Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
using ARMeilleure.Decoders;
|
|
|
|
using ARMeilleure.IntermediateRepresentation;
|
|
|
|
using ARMeilleure.State;
|
|
|
|
using ARMeilleure.Translation;
|
2019-12-29 17:22:47 -08:00
|
|
|
using System.Diagnostics;
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
|
|
|
|
using static ARMeilleure.Instructions.InstEmitAluHelper;
|
|
|
|
using static ARMeilleure.Instructions.InstEmitHelper;
|
|
|
|
using static ARMeilleure.IntermediateRepresentation.OperandHelper;
|
|
|
|
|
|
|
|
namespace ARMeilleure.Instructions
|
|
|
|
{
|
|
|
|
static partial class InstEmit
|
|
|
|
{
|
|
|
|
public static void Adc(ArmEmitterContext context) => EmitAdc(context, setFlags: false);
|
|
|
|
public static void Adcs(ArmEmitterContext context) => EmitAdc(context, setFlags: true);
|
|
|
|
|
|
|
|
private static void EmitAdc(ArmEmitterContext context, bool setFlags)
|
|
|
|
{
|
|
|
|
Operand n = GetAluN(context);
|
|
|
|
Operand m = GetAluM(context);
|
|
|
|
|
|
|
|
Operand d = context.Add(n, m);
|
|
|
|
|
|
|
|
Operand carry = GetFlag(PState.CFlag);
|
|
|
|
|
|
|
|
if (context.CurrOp.RegisterSize == RegisterSize.Int64)
|
|
|
|
{
|
|
|
|
carry = context.ZeroExtend32(OperandType.I64, carry);
|
|
|
|
}
|
|
|
|
|
|
|
|
d = context.Add(d, carry);
|
|
|
|
|
|
|
|
if (setFlags)
|
|
|
|
{
|
|
|
|
EmitNZFlagsCheck(context, d);
|
|
|
|
|
|
|
|
EmitAdcsCCheck(context, n, d);
|
|
|
|
EmitAddsVCheck(context, n, m, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
SetAluDOrZR(context, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Add(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetAluD(context, context.Add(GetAluN(context), GetAluM(context)));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Adds(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
Operand n = GetAluN(context);
|
|
|
|
Operand m = GetAluM(context);
|
|
|
|
|
|
|
|
context.MarkComparison(n, m);
|
|
|
|
|
|
|
|
Operand d = context.Add(n, m);
|
|
|
|
|
|
|
|
EmitNZFlagsCheck(context, d);
|
|
|
|
|
|
|
|
EmitAddsCCheck(context, n, d);
|
|
|
|
EmitAddsVCheck(context, n, m, d);
|
|
|
|
|
|
|
|
SetAluDOrZR(context, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void And(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetAluD(context, context.BitwiseAnd(GetAluN(context), GetAluM(context)));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Ands(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
Operand n = GetAluN(context);
|
|
|
|
Operand m = GetAluM(context);
|
|
|
|
|
|
|
|
Operand d = context.BitwiseAnd(n, m);
|
|
|
|
|
|
|
|
EmitNZFlagsCheck(context, d);
|
|
|
|
EmitCVFlagsClear(context);
|
|
|
|
|
|
|
|
SetAluDOrZR(context, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Asrv(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetAluDOrZR(context, context.ShiftRightSI(GetAluN(context), GetAluMShift(context)));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Bic(ArmEmitterContext context) => EmitBic(context, setFlags: false);
|
|
|
|
public static void Bics(ArmEmitterContext context) => EmitBic(context, setFlags: true);
|
|
|
|
|
|
|
|
private static void EmitBic(ArmEmitterContext context, bool setFlags)
|
|
|
|
{
|
|
|
|
Operand n = GetAluN(context);
|
|
|
|
Operand m = GetAluM(context);
|
|
|
|
|
|
|
|
Operand d = context.BitwiseAnd(n, context.BitwiseNot(m));
|
|
|
|
|
|
|
|
if (setFlags)
|
|
|
|
{
|
|
|
|
EmitNZFlagsCheck(context, d);
|
|
|
|
EmitCVFlagsClear(context);
|
|
|
|
}
|
|
|
|
|
|
|
|
SetAluD(context, d, setFlags);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Cls(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
OpCodeAlu op = (OpCodeAlu)context.CurrOp;
|
|
|
|
|
|
|
|
Operand n = GetIntOrZR(context, op.Rn);
|
|
|
|
|
|
|
|
Operand nHigh = context.ShiftRightUI(n, Const(1));
|
|
|
|
|
|
|
|
bool is32Bits = op.RegisterSize == RegisterSize.Int32;
|
|
|
|
|
|
|
|
Operand mask = is32Bits ? Const(int.MaxValue) : Const(long.MaxValue);
|
|
|
|
|
|
|
|
Operand nLow = context.BitwiseAnd(n, mask);
|
|
|
|
|
|
|
|
Operand res = context.CountLeadingZeros(context.BitwiseExclusiveOr(nHigh, nLow));
|
|
|
|
|
|
|
|
res = context.Subtract(res, Const(res.Type, 1));
|
|
|
|
|
|
|
|
SetAluDOrZR(context, res);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Clz(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
OpCodeAlu op = (OpCodeAlu)context.CurrOp;
|
|
|
|
|
|
|
|
Operand n = GetIntOrZR(context, op.Rn);
|
|
|
|
|
|
|
|
Operand d = context.CountLeadingZeros(n);
|
|
|
|
|
|
|
|
SetAluDOrZR(context, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Eon(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
Operand n = GetAluN(context);
|
|
|
|
Operand m = GetAluM(context);
|
|
|
|
|
|
|
|
Operand d = context.BitwiseExclusiveOr(n, context.BitwiseNot(m));
|
|
|
|
|
|
|
|
SetAluD(context, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Eor(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetAluD(context, context.BitwiseExclusiveOr(GetAluN(context), GetAluM(context)));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Extr(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
OpCodeAluRs op = (OpCodeAluRs)context.CurrOp;
|
|
|
|
|
|
|
|
Operand res = GetIntOrZR(context, op.Rm);
|
|
|
|
|
|
|
|
if (op.Shift != 0)
|
|
|
|
{
|
|
|
|
if (op.Rn == op.Rm)
|
|
|
|
{
|
|
|
|
res = context.RotateRight(res, Const(op.Shift));
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
res = context.ShiftRightUI(res, Const(op.Shift));
|
|
|
|
|
|
|
|
Operand n = GetIntOrZR(context, op.Rn);
|
|
|
|
|
|
|
|
int invShift = op.GetBitsCount() - op.Shift;
|
|
|
|
|
|
|
|
res = context.BitwiseOr(res, context.ShiftLeft(n, Const(invShift)));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
SetAluDOrZR(context, res);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Lslv(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetAluDOrZR(context, context.ShiftLeft(GetAluN(context), GetAluMShift(context)));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Lsrv(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetAluDOrZR(context, context.ShiftRightUI(GetAluN(context), GetAluMShift(context)));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Sbc(ArmEmitterContext context) => EmitSbc(context, setFlags: false);
|
|
|
|
public static void Sbcs(ArmEmitterContext context) => EmitSbc(context, setFlags: true);
|
|
|
|
|
|
|
|
private static void EmitSbc(ArmEmitterContext context, bool setFlags)
|
|
|
|
{
|
|
|
|
Operand n = GetAluN(context);
|
|
|
|
Operand m = GetAluM(context);
|
|
|
|
|
|
|
|
Operand d = context.Subtract(n, m);
|
|
|
|
|
|
|
|
Operand borrow = context.BitwiseExclusiveOr(GetFlag(PState.CFlag), Const(1));
|
|
|
|
|
|
|
|
if (context.CurrOp.RegisterSize == RegisterSize.Int64)
|
|
|
|
{
|
|
|
|
borrow = context.ZeroExtend32(OperandType.I64, borrow);
|
|
|
|
}
|
|
|
|
|
|
|
|
d = context.Subtract(d, borrow);
|
|
|
|
|
|
|
|
if (setFlags)
|
|
|
|
{
|
|
|
|
EmitNZFlagsCheck(context, d);
|
|
|
|
|
|
|
|
EmitSbcsCCheck(context, n, m);
|
|
|
|
EmitSubsVCheck(context, n, m, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
SetAluDOrZR(context, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Sub(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetAluD(context, context.Subtract(GetAluN(context), GetAluM(context)));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Subs(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
Operand n = GetAluN(context);
|
|
|
|
Operand m = GetAluM(context);
|
|
|
|
|
|
|
|
context.MarkComparison(n, m);
|
|
|
|
|
|
|
|
Operand d = context.Subtract(n, m);
|
|
|
|
|
|
|
|
EmitNZFlagsCheck(context, d);
|
|
|
|
|
|
|
|
EmitSubsCCheck(context, n, m);
|
|
|
|
EmitSubsVCheck(context, n, m, d);
|
|
|
|
|
|
|
|
SetAluDOrZR(context, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Orn(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
Operand n = GetAluN(context);
|
|
|
|
Operand m = GetAluM(context);
|
|
|
|
|
|
|
|
Operand d = context.BitwiseOr(n, context.BitwiseNot(m));
|
|
|
|
|
|
|
|
SetAluD(context, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Orr(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetAluD(context, context.BitwiseOr(GetAluN(context), GetAluM(context)));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Rbit(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
OpCodeAlu op = (OpCodeAlu)context.CurrOp;
|
|
|
|
|
|
|
|
Operand n = GetIntOrZR(context, op.Rn);
|
|
|
|
Operand d;
|
|
|
|
|
|
|
|
if (op.RegisterSize == RegisterSize.Int32)
|
|
|
|
{
|
2019-12-29 17:22:47 -08:00
|
|
|
d = EmitReverseBits32Op(context, n);
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2019-12-29 17:22:47 -08:00
|
|
|
d = EmitReverseBits64Op(context, n);
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
SetAluDOrZR(context, d);
|
|
|
|
}
|
|
|
|
|
2019-12-29 17:22:47 -08:00
|
|
|
private static Operand EmitReverseBits32Op(ArmEmitterContext context, Operand op)
|
|
|
|
{
|
|
|
|
Debug.Assert(op.Type == OperandType.I32);
|
|
|
|
|
|
|
|
Operand val = context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(op, Const(0xaaaaaaaau)), Const(1)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(op, Const(0x55555555u)), Const(1)));
|
|
|
|
|
|
|
|
val = context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(val, Const(0xccccccccu)), Const(2)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(val, Const(0x33333333u)), Const(2)));
|
|
|
|
val = context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(val, Const(0xf0f0f0f0u)), Const(4)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(val, Const(0x0f0f0f0fu)), Const(4)));
|
|
|
|
val = context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(val, Const(0xff00ff00u)), Const(8)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(val, Const(0x00ff00ffu)), Const(8)));
|
|
|
|
|
|
|
|
return context.BitwiseOr(context.ShiftRightUI(val, Const(16)), context.ShiftLeft(val, Const(16)));
|
|
|
|
}
|
|
|
|
|
|
|
|
private static Operand EmitReverseBits64Op(ArmEmitterContext context, Operand op)
|
|
|
|
{
|
|
|
|
Debug.Assert(op.Type == OperandType.I64);
|
|
|
|
|
|
|
|
Operand val = context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(op, Const(0xaaaaaaaaaaaaaaaaul)), Const(1)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(op, Const(0x5555555555555555ul)), Const(1)));
|
|
|
|
|
|
|
|
val = context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(val, Const(0xccccccccccccccccul)), Const(2)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(val, Const(0x3333333333333333ul)), Const(2)));
|
|
|
|
val = context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(val, Const(0xf0f0f0f0f0f0f0f0ul)), Const(4)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(val, Const(0x0f0f0f0f0f0f0f0ful)), Const(4)));
|
|
|
|
val = context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(val, Const(0xff00ff00ff00ff00ul)), Const(8)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(val, Const(0x00ff00ff00ff00fful)), Const(8)));
|
|
|
|
val = context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(val, Const(0xffff0000ffff0000ul)), Const(16)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(val, Const(0x0000ffff0000fffful)), Const(16)));
|
|
|
|
|
|
|
|
return context.BitwiseOr(context.ShiftRightUI(val, Const(32)), context.ShiftLeft(val, Const(32)));
|
|
|
|
}
|
|
|
|
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
public static void Rev16(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
OpCodeAlu op = (OpCodeAlu)context.CurrOp;
|
|
|
|
|
|
|
|
Operand n = GetIntOrZR(context, op.Rn);
|
|
|
|
Operand d;
|
|
|
|
|
|
|
|
if (op.RegisterSize == RegisterSize.Int32)
|
|
|
|
{
|
2019-12-29 17:22:47 -08:00
|
|
|
d = EmitReverseBytes16_32Op(context, n);
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2019-12-29 17:22:47 -08:00
|
|
|
d = EmitReverseBytes16_64Op(context, n);
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
SetAluDOrZR(context, d);
|
|
|
|
}
|
|
|
|
|
2019-12-29 17:22:47 -08:00
|
|
|
private static Operand EmitReverseBytes16_32Op(ArmEmitterContext context, Operand op)
|
|
|
|
{
|
|
|
|
Debug.Assert(op.Type == OperandType.I32);
|
|
|
|
|
|
|
|
Operand val = EmitReverseBytes16_64Op(context, context.ZeroExtend32(OperandType.I64, op));
|
|
|
|
|
|
|
|
return context.ConvertI64ToI32(val);
|
|
|
|
}
|
|
|
|
|
|
|
|
private static Operand EmitReverseBytes16_64Op(ArmEmitterContext context, Operand op)
|
|
|
|
{
|
|
|
|
Debug.Assert(op.Type == OperandType.I64);
|
|
|
|
|
|
|
|
return context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(op, Const(0xff00ff00ff00ff00ul)), Const(8)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(op, Const(0x00ff00ff00ff00fful)), Const(8)));
|
|
|
|
}
|
|
|
|
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
public static void Rev32(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
OpCodeAlu op = (OpCodeAlu)context.CurrOp;
|
|
|
|
|
|
|
|
Operand n = GetIntOrZR(context, op.Rn);
|
2019-12-29 17:22:47 -08:00
|
|
|
Operand d;
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
|
|
|
|
if (op.RegisterSize == RegisterSize.Int32)
|
|
|
|
{
|
2019-12-29 17:22:47 -08:00
|
|
|
d = context.ByteSwap(n);
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2019-12-29 17:22:47 -08:00
|
|
|
d = EmitReverseBytes32_64Op(context, n);
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
}
|
2019-12-29 17:22:47 -08:00
|
|
|
|
|
|
|
SetAluDOrZR(context, d);
|
|
|
|
}
|
|
|
|
|
|
|
|
private static Operand EmitReverseBytes32_64Op(ArmEmitterContext context, Operand op)
|
|
|
|
{
|
|
|
|
Debug.Assert(op.Type == OperandType.I64);
|
|
|
|
|
|
|
|
Operand val = EmitReverseBytes16_64Op(context, op);
|
|
|
|
|
|
|
|
return context.BitwiseOr(context.ShiftRightUI(context.BitwiseAnd(val, Const(0xffff0000ffff0000ul)), Const(16)),
|
|
|
|
context.ShiftLeft (context.BitwiseAnd(val, Const(0x0000ffff0000fffful)), Const(16)));
|
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
2019-08-08 11:56:22 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
public static void Rev64(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
OpCodeAlu op = (OpCodeAlu)context.CurrOp;
|
|
|
|
|
|
|
|
SetAluDOrZR(context, context.ByteSwap(GetIntOrZR(context, op.Rn)));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void Rorv(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetAluDOrZR(context, context.RotateRight(GetAluN(context), GetAluMShift(context)));
|
|
|
|
}
|
|
|
|
|
|
|
|
private static Operand GetAluMShift(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
IOpCodeAluRs op = (IOpCodeAluRs)context.CurrOp;
|
|
|
|
|
|
|
|
Operand m = GetIntOrZR(context, op.Rm);
|
|
|
|
|
|
|
|
if (op.RegisterSize == RegisterSize.Int64)
|
|
|
|
{
|
|
|
|
m = context.ConvertI64ToI32(m);
|
|
|
|
}
|
|
|
|
|
|
|
|
return context.BitwiseAnd(m, Const(context.CurrOp.GetBitsCount() - 1));
|
|
|
|
}
|
|
|
|
|
|
|
|
private static void EmitCVFlagsClear(ArmEmitterContext context)
|
|
|
|
{
|
|
|
|
SetFlag(context, PState.CFlag, Const(0));
|
|
|
|
SetFlag(context, PState.VFlag, Const(0));
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void SetAluD(ArmEmitterContext context, Operand d)
|
|
|
|
{
|
|
|
|
SetAluD(context, d, x31IsZR: false);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void SetAluDOrZR(ArmEmitterContext context, Operand d)
|
|
|
|
{
|
|
|
|
SetAluD(context, d, x31IsZR: true);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void SetAluD(ArmEmitterContext context, Operand d, bool x31IsZR)
|
|
|
|
{
|
|
|
|
IOpCodeAlu op = (IOpCodeAlu)context.CurrOp;
|
|
|
|
|
|
|
|
if ((x31IsZR || op is IOpCodeAluRs) && op.Rd == RegisterConsts.ZeroIndex)
|
|
|
|
{
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
SetIntOrSP(context, op.Rd, d);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|