A64: Implement AESE

src/backend_x64/emit_x64_aes.cpp

@@ -16,10 +16,10 @@ namespace Dynarmic::BackendX64 {
 
 using namespace Xbyak::util;
     
-using MixColumnsFn = void(Common::AESState&, const Common::AESState&);
+using AESFn = void(Common::AESState&, const Common::AESState&);
 
-static void EmitMixColumns(std::array<Argument, 3> args, EmitContext& ctx, BlockOfCode& code,
+static void EmitAESFunction(std::array<Argument, 3> args, EmitContext& ctx, BlockOfCode& code,
-                           IR::Inst* inst, MixColumnsFn fn) {
+                            IR::Inst* inst, AESFn fn) {
     constexpr u32 stack_space = static_cast<u32>(sizeof(Common::AESState)) * 2;
     const Xbyak::Xmm input = ctx.reg_alloc.UseXmm(args[0]);
     ctx.reg_alloc.EndOfAllocScope();
@@ -41,6 +41,12 @@ static void EmitMixColumns(std::array<Argument, 3> args, EmitContext& ctx, Block
     ctx.reg_alloc.DefineValue(inst, xmm0);
 }
 
+void EmitX64::EmitAESEncryptSingleRound(EmitContext& ctx, IR::Inst* inst) {
+    auto args = ctx.reg_alloc.GetArgumentInfo(inst);
+
+    EmitAESFunction(args, ctx, code, inst, Common::EncryptSingleRound);
+}
+
 void EmitX64::EmitAESInverseMixColumns(EmitContext& ctx, IR::Inst* inst) {
      auto args = ctx.reg_alloc.GetArgumentInfo(inst);
 
@@ -52,13 +58,13 @@ void EmitX64::EmitAESInverseMixColumns(EmitContext& ctx, IR::Inst* inst) {
 
         ctx.reg_alloc.DefineValue(inst, result);
     } else {
-        EmitMixColumns(args, ctx, code, inst, Common::InverseMixColumns);
+        EmitAESFunction(args, ctx, code, inst, Common::InverseMixColumns);
     }
 }
 
 void EmitX64::EmitAESMixColumns(EmitContext& ctx, IR::Inst* inst) {
     auto args = ctx.reg_alloc.GetArgumentInfo(inst);
-    EmitMixColumns(args, ctx, code, inst, Common::MixColumns);
+    EmitAESFunction(args, ctx, code, inst, Common::MixColumns);
 }
 
 } // namespace Dynarmic::BackendX64

src/common/aes.cpp

@@ -11,6 +11,27 @@
 
 namespace Dynarmic::Common {
 
+// See section 5.1.1 Figure 7 in FIPS 197
+constexpr std::array<u8, 256> substitution_box{{
+    // 0   1     2     3     4     5     6     7     8     9     A     B     C     D     E     F
+    0x63, 0x7C, 0x77, 0x7B, 0xF2, 0x6B, 0x6F, 0xC5, 0x30, 0x01, 0x67, 0x2B, 0xFE, 0xD7, 0xAB, 0x76,
+    0xCA, 0x82, 0xC9, 0x7D, 0xFA, 0x59, 0x47, 0xF0, 0xAD, 0xD4, 0xA2, 0xAF, 0x9C, 0xA4, 0x72, 0xC0,
+    0xB7, 0xFD, 0x93, 0x26, 0x36, 0x3F, 0xF7, 0xCC, 0x34, 0xA5, 0xE5, 0xF1, 0x71, 0xD8, 0x31, 0x15,
+    0x04, 0xC7, 0x23, 0xC3, 0x18, 0x96, 0x05, 0x9A, 0x07, 0x12, 0x80, 0xE2, 0xEB, 0x27, 0xB2, 0x75,
+    0x09, 0x83, 0x2C, 0x1A, 0x1B, 0x6E, 0x5A, 0xA0, 0x52, 0x3B, 0xD6, 0xB3, 0x29, 0xE3, 0x2F, 0x84,
+    0x53, 0xD1, 0x00, 0xED, 0x20, 0xFC, 0xB1, 0x5B, 0x6A, 0xCB, 0xBE, 0x39, 0x4A, 0x4C, 0x58, 0xCF,
+    0xD0, 0xEF, 0xAA, 0xFB, 0x43, 0x4D, 0x33, 0x85, 0x45, 0xF9, 0x02, 0x7F, 0x50, 0x3C, 0x9F, 0xA8,
+    0x51, 0xA3, 0x40, 0x8F, 0x92, 0x9D, 0x38, 0xF5, 0xBC, 0xB6, 0xDA, 0x21, 0x10, 0xFF, 0xF3, 0xD2,
+    0xCD, 0x0C, 0x13, 0xEC, 0x5F, 0x97, 0x44, 0x17, 0xC4, 0xA7, 0x7E, 0x3D, 0x64, 0x5D, 0x19, 0x73,
+    0x60, 0x81, 0x4F, 0xDC, 0x22, 0x2A, 0x90, 0x88, 0x46, 0xEE, 0xB8, 0x14, 0xDE, 0x5E, 0x0B, 0xDB,
+    0xE0, 0x32, 0x3A, 0x0A, 0x49, 0x06, 0x24, 0x5C, 0xC2, 0xD3, 0xAC, 0x62, 0x91, 0x95, 0xE4, 0x79,
+    0xE7, 0xC8, 0x37, 0x6D, 0x8D, 0xD5, 0x4E, 0xA9, 0x6C, 0x56, 0xF4, 0xEA, 0x65, 0x7A, 0xAE, 0x08,
+    0xBA, 0x78, 0x25, 0x2E, 0x1C, 0xA6, 0xB4, 0xC6, 0xE8, 0xDD, 0x74, 0x1F, 0x4B, 0xBD, 0x8B, 0x8A,
+    0x70, 0x3E, 0xB5, 0x66, 0x48, 0x03, 0xF6, 0x0E, 0x61, 0x35, 0x57, 0xB9, 0x86, 0xC1, 0x1D, 0x9E,
+    0xE1, 0xF8, 0x98, 0x11, 0x69, 0xD9, 0x8E, 0x94, 0x9B, 0x1E, 0x87, 0xE9, 0xCE, 0x55, 0x28, 0xDF,
+    0x8C, 0xA1, 0x89, 0x0D, 0xBF, 0xE6, 0x42, 0x68, 0x41, 0x99, 0x2D, 0x0F, 0xB0, 0x54, 0xBB, 0x16
+}};
+
 // See section 4.2.1 in FIPS 197.
 static constexpr u8 xtime(u8 x) {
     return static_cast<u8>((x << 1) ^ (((x >> 7) & 1) * 0x1B));
@@ -25,6 +46,50 @@ static constexpr u8 Multiply(u8 x, u8 y) {
             ((y >> 4 & 1) * xtime(xtime(xtime(xtime(x))))));
 }
 
+static void ShiftRows(AESState& out_state, const AESState& state) {
+    // Move zeroth row over
+    out_state[0] = state[0];
+    out_state[4] = state[4];
+    out_state[8] = state[8];
+    out_state[12] = state[12];
+
+    // Rotate first row 1 column left.
+    u8 temp = state[1];
+    out_state[1] = state[5];
+    out_state[5] = state[9];
+    out_state[9] = state[13];
+    out_state[13] = temp;
+
+    // Rotate second row 2 columns left
+    temp = state[2];
+    out_state[2] = state[10];
+    out_state[10] = temp;
+
+    temp = state[6];
+    out_state[6] = state[14];
+    out_state[14] = temp;
+
+    // Rotate third row 3 columns left
+    temp = state[3];
+    out_state[3] = state[15];
+    out_state[15] = state[11];
+    out_state[11] = state[7];
+    out_state[7] = temp;
+}
+
+static void SubBytes(AESState& state) {
+    for (size_t i = 0; i < 4; i++) {
+        for (size_t j = 0; j < 4; j++) {
+            state[4 * i + j] = substitution_box[state[4 * i + j]];
+        }
+    }
+}
+
+void EncryptSingleRound(AESState& out_state, const AESState& state) {
+    ShiftRows(out_state, state);
+    SubBytes(out_state);
+}
+
 void MixColumns(AESState& out_state, const AESState& state) {
     for (size_t i = 0; i < out_state.size(); i += 4) {
         const u8 a = state[i];

src/common/aes.h

@@ -13,6 +13,9 @@ namespace Dynarmic::Common {
 
 using AESState = std::array<u8, 16>;
 
+// Assumes the state has already been XORed by the round key.
+void EncryptSingleRound(AESState& out_state, const AESState& state);
+
 void MixColumns(AESState& out_state, const AESState& state);
 void InverseMixColumns(AESState& out_state, const AESState& state);
 

src/frontend/A64/decoder/a64.inc

@@ -329,7 +329,7 @@ INST(UMSUBL,                 "UMSUBL",                                    "10011
 //INST(UMULH,                  "UMULH",                                     "10011011110mmmmm011111nnnnnddddd")
 
 // Data Processing - FP and SIMD - AES
-//INST(AESE,                   "AESE",                                      "0100111000101000010010nnnnnddddd")
+INST(AESE,                   "AESE",                                      "0100111000101000010010nnnnnddddd")
 //INST(AESD,                   "AESD",                                      "0100111000101000010110nnnnnddddd")
 INST(AESMC,                  "AESMC",                                     "0100111000101000011010nnnnnddddd")
 INST(AESIMC,                 "AESIMC",                                    "0100111000101000011110nnnnnddddd")

src/frontend/A64/translate/impl/simd_aes.cpp

@@ -8,6 +8,16 @@
 
 namespace Dynarmic::A64 {
 
+bool TranslatorVisitor::AESE(Vec Vn, Vec Vd) {
+    const IR::U128 operand1 = ir.GetQ(Vd);
+    const IR::U128 operand2 = ir.GetQ(Vn);
+
+    const IR::U128 result = ir.AESEncrypt(ir.VectorEor(operand1, operand2));
+
+    ir.SetQ(Vd, result);
+    return true;
+}
+
 bool TranslatorVisitor::AESIMC(Vec Vn, Vec Vd) {
     const IR::U128 operand = ir.GetQ(Vn);
     const IR::U128 result = ir.AESInverseMixColumns(operand);

src/frontend/ir/ir_emitter.cpp

@@ -723,6 +723,10 @@ U32 IREmitter::CRC32ISO64(const U32& a, const U64& b) {
     return Inst<U32>(Opcode::CRC32ISO64, a, b);
 }
 
+U128 IREmitter::AESEncrypt(const U128& a) {
+    return Inst<U128>(Opcode::AESEncrypt, a);
+}
+
 U128 IREmitter::AESInverseMixColumns(const U128& a) {
     return Inst<U128>(Opcode::AESInverseMixColumns, a);
 }

src/frontend/ir/ir_emitter.h

@@ -197,6 +197,7 @@ public:
     U32 CRC32ISO32(const U32& a, const U32& b);
     U32 CRC32ISO64(const U32& a, const U64& b);
 
+    U128 AESEncrypt(const U128& a);
     U128 AESInverseMixColumns(const U128& a);
     U128 AESMixColumns(const U128& a);
 

src/frontend/ir/opcodes.inc

@@ -181,6 +181,7 @@ OPCODE(CRC32ISO32,              T::U32,         T::U32,         T::U32
 OPCODE(CRC32ISO64,              T::U32,         T::U32,         T::U64                          )
 
 // AES instructions
+OPCODE(AESEncrypt,              T::U128,        T::U128                                         )
 OPCODE(AESInverseMixColumns,    T::U128,        T::U128                                         )
 OPCODE(AESMixColumns,           T::U128,        T::U128                                         )