ir: separate components out a little more

src/CMakeLists.txt

@@ -15,9 +15,11 @@ set(SRCS
     frontend/arm_types.cpp
     frontend/disassembler/disassembler_arm.cpp
     frontend/disassembler/disassembler_thumb.cpp
-    frontend/ir/ir.cpp
+    frontend/ir/basic_block.cpp
     frontend/ir/ir_emitter.cpp
+    frontend/ir/microinstruction.cpp
     frontend/ir/opcodes.cpp
+    frontend/ir/value.cpp
     frontend/translate/translate.cpp
     frontend/translate/translate_arm.cpp
     frontend/translate/translate_arm/branch.cpp
@@ -63,9 +65,12 @@ set(HEADERS
     frontend/decoder/thumb32.h
     frontend/decoder/vfp2.h
     frontend/disassembler/disassembler.h
-    frontend/ir/ir.h
+    frontend/ir/basic_block.h
     frontend/ir/ir_emitter.h
+    frontend/ir/microinstruction.h
     frontend/ir/opcodes.h
+    frontend/ir/terminal.h
+    frontend/ir/value.h
     frontend/translate/translate.h
     frontend/translate/translate_arm/translate_arm.h
     interface/interface.h

src/backend_x64/emit_x64.h

@@ -8,13 +8,17 @@
 
 #include <set>
 #include <unordered_map>
+#include <vector>
 
 #include <boost/optional.hpp>
 
 #include "backend_x64/block_of_code.h"
 #include "backend_x64/reg_alloc.h"
 #include "common/x64/emitter.h"
-#include "frontend/ir/ir.h"
+#include "frontend/arm_types.h"
+#include "frontend/ir/basic_block.h"
+#include "frontend/ir/microinstruction.h"
+#include "frontend/ir/terminal.h"
 #include "interface/interface.h"
 
 namespace Dynarmic {

src/backend_x64/reg_alloc.h

@@ -6,13 +6,18 @@
 
 #pragma once
 
+#include <array>
 #include <map>
+#include <vector>
+
+#include <boost/optional.hpp>
 
 #include "backend_x64/block_of_code.h"
 #include "backend_x64/jitstate.h"
 #include "common/common_types.h"
 #include "common/x64/emitter.h"
-#include "frontend/ir/ir.h"
+#include "frontend/ir/microinstruction.h"
+#include "frontend/ir/value.h"
 
 namespace Dynarmic {
 namespace BackendX64 {

src/frontend/ir/basic_block.cpp

@@ -0,0 +1,85 @@
+/* This file is part of the dynarmic project.
+ * Copyright (c) 2016 MerryMage
+ * This software may be used and distributed according to the terms of the GNU
+ * General Public License version 2 or any later version.
+ */
+
+#include <map>
+#include <string>
+
+#include "common/string_util.h"
+#include "frontend/ir/basic_block.h"
+
+namespace Dynarmic {
+namespace IR {
+
+std::string DumpBlock(const IR::Block& block) {
+    std::string ret;
+
+    const auto loc_to_string = [](Arm::LocationDescriptor loc) -> std::string {
+        return Common::StringFromFormat("{%u,%s,%s,%u}",
+                                        loc.PC(),
+                                        loc.TFlag() ? "T" : "!T",
+                                        loc.EFlag() ? "E" : "!E",
+                                        loc.FPSCR());
+    };
+
+    ret += Common::StringFromFormat("Block: location=%s\n", loc_to_string(block.location).c_str());
+    ret += Common::StringFromFormat("cycles=%zu", block.cycle_count);
+    ret += Common::StringFromFormat(", entry_cond=%s", Arm::CondToString(block.cond, true));
+    if (block.cond != Arm::Cond::AL) {
+        ret += Common::StringFromFormat(", cond_fail=%s", loc_to_string(block.cond_failed.get()).c_str());
+    }
+    ret += "\n";
+
+    std::map<const IR::Inst*, size_t> inst_to_index;
+    size_t index = 0;
+
+    const auto arg_to_string = [&inst_to_index](const IR::Value& arg) -> std::string {
+        if (arg.IsEmpty()) {
+            return "<null>";
+        } else if (!arg.IsImmediate()) {
+            return Common::StringFromFormat("%%%zu", inst_to_index.at(arg.GetInst()));
+        }
+        switch (arg.GetType()) {
+            case Type::U1:
+                return Common::StringFromFormat("#%s", arg.GetU1() ? "1" : "0");
+            case Type::U8:
+                return Common::StringFromFormat("#%u", arg.GetU8());
+            case Type::U32:
+                return Common::StringFromFormat("#%#x", arg.GetU32());
+            case Type::RegRef:
+                return Arm::RegToString(arg.GetRegRef());
+            case Type::ExtRegRef:
+                return Arm::ExtRegToString(arg.GetExtRegRef());
+            default:
+                return "<unknown immediate type>";
+        }
+    };
+
+    for (auto inst = block.instructions.begin(); inst != block.instructions.end(); ++inst) {
+        const Opcode op = inst->GetOpcode();
+
+        if (GetTypeOf(op) != Type::Void) {
+            ret += Common::StringFromFormat("%%%-5zu = ", index);
+        } else {
+            ret += "         "; // '%00000 = ' -> 1 + 5 + 3 = 9 spaces
+        }
+
+        ret += GetNameOf(op);
+
+        const size_t arg_count = GetNumArgsOf(op);
+        for (size_t arg_index = 0; arg_index < arg_count; arg_index++) {
+            ret += arg_index != 0 ? ", " : " ";
+            ret += arg_to_string(inst->GetArg(arg_index));
+        }
+
+        ret += "\n";
+        inst_to_index[&*inst] = index++;
+    }
+
+    return ret;
+}
+
+} // namespace IR
+} // namespace Dynarmic

src/frontend/ir/basic_block.h

@@ -0,0 +1,56 @@
+/* This file is part of the dynarmic project.
+ * Copyright (c) 2016 MerryMage
+ * This software may be used and distributed according to the terms of the GNU
+ * General Public License version 2 or any later version.
+ */
+
+#pragma once
+
+#include <memory>
+#include <string>
+
+#include <boost/optional.hpp>
+
+#include "common/common_types.h"
+#include "common/intrusive_list.h"
+#include "common/memory_pool.h"
+#include "frontend/arm_types.h"
+#include "frontend/ir/microinstruction.h"
+#include "frontend/ir/terminal.h"
+
+namespace Dynarmic {
+namespace IR {
+
+/**
+ * A basic block. It consists of zero or more instructions followed by exactly one terminal.
+ * Note that this is a linear IR and not a pure tree-based IR: i.e.: there is an ordering to
+ * the microinstructions. This only matters before chaining is done in order to correctly
+ * order memory accesses.
+ */
+class Block final {
+public:
+    explicit Block(const Arm::LocationDescriptor& location) : location(location) {}
+
+    /// Description of the starting location of this block
+    Arm::LocationDescriptor location;
+    /// Conditional to pass in order to execute this block
+    Arm::Cond cond = Arm::Cond::AL;
+    /// Block to execute next if `cond` did not pass.
+    boost::optional<Arm::LocationDescriptor> cond_failed = {};
+
+    /// List of instructions in this block.
+    Common::IntrusiveList<Inst> instructions;
+    /// Memory pool for instruction list
+    std::unique_ptr<Common::Pool> instruction_alloc_pool = std::make_unique<Common::Pool>(sizeof(Inst), 4096);
+    /// Terminal instruction of this block.
+    Terminal terminal = Term::Invalid{};
+
+    /// Number of cycles this block takes to execute.
+    size_t cycle_count = 0;
+};
+
+/// Returns a string representation of the contents of block. Intended for debugging.
+std::string DumpBlock(const IR::Block& block);
+
+} // namespace IR
+} // namespace Dynarmic

src/frontend/ir/ir.cpp

@@ -1,254 +0,0 @@
-/* This file is part of the dynarmic project.
- * Copyright (c) 2016 MerryMage
- * This software may be used and distributed according to the terms of the GNU
- * General Public License version 2 or any later version.
- */
-
-#include <algorithm>
-#include <map>
-
-#include "common/assert.h"
-#include "common/string_util.h"
-#include "frontend/ir/ir.h"
-#include "frontend/ir/opcodes.h"
-
-namespace Dynarmic {
-namespace IR {
-
-// Value class member definitions
-
-Value::Value(Inst* value) : type(Type::Opaque) {
-    inner.inst = value;
-}
-
-Value::Value(Arm::Reg value) : type(Type::RegRef) {
-    inner.imm_regref = value;
-}
-
-Value::Value(Arm::ExtReg value) : type(Type::ExtRegRef) {
-    inner.imm_extregref = value;
-}
-
-Value::Value(bool value) : type(Type::U1) {
-    inner.imm_u1 = value;
-}
-
-Value::Value(u8 value) : type(Type::U8) {
-    inner.imm_u8 = value;
-}
-
-Value::Value(u32 value) : type(Type::U32) {
-    inner.imm_u32 = value;
-}
-
-Value::Value(u64 value) : type(Type::U64) {
-    inner.imm_u64 = value;
-}
-
-bool Value::IsImmediate() const {
-    if (type == Type::Opaque)
-        return inner.inst->GetOpcode() == Opcode::Identity ? inner.inst->GetArg(0).IsImmediate() : false;
-    return true;
-}
-
-bool Value::IsEmpty() const {
-    return type == Type::Void;
-}
-
-Type Value::GetType() const {
-    if (type == Type::Opaque) {
-        if (inner.inst->GetOpcode() == Opcode::Identity) {
-            return inner.inst->GetArg(0).GetType();
-        } else {
-            return inner.inst->GetType();
-        }
-    }
-    return type;
-}
-
-Arm::Reg Value::GetRegRef() const {
-    DEBUG_ASSERT(type == Type::RegRef);
-    return inner.imm_regref;
-}
-
-Arm::ExtReg Value::GetExtRegRef() const {
-    DEBUG_ASSERT(type == Type::ExtRegRef);
-    return inner.imm_extregref;
-}
-
-Inst* Value::GetInst() const {
-    DEBUG_ASSERT(type == Type::Opaque);
-    return inner.inst;
-}
-
-bool Value::GetU1() const {
-    if (type == Type::Opaque && inner.inst->GetOpcode() == Opcode::Identity)
-        return inner.inst->GetArg(0).GetU1();
-    DEBUG_ASSERT(type == Type::U1);
-    return inner.imm_u1;
-}
-
-u8 Value::GetU8() const {
-    if (type == Type::Opaque && inner.inst->GetOpcode() == Opcode::Identity)
-        return inner.inst->GetArg(0).GetU8();
-    DEBUG_ASSERT(type == Type::U8);
-    return inner.imm_u8;
-}
-
-u32 Value::GetU32() const {
-    if (type == Type::Opaque && inner.inst->GetOpcode() == Opcode::Identity)
-        return inner.inst->GetArg(0).GetU32();
-    DEBUG_ASSERT(type == Type::U32);
-    return inner.imm_u32;
-}
-
-u64 Value::GetU64() const {
-    if (type == Type::Opaque && inner.inst->GetOpcode() == Opcode::Identity)
-        return inner.inst->GetArg(0).GetU64();
-    DEBUG_ASSERT(type == Type::U64);
-    return inner.imm_u64;
-}
-
-// Inst class member definitions
-
-Value Inst::GetArg(size_t index) const {
-    DEBUG_ASSERT(index < GetNumArgsOf(op));
-    DEBUG_ASSERT(!args[index].IsEmpty());
-
-    return args[index];
-}
-
-void Inst::SetArg(size_t index, Value value) {
-    DEBUG_ASSERT(index < GetNumArgsOf(op));
-    DEBUG_ASSERT(value.GetType() == GetArgTypeOf(op, index) || Type::Opaque == GetArgTypeOf(op, index));
-
-    if (!args[index].IsImmediate()) {
-        UndoUse(args[index]);
-    }
-    if (!value.IsImmediate()) {
-        Use(value);
-    }
-
-    args[index] = value;
-}
-
-void Inst::Invalidate() {
-    for (auto& value : args) {
-        if (!value.IsImmediate()) {
-            UndoUse(value);
-        }
-    }
-}
-
-void Inst::ReplaceUsesWith(Value& replacement) {
-    Invalidate();
-
-    op = Opcode::Identity;
-
-    if (!replacement.IsImmediate()) {
-        Use(replacement);
-    }
-
-    args[0] = replacement;
-}
-
-void Inst::Use(Value& value) {
-    value.GetInst()->use_count++;
-
-    switch (op){
-        case Opcode::GetCarryFromOp:
-            value.GetInst()->carry_inst = this;
-            break;
-        case Opcode::GetOverflowFromOp:
-            value.GetInst()->overflow_inst = this;
-            break;
-        default:
-            break;
-    }
-}
-
-void Inst::UndoUse(Value& value) {
-    value.GetInst()->use_count--;
-
-    switch (op){
-        case Opcode::GetCarryFromOp:
-            value.GetInst()->carry_inst = nullptr;
-            break;
-        case Opcode::GetOverflowFromOp:
-            value.GetInst()->overflow_inst = nullptr;
-            break;
-        default:
-            break;
-    }
-}
-
-std::string DumpBlock(const IR::Block& block) {
-    std::string ret;
-
-    const auto loc_to_string = [](Arm::LocationDescriptor loc) -> std::string {
-        return Common::StringFromFormat("{%u,%s,%s,%u}",
-                                        loc.PC(),
-                                        loc.TFlag() ? "T" : "!T",
-                                        loc.EFlag() ? "E" : "!E",
-                                        loc.FPSCR());
-    };
-
-    ret += Common::StringFromFormat("Block: location=%s\n", loc_to_string(block.location).c_str());
-    ret += Common::StringFromFormat("cycles=%zu", block.cycle_count);
-    ret += Common::StringFromFormat(", entry_cond=%s", Arm::CondToString(block.cond, true));
-    if (block.cond != Arm::Cond::AL) {
-        ret += Common::StringFromFormat(", cond_fail=%s", loc_to_string(block.cond_failed.get()).c_str());
-    }
-    ret += "\n";
-
-    std::map<const IR::Inst*, size_t> inst_to_index;
-    size_t index = 0;
-
-    const auto arg_to_string = [&inst_to_index](const IR::Value& arg) -> std::string {
-        if (arg.IsEmpty()) {
-            return "<null>";
-        } else if (!arg.IsImmediate()) {
-            return Common::StringFromFormat("%%%zu", inst_to_index.at(arg.GetInst()));
-        }
-        switch (arg.GetType()) {
-            case Type::U1:
-                return Common::StringFromFormat("#%s", arg.GetU1() ? "1" : "0");
-            case Type::U8:
-                return Common::StringFromFormat("#%u", arg.GetU8());
-            case Type::U32:
-                return Common::StringFromFormat("#%#x", arg.GetU32());
-            case Type::RegRef:
-                return Arm::RegToString(arg.GetRegRef());
-            case Type::ExtRegRef:
-                return Arm::ExtRegToString(arg.GetExtRegRef());
-            default:
-                return "<unknown immediate type>";
-        }
-    };
-
-    for (auto inst = block.instructions.begin(); inst != block.instructions.end(); ++inst) {
-        const Opcode op = inst->GetOpcode();
-
-        if (GetTypeOf(op) != Type::Void) {
-            ret += Common::StringFromFormat("%%%-5zu = ", index);
-        } else {
-            ret += "         "; // '%00000 = ' -> 1 + 5 + 3 = 9 spaces
-        }
-
-        ret += GetNameOf(op);
-
-        const size_t arg_count = GetNumArgsOf(op);
-        for (size_t arg_index = 0; arg_index < arg_count; arg_index++) {
-            ret += arg_index != 0 ? ", " : " ";
-            ret += arg_to_string(inst->GetArg(arg_index));
-        }
-
-        ret += "\n";
-        inst_to_index[&*inst] = index++;
-    }
-
-    return ret;
-}
-
-} // namespace IR
-} // namespace Dynarmic

src/frontend/ir/ir.h

@@ -1,238 +0,0 @@
-/* This file is part of the dynarmic project.
- * Copyright (c) 2016 MerryMage
- * This software may be used and distributed according to the terms of the GNU
- * General Public License version 2 or any later version.
- */
-
-#pragma once
-
-#include <list>
-#include <memory>
-#include <vector>
-
-#include <boost/optional.hpp>
-#include <boost/variant.hpp>
-
-#include "common/assert.h"
-#include "common/common_types.h"
-#include "common/intrusive_list.h"
-#include "common/memory_pool.h"
-#include "frontend/arm_types.h"
-#include "frontend/ir/opcodes.h"
-
-namespace Dynarmic {
-namespace IR {
-
-// ARM JIT Microinstruction Intermediate Representation
-//
-// This intermediate representation is an SSA IR. It is designed primarily for analysis,
-// though it can be lowered into a reduced form for interpretation. Each IR node (Value)
-// is a microinstruction of an idealised ARM CPU. The choice of microinstructions is made
-// not based on any existing microarchitecture but on ease of implementation.
-//
-// A basic block is represented as an IR::Block.
-
-// Type declarations
-
-struct Value;
-class Inst;
-
-/**
- * A representation of a value in the IR.
- * A value may either be an immediate or the result of a microinstruction.
- */
-struct Value final {
-public:
-    Value() : type(Type::Void) {}
-    explicit Value(Inst* value);
-    explicit Value(Arm::Reg value);
-    explicit Value(Arm::ExtReg value);
-    explicit Value(bool value);
-    explicit Value(u8 value);
-    explicit Value(u32 value);
-    explicit Value(u64 value);
-
-    bool IsEmpty() const;
-    bool IsImmediate() const;
-    Type GetType() const;
-
-    Inst* GetInst() const;
-    Arm::Reg GetRegRef() const;
-    Arm::ExtReg GetExtRegRef() const;
-    bool GetU1() const;
-    u8 GetU8() const;
-    u32 GetU32() const;
-    u64 GetU64() const;
-
-private:
-    Type type;
-
-    union {
-        Inst* inst; // type == Type::Opaque
-        Arm::Reg imm_regref;
-        Arm::ExtReg imm_extregref;
-        bool imm_u1;
-        u8 imm_u8;
-        u32 imm_u32;
-        u64 imm_u64;
-    } inner;
-};
-
-/**
- * A representation of a microinstruction. A single ARM/Thumb instruction may be
- * converted into zero or more microinstructions.
- */
-class Inst final : public Common::IntrusiveListNode<Inst> {
-public:
-    Inst(Opcode op) : op(op) {}
-
-    bool HasUses() const { return use_count > 0; }
-
-    /// Get the microop this microinstruction represents.
-    Opcode GetOpcode() const { return op; }
-    /// Get the type this instruction returns.
-    Type GetType() const { return GetTypeOf(op); }
-    /// Get the number of arguments this instruction has.
-    size_t NumArgs() const { return GetNumArgsOf(op); }
-
-    Value GetArg(size_t index) const;
-    void SetArg(size_t index, Value value);
-
-    void Invalidate();
-
-    void ReplaceUsesWith(Value& replacement);
-
-    size_t use_count = 0;
-    Inst* carry_inst = nullptr;
-    Inst* overflow_inst = nullptr;
-
-private:
-    void Use(Value& value);
-    void UndoUse(Value& value);
-
-    Opcode op;
-    std::array<Value, 3> args;
-};
-
-namespace Term {
-
-struct Invalid {};
-
-/**
- * This terminal instruction calls the interpreter, starting at `next`.
- * The interpreter must interpret exactly one instruction.
- */
-struct Interpret {
-    explicit Interpret(const Arm::LocationDescriptor& next_) : next(next_) {}
-    Arm::LocationDescriptor next; ///< Location at which interpretation starts.
-};
-
-/**
- * This terminal instruction returns control to the dispatcher.
- * The dispatcher will use the value in R15 to determine what comes next.
- */
-struct ReturnToDispatch {};
-
-/**
- * This terminal instruction jumps to the basic block described by `next` if we have enough
- * cycles remaining. If we do not have enough cycles remaining, we return to the
- * dispatcher, which will return control to the host.
- */
-struct LinkBlock {
-    explicit LinkBlock(const Arm::LocationDescriptor& next_) : next(next_) {}
-    Arm::LocationDescriptor next; ///< Location descriptor for next block.
-};
-
-/**
- * This terminal instruction jumps to the basic block described by `next` unconditionally.
- * This is an optimization and MUST only be emitted when this is guaranteed not to result
- * in hanging, even in the face of other optimizations. (In practice, this means that only
- * forward jumps to short-ish blocks would use this instruction.)
- * A backend that doesn't support this optimization may choose to implement this exactly
- * as LinkBlock.
- */
-struct LinkBlockFast {
-    explicit LinkBlockFast(const Arm::LocationDescriptor& next_) : next(next_) {}
-    Arm::LocationDescriptor next; ///< Location descriptor for next block.
-};
-
-/**
- * This terminal instruction checks the top of the Return Stack Buffer against R15.
- * If RSB lookup fails, control is returned to the dispatcher.
- * This is an optimization for faster function calls. A backend that doesn't support
- * this optimization or doesn't have a RSB may choose to implement this exactly as
- * ReturnToDispatch.
- */
-struct PopRSBHint {};
-
-struct If;
-struct CheckHalt;
-/// A Terminal is the terminal instruction in a MicroBlock.
-using Terminal = boost::variant<
-        Invalid,
-        Interpret,
-        ReturnToDispatch,
-        LinkBlock,
-        LinkBlockFast,
-        PopRSBHint,
-        boost::recursive_wrapper<If>,
-        boost::recursive_wrapper<CheckHalt>
->;
-
-/**
- * This terminal instruction conditionally executes one terminal or another depending
- * on the run-time state of the ARM flags.
- */
-struct If {
-    If(Arm::Cond if_, Terminal then_, Terminal else_) : if_(if_), then_(then_), else_(else_) {}
-    Arm::Cond if_;
-    Terminal then_;
-    Terminal else_;
-};
-
-/**
- * This terminal instruction checks if a halt was requested. If it wasn't, else_ is
- * executed.
- */
-struct CheckHalt {
-    CheckHalt(Terminal else_) : else_(else_) {}
-    Terminal else_;
-};
-
-} // namespace Term
-
-using Term::Terminal;
-
-/**
- * A basic block. It consists of zero or more instructions followed by exactly one terminal.
- * Note that this is a linear IR and not a pure tree-based IR: i.e.: there is an ordering to
- * the microinstructions. This only matters before chaining is done in order to correctly
- * order memory accesses.
- */
-class Block final {
-public:
-    explicit Block(const Arm::LocationDescriptor& location) : location(location) {}
-
-    /// Description of the starting location of this block
-    Arm::LocationDescriptor location;
-    /// Conditional to pass in order to execute this block
-    Arm::Cond cond = Arm::Cond::AL;
-    /// Block to execute next if `cond` did not pass.
-    boost::optional<Arm::LocationDescriptor> cond_failed = {};
-
-    /// List of instructions in this block.
-    Common::IntrusiveList<Inst> instructions;
-    /// Memory pool for instruction list
-    std::unique_ptr<Common::Pool> instruction_alloc_pool = std::make_unique<Common::Pool>(sizeof(Inst), 4096);
-    /// Terminal instruction of this block.
-    Terminal terminal = Term::Invalid{};
-
-    /// Number of cycles this block takes to execute.
-    size_t cycle_count = 0;
-};
-
-/// Returns a string representation of the contents of block. Intended for debugging.
-std::string DumpBlock(const IR::Block& block);
-
-} // namespace IR
-} // namespace Dynarmic

src/frontend/ir/ir_emitter.h

@@ -6,9 +6,21 @@
 
 #pragma once
 
+#include <initializer_list>
+
+#include "common/common_types.h"
 #include "frontend/arm_types.h"
-#include "frontend/ir/ir.h"
+#include "frontend/ir/basic_block.h"
 #include "frontend/ir/opcodes.h"
+#include "frontend/ir/terminal.h"
+#include "frontend/ir/value.h"
+
+// ARM JIT Microinstruction Intermediate Representation
+//
+// This intermediate representation is an SSA IR. It is designed primarily for analysis,
+// though it can be lowered into a reduced form for interpretation. Each IR node (Value)
+// is a microinstruction of an idealised ARM CPU. The choice of microinstructions is made
+// not based on any existing microarchitecture but on ease of implementation.
 
 namespace Dynarmic {
 namespace Arm {

src/frontend/ir/microinstruction.cpp

@@ -0,0 +1,85 @@
+/* This file is part of the dynarmic project.
+ * Copyright (c) 2016 MerryMage
+ * This software may be used and distributed according to the terms of the GNU
+ * General Public License version 2 or any later version.
+ */
+
+#include "common/assert.h"
+#include "frontend/ir/microinstruction.h"
+
+namespace Dynarmic {
+namespace IR {
+
+Value Inst::GetArg(size_t index) const {
+    DEBUG_ASSERT(index < GetNumArgsOf(op));
+    DEBUG_ASSERT(!args[index].IsEmpty());
+
+    return args[index];
+}
+
+void Inst::SetArg(size_t index, Value value) {
+    DEBUG_ASSERT(index < GetNumArgsOf(op));
+    DEBUG_ASSERT(value.GetType() == GetArgTypeOf(op, index) || Type::Opaque == GetArgTypeOf(op, index));
+
+    if (!args[index].IsImmediate()) {
+        UndoUse(args[index]);
+    }
+    if (!value.IsImmediate()) {
+        Use(value);
+    }
+
+    args[index] = value;
+}
+
+void Inst::Invalidate() {
+    for (auto& value : args) {
+        if (!value.IsImmediate()) {
+            UndoUse(value);
+        }
+    }
+}
+
+void Inst::ReplaceUsesWith(Value& replacement) {
+    Invalidate();
+
+    op = Opcode::Identity;
+
+    if (!replacement.IsImmediate()) {
+        Use(replacement);
+    }
+
+    args[0] = replacement;
+}
+
+void Inst::Use(Value& value) {
+    value.GetInst()->use_count++;
+
+    switch (op){
+        case Opcode::GetCarryFromOp:
+            value.GetInst()->carry_inst = this;
+            break;
+        case Opcode::GetOverflowFromOp:
+            value.GetInst()->overflow_inst = this;
+            break;
+        default:
+            break;
+    }
+}
+
+void Inst::UndoUse(Value& value) {
+    value.GetInst()->use_count--;
+
+    switch (op){
+        case Opcode::GetCarryFromOp:
+            value.GetInst()->carry_inst = nullptr;
+            break;
+        case Opcode::GetOverflowFromOp:
+            value.GetInst()->overflow_inst = nullptr;
+            break;
+        default:
+            break;
+    }
+}
+
+} // namespace IR
+} // namespace Dynarmic

src/frontend/ir/microinstruction.h

@@ -0,0 +1,56 @@
+/* This file is part of the dynarmic project.
+ * Copyright (c) 2016 MerryMage
+ * This software may be used and distributed according to the terms of the GNU
+ * General Public License version 2 or any later version.
+ */
+
+#pragma once
+
+#include <array>
+
+#include "common/common_types.h"
+#include "common/intrusive_list.h"
+#include "frontend/ir/opcodes.h"
+#include "frontend/ir/value.h"
+
+namespace Dynarmic {
+namespace IR {
+
+/**
+ * A representation of a microinstruction. A single ARM/Thumb instruction may be
+ * converted into zero or more microinstructions.
+ */
+class Inst final : public Common::IntrusiveListNode<Inst> {
+public:
+    Inst(Opcode op) : op(op) {}
+
+    bool HasUses() const { return use_count > 0; }
+
+    /// Get the microop this microinstruction represents.
+    Opcode GetOpcode() const { return op; }
+    /// Get the type this instruction returns.
+    Type GetType() const { return GetTypeOf(op); }
+    /// Get the number of arguments this instruction has.
+    size_t NumArgs() const { return GetNumArgsOf(op); }
+
+    Value GetArg(size_t index) const;
+    void SetArg(size_t index, Value value);
+
+    void Invalidate();
+
+    void ReplaceUsesWith(Value& replacement);
+
+    size_t use_count = 0;
+    Inst* carry_inst = nullptr;
+    Inst* overflow_inst = nullptr;
+
+private:
+    void Use(Value& value);
+    void UndoUse(Value& value);
+
+    Opcode op;
+    std::array<Value, 3> args;
+};
+
+} // namespace IR
+} // namespace Dynarmic

src/frontend/ir/terminal.h

@@ -0,0 +1,106 @@
+/* This file is part of the dynarmic project.
+ * Copyright (c) 2016 MerryMage
+ * This software may be used and distributed according to the terms of the GNU
+ * General Public License version 2 or any later version.
+ */
+
+#pragma once
+
+#include <boost/variant.hpp>
+
+#include "common/common_types.h"
+#include "frontend/arm_types.h"
+
+namespace Dynarmic {
+namespace IR {
+namespace Term {
+
+struct Invalid {};
+
+/**
+ * This terminal instruction calls the interpreter, starting at `next`.
+ * The interpreter must interpret exactly one instruction.
+ */
+struct Interpret {
+    explicit Interpret(const Arm::LocationDescriptor& next_) : next(next_) {}
+    Arm::LocationDescriptor next; ///< Location at which interpretation starts.
+};
+
+/**
+ * This terminal instruction returns control to the dispatcher.
+ * The dispatcher will use the value in R15 to determine what comes next.
+ */
+struct ReturnToDispatch {};
+
+/**
+ * This terminal instruction jumps to the basic block described by `next` if we have enough
+ * cycles remaining. If we do not have enough cycles remaining, we return to the
+ * dispatcher, which will return control to the host.
+ */
+struct LinkBlock {
+    explicit LinkBlock(const Arm::LocationDescriptor& next_) : next(next_) {}
+    Arm::LocationDescriptor next; ///< Location descriptor for next block.
+};
+
+/**
+ * This terminal instruction jumps to the basic block described by `next` unconditionally.
+ * This is an optimization and MUST only be emitted when this is guaranteed not to result
+ * in hanging, even in the face of other optimizations. (In practice, this means that only
+ * forward jumps to short-ish blocks would use this instruction.)
+ * A backend that doesn't support this optimization may choose to implement this exactly
+ * as LinkBlock.
+ */
+struct LinkBlockFast {
+    explicit LinkBlockFast(const Arm::LocationDescriptor& next_) : next(next_) {}
+    Arm::LocationDescriptor next; ///< Location descriptor for next block.
+};
+
+/**
+ * This terminal instruction checks the top of the Return Stack Buffer against R15.
+ * If RSB lookup fails, control is returned to the dispatcher.
+ * This is an optimization for faster function calls. A backend that doesn't support
+ * this optimization or doesn't have a RSB may choose to implement this exactly as
+ * ReturnToDispatch.
+ */
+struct PopRSBHint {};
+
+struct If;
+struct CheckHalt;
+/// A Terminal is the terminal instruction in a MicroBlock.
+using Terminal = boost::variant<
+        Invalid,
+        Interpret,
+        ReturnToDispatch,
+        LinkBlock,
+        LinkBlockFast,
+        PopRSBHint,
+        boost::recursive_wrapper<If>,
+        boost::recursive_wrapper<CheckHalt>
+>;
+
+/**
+ * This terminal instruction conditionally executes one terminal or another depending
+ * on the run-time state of the ARM flags.
+ */
+struct If {
+    If(Arm::Cond if_, Terminal then_, Terminal else_) : if_(if_), then_(then_), else_(else_) {}
+    Arm::Cond if_;
+    Terminal then_;
+    Terminal else_;
+};
+
+/**
+ * This terminal instruction checks if a halt was requested. If it wasn't, else_ is
+ * executed.
+ */
+struct CheckHalt {
+    CheckHalt(Terminal else_) : else_(else_) {}
+    Terminal else_;
+};
+
+} // namespace Term
+
+using Term::Terminal;
+
+} // namespace IR
+} // namespace Dynarmic

src/frontend/ir/value.cpp

@@ -0,0 +1,107 @@
+/* This file is part of the dynarmic project.
+ * Copyright (c) 2016 MerryMage
+ * This software may be used and distributed according to the terms of the GNU
+ * General Public License version 2 or any later version.
+ */
+
+#include "common/assert.h"
+#include "frontend/ir/microinstruction.h"
+#include "frontend/ir/value.h"
+
+namespace Dynarmic {
+namespace IR {
+
+Value::Value(Inst* value) : type(Type::Opaque) {
+    inner.inst = value;
+}
+
+Value::Value(Arm::Reg value) : type(Type::RegRef) {
+    inner.imm_regref = value;
+}
+
+Value::Value(Arm::ExtReg value) : type(Type::ExtRegRef) {
+    inner.imm_extregref = value;
+}
+
+Value::Value(bool value) : type(Type::U1) {
+    inner.imm_u1 = value;
+}
+
+Value::Value(u8 value) : type(Type::U8) {
+    inner.imm_u8 = value;
+}
+
+Value::Value(u32 value) : type(Type::U32) {
+    inner.imm_u32 = value;
+}
+
+Value::Value(u64 value) : type(Type::U64) {
+    inner.imm_u64 = value;
+}
+
+bool Value::IsImmediate() const {
+    if (type == Type::Opaque)
+        return inner.inst->GetOpcode() == Opcode::Identity ? inner.inst->GetArg(0).IsImmediate() : false;
+    return true;
+}
+
+bool Value::IsEmpty() const {
+    return type == Type::Void;
+}
+
+Type Value::GetType() const {
+    if (type == Type::Opaque) {
+        if (inner.inst->GetOpcode() == Opcode::Identity) {
+            return inner.inst->GetArg(0).GetType();
+        } else {
+            return inner.inst->GetType();
+        }
+    }
+    return type;
+}
+
+Arm::Reg Value::GetRegRef() const {
+    DEBUG_ASSERT(type == Type::RegRef);
+    return inner.imm_regref;
+}
+
+Arm::ExtReg Value::GetExtRegRef() const {
+    DEBUG_ASSERT(type == Type::ExtRegRef);
+    return inner.imm_extregref;
+}
+
+Inst* Value::GetInst() const {
+    DEBUG_ASSERT(type == Type::Opaque);
+    return inner.inst;
+}
+
+bool Value::GetU1() const {
+    if (type == Type::Opaque && inner.inst->GetOpcode() == Opcode::Identity)
+        return inner.inst->GetArg(0).GetU1();
+    DEBUG_ASSERT(type == Type::U1);
+    return inner.imm_u1;
+}
+
+u8 Value::GetU8() const {
+    if (type == Type::Opaque && inner.inst->GetOpcode() == Opcode::Identity)
+        return inner.inst->GetArg(0).GetU8();
+    DEBUG_ASSERT(type == Type::U8);
+    return inner.imm_u8;
+}
+
+u32 Value::GetU32() const {
+    if (type == Type::Opaque && inner.inst->GetOpcode() == Opcode::Identity)
+        return inner.inst->GetArg(0).GetU32();
+    DEBUG_ASSERT(type == Type::U32);
+    return inner.imm_u32;
+}
+
+u64 Value::GetU64() const {
+    if (type == Type::Opaque && inner.inst->GetOpcode() == Opcode::Identity)
+        return inner.inst->GetArg(0).GetU64();
+    DEBUG_ASSERT(type == Type::U64);
+    return inner.imm_u64;
+}
+
+} // namespace IR
+} // namespace Dynarmic

src/frontend/ir/value.h

@@ -0,0 +1,59 @@
+/* This file is part of the dynarmic project.
+ * Copyright (c) 2016 MerryMage
+ * This software may be used and distributed according to the terms of the GNU
+ * General Public License version 2 or any later version.
+ */
+
+#pragma once
+
+#include "common/common_types.h"
+#include "frontend/arm_types.h"
+
+namespace Dynarmic {
+namespace IR {
+
+class Inst;
+
+/**
+ * A representation of a value in the IR.
+ * A value may either be an immediate or the result of a microinstruction.
+ */
+struct Value final {
+public:
+    Value() : type(Type::Void) {}
+    explicit Value(Inst* value);
+    explicit Value(Arm::Reg value);
+    explicit Value(Arm::ExtReg value);
+    explicit Value(bool value);
+    explicit Value(u8 value);
+    explicit Value(u32 value);
+    explicit Value(u64 value);
+
+    bool IsEmpty() const;
+    bool IsImmediate() const;
+    Type GetType() const;
+
+    Inst* GetInst() const;
+    Arm::Reg GetRegRef() const;
+    Arm::ExtReg GetExtRegRef() const;
+    bool GetU1() const;
+    u8 GetU8() const;
+    u32 GetU32() const;
+    u64 GetU64() const;
+
+private:
+    Type type;
+
+    union {
+        Inst* inst; // type == Type::Opaque
+        Arm::Reg imm_regref;
+        Arm::ExtReg imm_extregref;
+        bool imm_u1;
+        u8 imm_u8;
+        u32 imm_u32;
+        u64 imm_u64;
+    } inner;
+};
+
+} // namespace IR
+} // namespace Dynarmic

src/frontend/translate/translate.cpp

@@ -5,8 +5,8 @@
  */
 
 #include "frontend/arm_types.h"
-#include "frontend/ir/ir.h"
-#include "translate.h"
+#include "frontend/ir/basic_block.h"
+#include "frontend/translate/translate.h"
 
 namespace Dynarmic {
 namespace Arm {

src/frontend/translate/translate.h

@@ -5,10 +5,15 @@
  */
 #pragma once
 
+#include "common/common_types.h"
 #include "frontend/arm_types.h"
-#include "frontend/ir/ir.h"
 
 namespace Dynarmic {
+
+namespace IR {
+    class Block;
+} // namespace IR
+
 namespace Arm {
 
 using MemoryRead32FuncType = u32 (*)(u32 vaddr);

src/frontend/translate/translate_arm.cpp

@@ -8,7 +8,7 @@
 #include "frontend/arm_types.h"
 #include "frontend/decoder/arm.h"
 #include "frontend/decoder/vfp2.h"
-#include "frontend/ir/ir.h"
+#include "frontend/ir/basic_block.h"
 #include "frontend/translate/translate.h"
 #include "frontend/translate/translate_arm/translate_arm.h"
 

src/ir_opt/dead_code_elimination_pass.cpp

@@ -5,7 +5,8 @@
  */
 
 #include "common/assert.h"
-#include "frontend/ir/ir.h"
+#include "frontend/ir/basic_block.h"
+#include "frontend/ir/opcodes.h"
 #include "ir_opt/passes.h"
 
 namespace Dynarmic {

src/ir_opt/get_set_elimination_pass.cpp

@@ -4,8 +4,12 @@
  * General Public License version 2 or any later version.
  */
 
+#include <array>
+
 #include "common/assert.h"
-#include "frontend/ir/ir.h"
+#include "common/common_types.h"
+#include "frontend/ir/basic_block.h"
+#include "frontend/ir/value.h"
 #include "ir_opt/passes.h"
 
 namespace Dynarmic {

src/ir_opt/verification_pass.cpp

@@ -7,7 +7,9 @@
 #include <map>
 
 #include "common/assert.h"
-#include "frontend/ir/ir.h"
+#include "common/common_types.h"
+#include "frontend/ir/basic_block.h"
+#include "frontend/ir/microinstruction.h"
 #include "ir_opt/passes.h"
 
 namespace Dynarmic {

tests/arm/fuzz_arm.cpp

@@ -4,9 +4,14 @@
  * General Public License version 2 or any later version.
  */
 
+#include <algorithm>
+#include <array>
 #include <cinttypes>
+#include <cstdio>
 #include <cstring>
 #include <functional>
+#include <tuple>
+#include <vector>
 
 #include <catch.hpp>
 
@@ -14,7 +19,7 @@
 #include "common/common_types.h"
 #include "frontend/arm_types.h"
 #include "frontend/disassembler/disassembler.h"
-#include "frontend/ir/ir.h"
+#include "frontend/ir/basic_block.h"
 #include "frontend/translate/translate.h"
 #include "interface/interface.h"
 #include "ir_opt/passes.h"

tests/arm/fuzz_thumb.cpp

@@ -4,23 +4,27 @@
  * General Public License version 2 or any later version.
  */
 
+#include <algorithm>
+#include <array>
 #include <cinttypes>
+#include <cstdio>
 #include <cstring>
 #include <functional>
+#include <tuple>
 
 #include <catch.hpp>
 
 #include "common/bit_util.h"
 #include "common/common_types.h"
 #include "frontend/disassembler/disassembler.h"
+#include "frontend/ir/basic_block.h"
+#include "frontend/translate/translate.h"
 #include "interface/interface.h"
+#include "ir_opt/passes.h"
 #include "rand_int.h"
 #include "skyeye_interpreter/dyncom/arm_dyncom_interpreter.h"
 #include "skyeye_interpreter/skyeye_common/armstate.h"
 
-#include "frontend/translate/translate.h"
-#include "ir_opt/passes.h"
-
 struct WriteRecord {
     size_t size;
     u32 address;