LibWasm: Implement a few SIMD instructions

This commit is contained in:
Ali Mohammad Pur 2023-06-12 13:38:22 +03:30 committed by Ali Mohammad Pur
parent 5f013e5374
commit 3c176bafee
Notes: sideshowbarker 2024-07-16 17:12:03 +09:00
8 changed files with 356 additions and 11 deletions

View file

@ -6,6 +6,7 @@
#pragma once
#include <AK/Concepts.h>
#include <AK/SIMD.h>
// Functions returning vectors or accepting vector arguments have different calling conventions
@ -166,6 +167,31 @@ ALWAYS_INLINE static void store4_masked(VectorType v, UnderlyingType* a, Underly
*d = v[3];
}
// Shuffle
template<OneOf<i8x16, u8x16> T>
ALWAYS_INLINE static T shuffle(T a, T control)
{
// FIXME: This is probably not the fastest way to do this.
return T {
a[control[0] & 0xf],
a[control[1] & 0xf],
a[control[2] & 0xf],
a[control[3] & 0xf],
a[control[4] & 0xf],
a[control[5] & 0xf],
a[control[6] & 0xf],
a[control[7] & 0xf],
a[control[8] & 0xf],
a[control[9] & 0xf],
a[control[10] & 0xf],
a[control[11] & 0xf],
a[control[12] & 0xf],
a[control[13] & 0xf],
a[control[14] & 0xf],
a[control[15] & 0xf],
};
}
}
#pragma GCC diagnostic pop

View file

@ -16,7 +16,7 @@ if(INCLUDE_WASM_SPEC_TESTS)
find_program(WAT2WASM wat2wasm REQUIRED)
find_program(PRETTIER prettier OPTIONAL)
if (NOT SKIP_PRETTIER AND NOT PRETTIER_FOUND)
if (NOT SKIP_PRETTIER AND PRETTIER EQUAL "PRETTIER-NOTFOUND")
message(FATAL_ERROR "Prettier required to format Wasm spec tests! Install prettier or set WASM_SPEC_TEST_SKIP_FORMATTING to ON")
endif()

View file

@ -140,6 +140,7 @@ def parse_typed_value(ast):
value.frombytes(parse_v128_chunk(num[0], ast[1][0]))
assert len(value) - s == size, f'Expected {size} bytes, got {len(value) - s} bytes'
assert len(value) == 16, f'Expected 16 bytes, got {len(value)} bytes'
return {
'type': types[ast[0][0]],
'value': value.tobytes().hex()

View file

@ -140,10 +140,17 @@ public:
if constexpr (IsSame<T, decltype(value)> || (!IsFloatingPoint<T> && IsSame<decltype(value), MakeSigned<T>>)) {
result = static_cast<T>(value);
} else if constexpr (!IsFloatingPoint<T> && IsConvertible<decltype(value), T>) {
if (AK::is_within_range<T>(value))
result = static_cast<T>(value);
// NOTE: No implicit vector <-> scalar conversion.
if constexpr (!IsSame<T, u128>) {
if (AK::is_within_range<T>(value))
result = static_cast<T>(value);
}
}
},
[&](u128 value) {
if constexpr (IsSame<T, u128>)
result = value;
},
[&](Reference const& value) {
if constexpr (IsSame<T, Reference>) {
result = value;

View file

@ -5,9 +5,11 @@
* SPDX-License-Identifier: BSD-2-Clause
*/
#include <AK/ByteReader.h>
#include <AK/Debug.h>
#include <AK/Endian.h>
#include <AK/MemoryStream.h>
#include <AK/SIMDExtras.h>
#include <LibWasm/AbstractMachine/AbstractMachine.h>
#include <LibWasm/AbstractMachine/BytecodeInterpreter.h>
#include <LibWasm/AbstractMachine/Configuration.h>
@ -15,6 +17,8 @@
#include <LibWasm/Opcode.h>
#include <LibWasm/Printer/Printer.h>
using namespace AK::SIMD;
namespace Wasm {
#define TRAP_IF_NOT(x) \
@ -110,6 +114,148 @@ void BytecodeInterpreter::load_and_push(Configuration& configuration, Instructio
configuration.stack().peek() = Value(static_cast<PushType>(read_value<ReadType>(slice)));
}
template<typename TDst, typename TSrc>
ALWAYS_INLINE static TDst convert_vector(TSrc v)
{
return __builtin_convertvector(v, TDst);
}
template<size_t M, size_t N, template<typename> typename SetSign>
void BytecodeInterpreter::load_and_push_mxn(Configuration& configuration, Instruction const& instruction)
{
auto& address = configuration.frame().module().memories().first();
auto memory = configuration.store().get(address);
if (!memory) {
m_trap = Trap { "Nonexistent memory" };
return;
}
auto& arg = instruction.arguments().get<Instruction::MemoryArgument>();
auto& entry = configuration.stack().peek();
auto base = entry.get<Value>().to<i32>();
if (!base.has_value()) {
m_trap = Trap { "Memory access out of bounds" };
return;
}
u64 instance_address = static_cast<u64>(bit_cast<u32>(base.value())) + arg.offset;
Checked addition { instance_address };
addition += M * N / 8;
if (addition.has_overflow() || addition.value() > memory->size()) {
m_trap = Trap { "Memory access out of bounds" };
dbgln("LibWasm: Memory access out of bounds (expected {} to be less than or equal to {})", instance_address + M * N / 8, memory->size());
return;
}
dbgln_if(WASM_TRACE_DEBUG, "vec-load({} : {}) -> stack", instance_address, M * N / 8);
auto slice = memory->data().bytes().slice(instance_address, M * N / 8);
using V64 = NativeVectorType<M, N, SetSign>;
using V128 = NativeVectorType<M * 2, N, SetSign>;
V64 bytes { 0 };
if (bit_cast<FlatPtr>(slice.data()) % sizeof(V64) == 0)
bytes = *bit_cast<V64*>(slice.data());
else
ByteReader::load(slice.data(), bytes);
configuration.stack().peek() = Value(bit_cast<u128>(convert_vector<V128>(bytes)));
}
template<size_t M>
void BytecodeInterpreter::load_and_push_m_splat(Configuration& configuration, Instruction const& instruction)
{
auto& address = configuration.frame().module().memories().first();
auto memory = configuration.store().get(address);
if (!memory) {
m_trap = Trap { "Nonexistent memory" };
return;
}
auto& arg = instruction.arguments().get<Instruction::MemoryArgument>();
auto& entry = configuration.stack().peek();
auto base = entry.get<Value>().to<i32>();
if (!base.has_value()) {
m_trap = Trap { "Memory access out of bounds" };
return;
}
u64 instance_address = static_cast<u64>(bit_cast<u32>(base.value())) + arg.offset;
Checked addition { instance_address };
addition += M / 8;
if (addition.has_overflow() || addition.value() > memory->size()) {
m_trap = Trap { "Memory access out of bounds" };
dbgln("LibWasm: Memory access out of bounds (expected {} to be less than or equal to {})", instance_address + M / 8, memory->size());
return;
}
dbgln_if(WASM_TRACE_DEBUG, "vec-splat({} : {}) -> stack", instance_address, M / 8);
auto slice = memory->data().bytes().slice(instance_address, M / 8);
auto value = read_value<NativeIntegralType<M>>(slice);
set_top_m_splat<M, NativeIntegralType>(configuration, value);
}
template<size_t M, template<size_t> typename NativeType>
void BytecodeInterpreter::set_top_m_splat(Wasm::Configuration& configuration, NativeType<M> value)
{
auto push = [&](auto result) {
configuration.stack().peek() = Value(bit_cast<u128>(result));
};
if constexpr (IsFloatingPoint<NativeType<32>>) {
if constexpr (M == 32) // 32 -> 32x4
push(expand4(value));
else if constexpr (M == 64) // 64 -> 64x2
push(f64x2 { value, value });
else
static_assert(DependentFalse<NativeType<M>>, "Invalid vector size");
} else {
if constexpr (M == 8) // 8 -> 8x4 -> 32x4
push(expand4(bit_cast<u32>(u8x4 { value, value, value, value })));
else if constexpr (M == 16) // 16 -> 16x2 -> 32x4
push(expand4(bit_cast<u32>(u16x2 { value, value })));
else if constexpr (M == 32) // 32 -> 32x4
push(expand4(value));
else if constexpr (M == 64) // 64 -> 64x2
push(u64x2 { value, value });
else
static_assert(DependentFalse<NativeType<M>>, "Invalid vector size");
}
}
template<size_t M, template<size_t> typename NativeType>
void BytecodeInterpreter::pop_and_push_m_splat(Wasm::Configuration& configuration, Instruction const&)
{
using PopT = Conditional<M <= 32, NativeType<32>, NativeType<64>>;
using ReadT = NativeType<M>;
auto entry = configuration.stack().peek();
auto value = static_cast<ReadT>(*entry.get<Value>().to<PopT>());
dbgln_if(WASM_TRACE_DEBUG, "stack({}) -> splat({})", value, M);
set_top_m_splat<M, NativeType>(configuration, value);
}
template<typename M, template<typename> typename SetSign, typename VectorType>
Optional<VectorType> BytecodeInterpreter::pop_vector(Configuration& configuration)
{
auto value = peek_vector<M, SetSign, VectorType>(configuration);
if (value.has_value())
configuration.stack().pop();
return value;
}
template<typename M, template<typename> typename SetSign, typename VectorType>
Optional<VectorType> BytecodeInterpreter::peek_vector(Configuration& configuration)
{
auto& entry = configuration.stack().peek();
auto value = entry.get<Value>().value().get_pointer<u128>();
if (!value)
return {};
auto vector = bit_cast<VectorType>(*value);
dbgln_if(WASM_TRACE_DEBUG, "stack({}) peek-> vector({:x})", *value, bit_cast<u128>(vector));
return vector;
}
template<typename VectorType>
static u128 shuffle_vector(VectorType values, VectorType indices)
{
auto vector = bit_cast<VectorType>(values);
auto indices_vector = bit_cast<VectorType>(indices);
return bit_cast<u128>(shuffle(vector, indices_vector));
}
void BytecodeInterpreter::call_address(Configuration& configuration, FunctionAddress address)
{
TRAP_IF_NOT(m_stack_info.size_free() >= Constants::minimum_stack_space_to_keep_free);
@ -150,15 +296,15 @@ void BytecodeInterpreter::call_address(Configuration& configuration, FunctionAdd
configuration.stack().entries().unchecked_append(move(entry));
}
template<typename PopType, typename PushType, typename Operator>
template<typename PopTypeLHS, typename PushType, typename Operator, typename PopTypeRHS>
void BytecodeInterpreter::binary_numeric_operation(Configuration& configuration)
{
auto rhs_entry = configuration.stack().pop();
auto& lhs_entry = configuration.stack().peek();
auto rhs_ptr = rhs_entry.get_pointer<Value>();
auto lhs_ptr = lhs_entry.get_pointer<Value>();
auto rhs = rhs_ptr->to<PopType>();
auto lhs = lhs_ptr->to<PopType>();
auto rhs = rhs_ptr->to<PopTypeRHS>();
auto lhs = lhs_ptr->to<PopTypeLHS>();
PushType result;
auto call_result = Operator {}(lhs.value(), rhs.value());
if constexpr (IsSpecializationOf<decltype(call_result), AK::Result>) {
@ -1016,6 +1162,78 @@ void BytecodeInterpreter::interpret(Configuration& configuration, InstructionPoi
return unary_operation<double, i64, Operators::SaturatingTruncate<i64>>(configuration);
case Instructions::i64_trunc_sat_f64_u.value():
return unary_operation<double, i64, Operators::SaturatingTruncate<u64>>(configuration);
case Instructions::v128_const.value():
configuration.stack().push(Value(instruction.arguments().get<u128>()));
return;
case Instructions::v128_load.value():
return load_and_push<u128, u128>(configuration, instruction);
case Instructions::v128_load8x8_s.value():
return load_and_push_mxn<8, 8, MakeSigned>(configuration, instruction);
case Instructions::v128_load8x8_u.value():
return load_and_push_mxn<8, 8, MakeUnsigned>(configuration, instruction);
case Instructions::v128_load16x4_s.value():
return load_and_push_mxn<16, 4, MakeSigned>(configuration, instruction);
case Instructions::v128_load16x4_u.value():
return load_and_push_mxn<16, 4, MakeUnsigned>(configuration, instruction);
case Instructions::v128_load32x2_s.value():
return load_and_push_mxn<32, 2, MakeSigned>(configuration, instruction);
case Instructions::v128_load32x2_u.value():
return load_and_push_mxn<32, 2, MakeUnsigned>(configuration, instruction);
case Instructions::v128_load8_splat.value():
return load_and_push_m_splat<8>(configuration, instruction);
case Instructions::v128_load16_splat.value():
return load_and_push_m_splat<16>(configuration, instruction);
case Instructions::v128_load32_splat.value():
return load_and_push_m_splat<32>(configuration, instruction);
case Instructions::v128_load64_splat.value():
return load_and_push_m_splat<64>(configuration, instruction);
case Instructions::i8x16_splat.value():
return pop_and_push_m_splat<8, NativeIntegralType>(configuration, instruction);
case Instructions::i16x8_splat.value():
return pop_and_push_m_splat<16, NativeIntegralType>(configuration, instruction);
case Instructions::i32x4_splat.value():
return pop_and_push_m_splat<32, NativeIntegralType>(configuration, instruction);
case Instructions::i64x2_splat.value():
return pop_and_push_m_splat<64, NativeIntegralType>(configuration, instruction);
case Instructions::f32x4_splat.value():
return pop_and_push_m_splat<32, NativeFloatingType>(configuration, instruction);
case Instructions::f64x2_splat.value():
return pop_and_push_m_splat<64, NativeFloatingType>(configuration, instruction);
case Instructions::i8x16_shuffle.value(): {
auto indices = pop_vector<u8, MakeSigned>(configuration);
TRAP_IF_NOT(indices.has_value());
auto vector = peek_vector<u8, MakeSigned>(configuration);
TRAP_IF_NOT(vector.has_value());
auto result = shuffle_vector(vector.value(), indices.value());
configuration.stack().peek() = Value(result);
return;
}
case Instructions::v128_store.value():
return pop_and_store<u128, u128>(configuration, instruction);
case Instructions::i8x16_shl.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftLeft<16>, i32>(configuration);
case Instructions::i8x16_shr_u.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftRight<16, MakeUnsigned>, i32>(configuration);
case Instructions::i8x16_shr_s.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftRight<16, MakeSigned>, i32>(configuration);
case Instructions::i16x8_shl.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftLeft<8>, i32>(configuration);
case Instructions::i16x8_shr_u.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftRight<8, MakeUnsigned>, i32>(configuration);
case Instructions::i16x8_shr_s.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftRight<8, MakeSigned>, i32>(configuration);
case Instructions::i32x4_shl.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftLeft<4>, i32>(configuration);
case Instructions::i32x4_shr_u.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftRight<4, MakeUnsigned>, i32>(configuration);
case Instructions::i32x4_shr_s.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftRight<4, MakeSigned>, i32>(configuration);
case Instructions::i64x2_shl.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftLeft<2>, i32>(configuration);
case Instructions::i64x2_shr_u.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftRight<2, MakeUnsigned>, i32>(configuration);
case Instructions::i64x2_shr_s.value():
return binary_numeric_operation<u128, u128, Operators::VectorShiftRight<2, MakeSigned>, i32>(configuration);
case Instructions::table_init.value():
case Instructions::elem_drop.value():
case Instructions::table_copy.value():
@ -1024,7 +1242,7 @@ void BytecodeInterpreter::interpret(Configuration& configuration, InstructionPoi
case Instructions::table_fill.value():
default:
unimplemented:;
dbgln("Instruction '{}' not implemented", instruction_name(instruction.opcode()));
dbgln_if(WASM_TRACE_DEBUG, "Instruction '{}' not implemented", instruction_name(instruction.opcode()));
m_trap = Trap { DeprecatedString::formatted("Unimplemented instruction {}", instruction_name(instruction.opcode())) };
return;
}

View file

@ -50,10 +50,22 @@ protected:
void load_and_push(Configuration&, Instruction const&);
template<typename PopT, typename StoreT>
void pop_and_store(Configuration&, Instruction const&);
template<size_t M, size_t N, template<typename> typename SetSign>
void load_and_push_mxn(Configuration&, Instruction const&);
template<size_t M>
void load_and_push_m_splat(Configuration&, Instruction const&);
template<size_t M, template<size_t> typename NativeType>
void set_top_m_splat(Configuration&, NativeType<M>);
template<size_t M, template<size_t> typename NativeType>
void pop_and_push_m_splat(Configuration&, Instruction const&);
template<typename M, template<typename> typename SetSign, typename VectorType = Native128ByteVectorOf<M, SetSign>>
Optional<VectorType> pop_vector(Configuration&);
template<typename M, template<typename> typename SetSign, typename VectorType = Native128ByteVectorOf<M, SetSign>>
Optional<VectorType> peek_vector(Configuration&);
void store_to_memory(Configuration&, Instruction const&, ReadonlyBytes data, i32 base);
void call_address(Configuration&, FunctionAddress);
template<typename PopType, typename PushType, typename Operator>
template<typename PopTypeLHS, typename PushType, typename Operator, typename PopTypeRHS = PopTypeLHS>
void binary_numeric_operation(Configuration&);
template<typename PopType, typename PushType, typename Operator>

View file

@ -1,5 +1,5 @@
/*
* Copyright (c) 2021, Ali Mohammad Pur <mpfard@serenityos.org>
* Copyright (c) 2021-2023, Ali Mohammad Pur <mpfard@serenityos.org>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
@ -9,12 +9,15 @@
#include <AK/BitCast.h>
#include <AK/BuiltinWrappers.h>
#include <AK/Result.h>
#include <AK/SIMD.h>
#include <AK/StringView.h>
#include <AK/Types.h>
#include <limits.h>
#include <math.h>
namespace Operators {
namespace Wasm::Operators {
using namespace AK::SIMD;
#define DEFINE_BINARY_OPERATOR(Name, operation) \
struct Name { \
@ -62,6 +65,7 @@ struct Divide {
static StringView name() { return "/"sv; }
};
struct Modulo {
template<typename Lhs, typename Rhs>
auto operator()(Lhs lhs, Rhs rhs) const
@ -77,18 +81,21 @@ struct Modulo {
static StringView name() { return "%"sv; }
};
struct BitShiftLeft {
template<typename Lhs, typename Rhs>
auto operator()(Lhs lhs, Rhs rhs) const { return lhs << (rhs % (sizeof(lhs) * 8)); }
static StringView name() { return "<<"sv; }
};
struct BitShiftRight {
template<typename Lhs, typename Rhs>
auto operator()(Lhs lhs, Rhs rhs) const { return lhs >> (rhs % (sizeof(lhs) * 8)); }
static StringView name() { return ">>"sv; }
};
struct BitRotateLeft {
template<typename Lhs, typename Rhs>
auto operator()(Lhs lhs, Rhs rhs) const
@ -102,6 +109,7 @@ struct BitRotateLeft {
static StringView name() { return "rotate_left"sv; }
};
struct BitRotateRight {
template<typename Lhs, typename Rhs>
auto operator()(Lhs lhs, Rhs rhs) const
@ -115,6 +123,55 @@ struct BitRotateRight {
static StringView name() { return "rotate_right"sv; }
};
template<size_t VectorSize>
struct VectorShiftLeft {
auto operator()(u128 lhs, i32 rhs) const
{
auto shift_value = rhs % (sizeof(lhs) * 8 / VectorSize);
return bit_cast<u128>(bit_cast<Native128ByteVectorOf<NativeIntegralType<128 / VectorSize>, MakeUnsigned>>(lhs) << shift_value);
}
static StringView name()
{
switch (VectorSize) {
case 16:
return "vec(8x16)<<"sv;
case 8:
return "vec(16x8)<<"sv;
case 4:
return "vec(32x4)<<"sv;
case 2:
return "vec(64x2)<<"sv;
default:
VERIFY_NOT_REACHED();
}
}
};
template<size_t VectorSize, template<typename> typename SetSign>
struct VectorShiftRight {
auto operator()(u128 lhs, i32 rhs) const
{
auto shift_value = rhs % (sizeof(lhs) * 8 / VectorSize);
return bit_cast<u128>(bit_cast<Native128ByteVectorOf<NativeIntegralType<128 / VectorSize>, SetSign>>(lhs) >> shift_value);
}
static StringView name()
{
switch (VectorSize) {
case 16:
return "vec(8x16)>>"sv;
case 8:
return "vec(16x8)>>"sv;
case 4:
return "vec(32x4)>>"sv;
case 2:
return "vec(64x2)>>"sv;
default:
VERIFY_NOT_REACHED();
}
}
};
struct Minimum {
template<typename Lhs, typename Rhs>
auto operator()(Lhs lhs, Rhs rhs) const
@ -134,6 +191,7 @@ struct Minimum {
static StringView name() { return "minimum"sv; }
};
struct Maximum {
template<typename Lhs, typename Rhs>
auto operator()(Lhs lhs, Rhs rhs) const
@ -153,6 +211,7 @@ struct Maximum {
static StringView name() { return "maximum"sv; }
};
struct CopySign {
template<typename Lhs, typename Rhs>
auto operator()(Lhs lhs, Rhs rhs) const
@ -176,6 +235,7 @@ struct EqualsZero {
static StringView name() { return "== 0"sv; }
};
struct CountLeadingZeros {
template<typename Lhs>
i32 operator()(Lhs lhs) const
@ -191,6 +251,7 @@ struct CountLeadingZeros {
static StringView name() { return "clz"sv; }
};
struct CountTrailingZeros {
template<typename Lhs>
i32 operator()(Lhs lhs) const
@ -206,6 +267,7 @@ struct CountTrailingZeros {
static StringView name() { return "ctz"sv; }
};
struct PopCount {
template<typename Lhs>
auto operator()(Lhs lhs) const
@ -218,18 +280,21 @@ struct PopCount {
static StringView name() { return "popcnt"sv; }
};
struct Absolute {
template<typename Lhs>
auto operator()(Lhs lhs) const { return AK::abs(lhs); }
static StringView name() { return "abs"sv; }
};
struct Negate {
template<typename Lhs>
auto operator()(Lhs lhs) const { return -lhs; }
static StringView name() { return "== 0"sv; }
};
struct Ceil {
template<typename Lhs>
auto operator()(Lhs lhs) const
@ -244,6 +309,7 @@ struct Ceil {
static StringView name() { return "ceil"sv; }
};
struct Floor {
template<typename Lhs>
auto operator()(Lhs lhs) const
@ -258,9 +324,10 @@ struct Floor {
static StringView name() { return "floor"sv; }
};
struct Truncate {
template<typename Lhs>
Result<Lhs, StringView> operator()(Lhs lhs) const
AK::Result<Lhs, StringView> operator()(Lhs lhs) const
{
if constexpr (IsSame<Lhs, float>)
return truncf(lhs);
@ -272,6 +339,7 @@ struct Truncate {
static StringView name() { return "truncate"sv; }
};
struct NearbyIntegral {
template<typename Lhs>
auto operator()(Lhs lhs) const
@ -286,6 +354,7 @@ struct NearbyIntegral {
static StringView name() { return "round"sv; }
};
struct SquareRoot {
template<typename Lhs>
auto operator()(Lhs lhs) const

View file

@ -20,6 +20,18 @@
namespace Wasm {
template<size_t M>
using NativeIntegralType = Conditional<M == 8, u8, Conditional<M == 16, u16, Conditional<M == 32, u32, Conditional<M == 64, u64, void>>>>;
template<size_t M>
using NativeFloatingType = Conditional<M == 32, f32, Conditional<M == 64, f64, void>>;
template<size_t M, size_t N, template<typename> typename SetSign, typename ElementType = SetSign<NativeIntegralType<M>>>
using NativeVectorType __attribute__((vector_size(N * sizeof(ElementType)))) = ElementType;
template<typename T, template<typename> typename SetSign>
using Native128ByteVectorOf = NativeVectorType<sizeof(T) * 8, 16 / sizeof(T), SetSign, T>;
enum class ParseError {
UnexpectedEof,
UnknownInstruction,