|
||||
|
||||
// G
|
||||
UBFM(scratch2_reg, scratch3_reg, 5, 10);
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 2));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSR, 6));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 8));
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 2));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSR, 6));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 8));
|
||||
|
||||
// R
|
||||
UBFM(scratch2_reg, scratch3_reg, 11, 15);
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 3));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSR, 2));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 3));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSR, 2));
|
||||
|
||||
// A
|
||||
ORRI2R(scratch1_reg, scratch1_reg, 0xFF000000);
|
||||
|
@ -286,18 +286,18 @@ void VertexLoaderARM64::ReadColor(u64 attribute, int format, s32 offset)
|
|||
|
||||
// G
|
||||
AND(scratch2_reg, scratch3_reg, 32, 3);
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 8));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 8));
|
||||
|
||||
// B
|
||||
UBFM(scratch2_reg, scratch3_reg, 12, 15);
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 16));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 16));
|
||||
|
||||
// A
|
||||
UBFM(scratch2_reg, scratch3_reg, 8, 11);
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 24));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 24));
|
||||
|
||||
// Final duplication
|
||||
ORR(scratch1_reg, scratch1_reg, scratch1_reg, ArithOption(scratch1_reg, ST_LSL, 4));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch1_reg, ArithOption(scratch1_reg, ShiftType::LSL, 4));
|
||||
|
||||
STR(IndexType::Unsigned, scratch1_reg, dst_reg, m_dst_ofs);
|
||||
load_bytes = 2;
|
||||
|
@ -323,26 +323,26 @@ void VertexLoaderARM64::ReadColor(u64 attribute, int format, s32 offset)
|
|||
|
||||
// A
|
||||
UBFM(scratch2_reg, scratch3_reg, 0, 5);
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 2));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSR, 6));
|
||||
ORR(scratch1_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 24));
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 2));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSR, 6));
|
||||
ORR(scratch1_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 24));
|
||||
|
||||
// B
|
||||
UBFM(scratch2_reg, scratch3_reg, 6, 11);
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 2));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSR, 6));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 16));
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 2));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSR, 6));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 16));
|
||||
|
||||
// G
|
||||
UBFM(scratch2_reg, scratch3_reg, 12, 17);
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 2));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSR, 6));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 8));
|
||||
ORR(scratch2_reg, WSP, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 2));
|
||||
ORR(scratch2_reg, scratch2_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSR, 6));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 8));
|
||||
|
||||
// R
|
||||
UBFM(scratch2_reg, scratch3_reg, 18, 23);
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSL, 2));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ST_LSR, 4));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSL, 2));
|
||||
ORR(scratch1_reg, scratch1_reg, scratch2_reg, ArithOption(scratch2_reg, ShiftType::LSR, 4));
|
||||
|
||||
STR(IndexType::Unsigned, scratch1_reg, dst_reg, m_dst_ofs);
|
||||
|
||||
|
|
Loading…
Reference in New Issue