ryujinx/ARMeilleure/Instructions/InstEmitMemoryHelper.cs

512 lines
16 KiB
C#
Raw Normal View History

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 14:56:22 -04:00
using ARMeilleure.Decoders;
using ARMeilleure.IntermediateRepresentation;
using ARMeilleure.Memory;
using ARMeilleure.Translation;
using System;
using static ARMeilleure.Instructions.InstEmitHelper;
using static ARMeilleure.IntermediateRepresentation.OperandHelper;
namespace ARMeilleure.Instructions
{
static class InstEmitMemoryHelper
{
private enum Extension
{
Zx,
Sx32,
Sx64
}
public static void EmitLoadZx(ArmEmitterContext context, Operand address, int rt, int size)
{
EmitLoad(context, address, Extension.Zx, rt, size);
}
public static void EmitLoadSx32(ArmEmitterContext context, Operand address, int rt, int size)
{
EmitLoad(context, address, Extension.Sx32, rt, size);
}
public static void EmitLoadSx64(ArmEmitterContext context, Operand address, int rt, int size)
{
EmitLoad(context, address, Extension.Sx64, rt, size);
}
private static void EmitLoad(ArmEmitterContext context, Operand address, Extension ext, int rt, int size)
{
bool isSimd = IsSimd(context);
if ((uint)size > (isSimd ? 4 : 3))
{
throw new ArgumentOutOfRangeException(nameof(size));
}
if (isSimd)
{
EmitReadVector(context, address, context.VectorZero(), rt, 0, size);
}
else
{
EmitReadInt(context, address, rt, size);
}
if (!isSimd)
{
Operand value = GetIntOrZR(context, rt);
if (ext == Extension.Sx32 || ext == Extension.Sx64)
{
OperandType destType = ext == Extension.Sx64 ? OperandType.I64 : OperandType.I32;
switch (size)
{
case 0: value = context.SignExtend8 (destType, value); break;
case 1: value = context.SignExtend16(destType, value); break;
case 2: value = context.SignExtend32(destType, value); break;
}
}
SetIntOrZR(context, rt, value);
}
}
public static void EmitLoadSimd(
ArmEmitterContext context,
Operand address,
Operand vector,
int rt,
int elem,
int size)
{
EmitReadVector(context, address, vector, rt, elem, size);
}
public static void EmitStore(ArmEmitterContext context, Operand address, int rt, int size)
{
bool isSimd = IsSimd(context);
if ((uint)size > (isSimd ? 4 : 3))
{
throw new ArgumentOutOfRangeException(nameof(size));
}
if (isSimd)
{
EmitWriteVector(context, address, rt, 0, size);
}
else
{
EmitWriteInt(context, address, rt, size);
}
}
public static void EmitStoreSimd(
ArmEmitterContext context,
Operand address,
int rt,
int elem,
int size)
{
EmitWriteVector(context, address, rt, elem, size);
}
private static bool IsSimd(ArmEmitterContext context)
{
return context.CurrOp is IOpCodeSimd &&
!(context.CurrOp is OpCodeSimdMemMs ||
context.CurrOp is OpCodeSimdMemSs);
}
private static void EmitReadInt(ArmEmitterContext context, Operand address, int rt, int size)
{
Operand isUnalignedAddr = EmitAddressCheck(context, address, size);
Operand lblFastPath = Label();
Operand lblSlowPath = Label();
Operand lblEnd = Label();
context.BranchIfFalse(lblFastPath, isUnalignedAddr);
context.MarkLabel(lblSlowPath);
EmitReadIntFallback(context, address, rt, size);
context.Branch(lblEnd);
context.MarkLabel(lblFastPath);
Operand physAddr = EmitPtPointerLoad(context, address, lblSlowPath);
Operand value = null;
switch (size)
{
case 0:
value = context.Load8(physAddr);
break;
case 1:
value = context.Load16(physAddr);
break;
case 2:
value = context.Load(OperandType.I32, physAddr);
break;
case 3:
value = context.Load(OperandType.I64, physAddr);
break;
}
SetInt(context, rt, value);
context.MarkLabel(lblEnd);
}
private static void EmitReadVector(
ArmEmitterContext context,
Operand address,
Operand vector,
int rt,
int elem,
int size)
{
Operand isUnalignedAddr = EmitAddressCheck(context, address, size);
Operand lblFastPath = Label();
Operand lblSlowPath = Label();
Operand lblEnd = Label();
context.BranchIfFalse(lblFastPath, isUnalignedAddr);
context.MarkLabel(lblSlowPath);
EmitReadVectorFallback(context, address, vector, rt, elem, size);
context.Branch(lblEnd);
context.MarkLabel(lblFastPath);
Operand physAddr = EmitPtPointerLoad(context, address, lblSlowPath);
Operand value = null;
switch (size)
{
case 0:
value = context.VectorInsert8(vector, context.Load8(physAddr), elem);
break;
case 1:
value = context.VectorInsert16(vector, context.Load16(physAddr), elem);
break;
case 2:
value = context.VectorInsert(vector, context.Load(OperandType.I32, physAddr), elem);
break;
case 3:
value = context.VectorInsert(vector, context.Load(OperandType.I64, physAddr), elem);
break;
case 4:
value = context.Load(OperandType.V128, physAddr);
break;
}
context.Copy(GetVec(rt), value);
context.MarkLabel(lblEnd);
}
private static Operand VectorCreate(ArmEmitterContext context, Operand value)
{
return context.VectorInsert(context.VectorZero(), value, 0);
}
private static void EmitWriteInt(ArmEmitterContext context, Operand address, int rt, int size)
{
Operand isUnalignedAddr = EmitAddressCheck(context, address, size);
Operand lblFastPath = Label();
Operand lblSlowPath = Label();
Operand lblEnd = Label();
context.BranchIfFalse(lblFastPath, isUnalignedAddr);
context.MarkLabel(lblSlowPath);
EmitWriteIntFallback(context, address, rt, size);
context.Branch(lblEnd);
context.MarkLabel(lblFastPath);
Operand physAddr = EmitPtPointerLoad(context, address, lblSlowPath);
Operand value = GetInt(context, rt);
if (size < 3 && value.Type == OperandType.I64)
{
value = context.ConvertI64ToI32(value);
}
switch (size)
{
case 0: context.Store8 (physAddr, value); break;
case 1: context.Store16(physAddr, value); break;
case 2: context.Store (physAddr, value); break;
case 3: context.Store (physAddr, value); break;
}
context.MarkLabel(lblEnd);
}
private static void EmitWriteVector(
ArmEmitterContext context,
Operand address,
int rt,
int elem,
int size)
{
Operand isUnalignedAddr = EmitAddressCheck(context, address, size);
Operand lblFastPath = Label();
Operand lblSlowPath = Label();
Operand lblEnd = Label();
context.BranchIfFalse(lblFastPath, isUnalignedAddr);
context.MarkLabel(lblSlowPath);
EmitWriteVectorFallback(context, address, rt, elem, size);
context.Branch(lblEnd);
context.MarkLabel(lblFastPath);
Operand physAddr = EmitPtPointerLoad(context, address, lblSlowPath);
Operand value = GetVec(rt);
switch (size)
{
case 0:
context.Store8(physAddr, context.VectorExtract8(value, elem));
break;
case 1:
context.Store16(physAddr, context.VectorExtract16(value, elem));
break;
case 2:
context.Store(physAddr, context.VectorExtract(OperandType.FP32, value, elem));
break;
case 3:
context.Store(physAddr, context.VectorExtract(OperandType.FP64, value, elem));
break;
case 4:
context.Store(physAddr, value);
break;
}
context.MarkLabel(lblEnd);
}
private static Operand EmitAddressCheck(ArmEmitterContext context, Operand address, int size)
{
long addressCheckMask = ~(context.Memory.AddressSpaceSize - 1);
addressCheckMask |= (1u << size) - 1;
return context.BitwiseAnd(address, Const(address.Type, addressCheckMask));
}
private static Operand EmitPtPointerLoad(ArmEmitterContext context, Operand address, Operand lblFallbackPath)
{
Operand pte = Const(context.Memory.PageTable.ToInt64());
int bit = MemoryManager.PageBits;
do
{
Operand addrPart = context.ShiftRightUI(address, Const(bit));
bit += context.Memory.PtLevelBits;
if (bit < context.Memory.AddressSpaceBits)
{
addrPart = context.BitwiseAnd(addrPart, Const(addrPart.Type, context.Memory.PtLevelMask));
}
Operand pteOffset = context.ShiftLeft(addrPart, Const(3));
if (pteOffset.Type == OperandType.I32)
{
pteOffset = context.ZeroExtend32(OperandType.I64, pteOffset);
}
Operand pteAddress = context.Add(pte, pteOffset);
pte = context.Load(OperandType.I64, pteAddress);
}
while (bit < context.Memory.AddressSpaceBits);
if (!context.Memory.HasWriteWatchSupport)
{
Operand hasFlagSet = context.BitwiseAnd(pte, Const((long)MemoryManager.PteFlagsMask));
context.BranchIfTrue(lblFallbackPath, hasFlagSet);
}
Operand pageOffset = context.BitwiseAnd(address, Const(address.Type, MemoryManager.PageMask));
if (pageOffset.Type == OperandType.I32)
{
pageOffset = context.ZeroExtend32(OperandType.I64, pageOffset);
}
Operand physAddr = context.Add(pte, pageOffset);
return physAddr;
}
private static void EmitReadIntFallback(ArmEmitterContext context, Operand address, int rt, int size)
{
Delegate fallbackMethodDlg = null;
switch (size)
{
case 0: fallbackMethodDlg = new _U8_U64 (NativeInterface.ReadByte); break;
case 1: fallbackMethodDlg = new _U16_U64(NativeInterface.ReadUInt16); break;
case 2: fallbackMethodDlg = new _U32_U64(NativeInterface.ReadUInt32); break;
case 3: fallbackMethodDlg = new _U64_U64(NativeInterface.ReadUInt64); break;
}
SetInt(context, rt, context.Call(fallbackMethodDlg, address));
}
private static void EmitReadVectorFallback(
ArmEmitterContext context,
Operand address,
Operand vector,
int rt,
int elem,
int size)
{
Delegate fallbackMethodDlg = null;
switch (size)
{
case 0: fallbackMethodDlg = new _U8_U64 (NativeInterface.ReadByte); break;
case 1: fallbackMethodDlg = new _U16_U64 (NativeInterface.ReadUInt16); break;
case 2: fallbackMethodDlg = new _U32_U64 (NativeInterface.ReadUInt32); break;
case 3: fallbackMethodDlg = new _U64_U64 (NativeInterface.ReadUInt64); break;
case 4: fallbackMethodDlg = new _V128_U64(NativeInterface.ReadVector128); break;
}
Operand value = context.Call(fallbackMethodDlg, address);
switch (size)
{
case 0: value = context.VectorInsert8 (vector, value, elem); break;
case 1: value = context.VectorInsert16(vector, value, elem); break;
case 2: value = context.VectorInsert (vector, value, elem); break;
case 3: value = context.VectorInsert (vector, value, elem); break;
}
context.Copy(GetVec(rt), value);
}
private static void EmitWriteIntFallback(ArmEmitterContext context, Operand address, int rt, int size)
{
Delegate fallbackMethodDlg = null;
switch (size)
{
case 0: fallbackMethodDlg = new _Void_U64_U8 (NativeInterface.WriteByte); break;
case 1: fallbackMethodDlg = new _Void_U64_U16(NativeInterface.WriteUInt16); break;
case 2: fallbackMethodDlg = new _Void_U64_U32(NativeInterface.WriteUInt32); break;
case 3: fallbackMethodDlg = new _Void_U64_U64(NativeInterface.WriteUInt64); break;
}
Operand value = GetInt(context, rt);
if (size < 3 && value.Type == OperandType.I64)
{
value = context.ConvertI64ToI32(value);
}
context.Call(fallbackMethodDlg, address, value);
}
private static void EmitWriteVectorFallback(
ArmEmitterContext context,
Operand address,
int rt,
int elem,
int size)
{
Delegate fallbackMethodDlg = null;
switch (size)
{
case 0: fallbackMethodDlg = new _Void_U64_U8 (NativeInterface.WriteByte); break;
case 1: fallbackMethodDlg = new _Void_U64_U16 (NativeInterface.WriteUInt16); break;
case 2: fallbackMethodDlg = new _Void_U64_U32 (NativeInterface.WriteUInt32); break;
case 3: fallbackMethodDlg = new _Void_U64_U64 (NativeInterface.WriteUInt64); break;
case 4: fallbackMethodDlg = new _Void_U64_V128(NativeInterface.WriteVector128); break;
}
Operand value = null;
if (size < 4)
{
switch (size)
{
case 0:
value = context.VectorExtract8(GetVec(rt), elem);
break;
case 1:
value = context.VectorExtract16(GetVec(rt), elem);
break;
case 2:
value = context.VectorExtract(OperandType.I32, GetVec(rt), elem);
break;
case 3:
value = context.VectorExtract(OperandType.I64, GetVec(rt), elem);
break;
}
}
else
{
value = GetVec(rt);
}
context.Call(fallbackMethodDlg, address, value);
}
private static Operand GetInt(ArmEmitterContext context, int rt)
{
return context.CurrOp is OpCode32 ? GetIntA32(context, rt) : GetIntOrZR(context, rt);
}
private static void SetInt(ArmEmitterContext context, int rt, Operand value)
{
if (context.CurrOp is OpCode32)
{
SetIntA32(context, rt, value);
}
else
{
SetIntOrZR(context, rt, value);
}
}
}
}