A64: Implement CRC32C

src/frontend/A64/decoder/a64.inc

@@ -275,7 +275,7 @@ INST(LSRV,                   "LSRV",                                      "z0011
 INST(ASRV,                   "ASRV",                                      "z0011010110mmmmm001010nnnnnddddd")
 INST(RORV,                   "RORV",                                      "z0011010110mmmmm001011nnnnnddddd")
 //INST(CRC32,                  "CRC32B, CRC32H, CRC32W, CRC32X",            "z0011010110mmmmm0100zznnnnnddddd")
-//INST(CRC32C,                 "CRC32CB, CRC32CH, CRC32CW, CRC32CX",        "z0011010110mmmmm0101zznnnnnddddd")
+INST(CRC32C,                 "CRC32CB, CRC32CH, CRC32CW, CRC32CX",        "z0011010110mmmmm0101zznnnnnddddd")
 //INST(PACGA,                  "PACGA",                                     "10011010110mmmmm001100nnnnnddddd")
 
 // Data Processing - Register - 1 source

src/frontend/A64/translate/impl/data_processing_crc32.cpp

@@ -0,0 +1,44 @@
+/* This file is part of the dynarmic project.
+ * Copyright (c) 2018 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 "frontend/A64/translate/impl/impl.h"
+
+namespace Dynarmic::A64 {
+
+bool TranslatorVisitor::CRC32C(bool sf, Reg Rm, Imm<2> sz, Reg Rn, Reg Rd) {
+    const u32 integral_size = sz.ZeroExtend();
+
+    if (sf && integral_size != 0b11) {
+        return UnallocatedEncoding();
+    }
+
+    if (!sf && integral_size == 0b11) {
+        return UnallocatedEncoding();
+    }
+
+    const IR::U32 result = [&] {
+        const size_t datasize = sf ? 64 : 32;
+        const IR::U32 accumulator = ir.GetW(Rn);
+        const IR::U32U64 data = X(datasize, Rm);
+
+        switch (integral_size) {
+        case 0b00:
+            return ir.CRC32Castagnoli8(accumulator, data);
+        case 0b01:
+            return ir.CRC32Castagnoli16(accumulator, data);
+        case 0b10:
+            return ir.CRC32Castagnoli32(accumulator, data);
+        case 0b11:
+        default:
+            return ir.CRC32Castagnoli64(accumulator, data);
+        }
+    }();
+
+    X(32, Rd, result);
+    return true;
+}
+
+} // namespace Dynarmic::A64

src/frontend/ir/ir_emitter.cpp

@@ -683,6 +683,22 @@ U32 IREmitter::PackedSelect(const U32& ge, const U32& a, const U32& b) {
     return Inst<U32>(Opcode::PackedSelect, ge, a, b);
 }
 
+U32 IREmitter::CRC32Castagnoli8(const U32& a, const U32& b) {
+    return Inst<U32>(Opcode::CRC32Castagnoli8, a, b);
+}
+
+U32 IREmitter::CRC32Castagnoli16(const U32& a, const U32& b) {
+    return Inst<U32>(Opcode::CRC32Castagnoli16, a, b);
+}
+
+U32 IREmitter::CRC32Castagnoli32(const U32& a, const U32& b) {
+    return Inst<U32>(Opcode::CRC32Castagnoli32, a, b);
+}
+
+U32 IREmitter::CRC32Castagnoli64(const U32& a, const U64& b) {
+    return Inst<U32>(Opcode::CRC32Castagnoli64, a, b);
+}
+
 UAny IREmitter::VectorGetElement(size_t esize, const U128& a, size_t index) {
     ASSERT_MSG(esize * index < 128, "Invalid index");
     switch (esize) {

src/frontend/ir/ir_emitter.h

@@ -186,6 +186,11 @@ public:
     U32 PackedAbsDiffSumS8(const U32& a, const U32& b);
     U32 PackedSelect(const U32& ge, const U32& a, const U32& b);
 
+    U32 CRC32Castagnoli8(const U32& a, const U32& b);
+    U32 CRC32Castagnoli16(const U32& a, const U32& b);
+    U32 CRC32Castagnoli32(const U32& a, const U32& b);
+    U32 CRC32Castagnoli64(const U32& a, const U64& b);
+
     UAny VectorGetElement(size_t esize, const U128& a, size_t index);
     U128 VectorAdd8(const U128& a, const U128& b);
     U128 VectorAdd16(const U128& a, const U128& b);

src/frontend/ir/opcodes.inc

@@ -167,6 +167,12 @@ OPCODE(PackedSaturatedSubS16,   T::U32,         T::U32,         T::U32
 OPCODE(PackedAbsDiffSumS8,      T::U32,         T::U32,         T::U32                          )
 OPCODE(PackedSelect,            T::U32,         T::U32,         T::U32,         T::U32          )
 
+// CRC instructions
+OPCODE(CRC32Castagnoli8,        T::U32,         T::U32,         T::U32                          )
+OPCODE(CRC32Castagnoli16,       T::U32,         T::U32,         T::U32                          )
+OPCODE(CRC32Castagnoli32,       T::U32,         T::U32,         T::U32                          )
+OPCODE(CRC32Castagnoli64,       T::U32,         T::U32,         T::U64                          )
+
 // Vector instructions
 OPCODE(VectorGetElement8,       T::U8,          T::U128,        T::U8                           )
 OPCODE(VectorGetElement16,      T::U16,         T::U128,        T::U8                           )