Basic disassembler

src/binary_translation/CMakeLists.txt

@@ -3,17 +3,23 @@ set(SRCS
 			CodeGen.cpp
 			ModuleGen.cpp
 			Disassembler.cpp
-			Instruction.cpp
+			
+			Instructions/Instruction.cpp
+			Instructions/DataProcessing.cpp
             )
 set(HEADERS
 			CodeGen.h
 			ModuleGen.h
 			Disassembler.h
-			Instruction.h
+			
+			Instructions/Types.h
+			Instructions/Instruction.h
+			Instructions/DataProcessing.h
             )
 			
 create_directory_groups(${SRCS} ${HEADERS})
 
+include_directories(.)
 add_executable(binary_translate ${SRCS} ${HEADERS})
 target_link_libraries(binary_translate ${llvm_libs})
 target_link_libraries(binary_translate core common video_core)

src/binary_translation/Disassembler.cpp

@@ -1,5 +1,5 @@
 #include "Disassembler.h"
-#include "Instruction.h"
+#include "Instructions/Instruction.h"
 #include <vector>
 
 std::vector<RegisterInstructionBase::CreateFunctionType> g_read_functions;
@@ -9,7 +9,7 @@ RegisterInstructionBase::RegisterInstructionBase(CreateFunctionType create_funct
     g_read_functions.push_back(create_function);
 }
 
-std::unique_ptr<Instruction> Disassmbler::Disassemble(u32 instruction, u32 address)
+std::unique_ptr<Instruction> Disassembler::Disassemble(u32 instruction, u32 address)
 {
     for (auto read_function : g_read_functions)
     {

src/binary_translation/Disassembler.h

@@ -3,7 +3,7 @@
 
 class Instruction;
 
-class Disassmbler
+class Disassembler
 {
 public:
     /*
@@ -32,8 +32,8 @@ public:
 private:
     static Instruction *Create(u32 instruction, u32 address)
     {
-        auto result = new DerivedInstruction(instruction, address);
+        auto result = new DerivedInstruction();
-        if (!result->Read())
+        if (!result->Read(instruction, address))
         {
             delete result;
             return nullptr;

src/binary_translation/Instructions/DataProcessing.cpp

@@ -0,0 +1,15 @@
+#include "DataProcessing.h"
+#include "Disassembler.h"
+
+static RegisterInstruction<DataProcessing> register_instruction;
+
+bool DataProcessing::Decode()
+{
+    if (ReadFields({ FieldDef<4>(&cond), FieldDef<3>(1), FieldDef<4>(&short_op), FieldDef<1>(&s), FieldDef<4>(&rn),
+                     FieldDef<4>(&rd), FieldDef<12>(&imm12) }))
+    {
+        form = Form::Immediate;
+        return true;
+    }
+    return false;
+}

src/binary_translation/Instructions/DataProcessing.h

@@ -0,0 +1,36 @@
+#include "Instruction.h"
+#include "Types.h"
+
+/*
+ * Data processing instructions
+ * ARMv7-A 5.2.1 (register), 5.2.2 (register-shifted register, 5.2.3 (immediate)
+ */
+
+class DataProcessing : public Instruction
+{
+public:
+    /*
+     * The 4 bit op types (1 = 0001x: BitwiseXor, etc...)
+     */
+    enum class ShortOpType
+    {
+        BitwiseAnd = 0, BitwiseXor, Subtract, RevSubtract, Add, AddWithCarry, SubtractWithCarry, ReverseSubtractWithCarry,
+        // Compare, Test, Misc
+        BitwiseOr = 12, Move, BitwiseBitClear, BitwiseNot
+    };
+    enum class Form
+    {
+        Register, RegisterShiftedRegister, Immediate
+    };
+
+public:
+    virtual bool Decode() override;
+private:
+    Form form;
+    Condition cond;
+    ShortOpType short_op;
+    bool s;
+    Register rn;
+    Register rd;
+    u32 imm12;
+};

src/binary_translation/Instructions/Instruction.cpp

@@ -1,7 +1,7 @@
 #include "Instruction.h"
 #include <cassert>
 
-Instruction::Instruction(u32 instruction, u32 address) : instruction(instruction), address(address)
+Instruction::Instruction()
 {
 }
 
@@ -9,6 +9,14 @@ Instruction::~Instruction()
 {
 }
 
+bool Instruction::Read(u32 instruction, u32 address)
+{
+    this->instruction = instruction;
+    this->address = address;
+    // Call the read of derived class
+    return Decode();
+}
+
 bool Instruction::ReadFields(const std::initializer_list<FieldDefObject> &fields)
 {
     size_t total_bit_count = 0;

src/binary_translation/Instructions/Instruction.h

@@ -7,15 +7,19 @@ class Instruction
 protected:
     class FieldDefObject;
 public:
-    Instruction(u32 instruction, u32 address);
+    Instruction();
     virtual ~Instruction();
 
     /*
      * Reads the instruction.
      * Returns true on success, or false otherwise
      */
-    virtual bool Read() = 0;
+    bool Read(u32 instruction, u32 address);
 protected:
+    /*
+     * Derived classes should override this, and implement it by calling ReadFields
+     */
+    virtual bool Decode() = 0;
     /*
      * Reads fields from the instruction
      * The fields come most significant first

src/binary_translation/Instructions/Types.h

@@ -0,0 +1,14 @@
+/*
+ * A register in a broad sense: R0-R15, and flags
+ */
+enum class Register
+{
+    R0, R1, R2, R3, R4, R5, R6, R7,
+    R8, R9, R10, R11, R12, SP, LR, PC,
+    N, Z, C, V
+};
+
+enum class Condition
+{
+    EQ, NE, CS, CC, MI, PL, VS, VC, HI, LS, GE, LT, GT, LE, AL, Invalid
+};

src/binary_translation/ModuleGen.cpp

@@ -1,4 +1,8 @@
 #include "ModuleGen.h"
+#include "Disassembler.h"
+#include "core/loader/loader.h"
+#include "core/mem_map.h"
+#include "Instructions/Instruction.h"
 #include <llvm/IR/Function.h>
 #include <llvm/IR/GlobalVariable.h>
 
@@ -19,6 +23,8 @@ void ModuleGen::Run()
     GenerateGlobals();
     GenerateCanRunFunction();
     GenerateRunFunction();
+
+    GenerateBlocks();
 }
 
 void ModuleGen::GenerateGlobals()
@@ -49,4 +55,16 @@ void ModuleGen::GenerateRunFunction()
 
     ir_builder->SetInsertPoint(basic_block);
     ir_builder->CreateRetVoid();
+}
+
+void ModuleGen::GenerateBlocks()
+{
+    for (auto i = Loader::ROMCodeStart; i <= Loader::ROMCodeStart + Loader::ROMCodeSize - 4; i += 4)
+    {
+        auto instruction = Disassembler::Disassemble(Memory::Read32(i), i);
+        if (instruction != nullptr)
+        {
+            LOG_DEBUG(BinaryTranslator, "Instruction at %08x", i);
+        }
+    }
 }

src/binary_translation/ModuleGen.h

@@ -15,6 +15,7 @@ public:
     void GenerateGlobals();
     void GenerateCanRunFunction();
     void GenerateRunFunction();
+    void GenerateBlocks();
 private:
     std::unique_ptr<llvm::IRBuilder<>> ir_builder;
     llvm::Module *module;