From af1384d7002ab9e4020fdb6fec6e28cac95a4524 Mon Sep 17 00:00:00 2001 From: Lioncash Date: Sun, 28 Jan 2018 01:41:58 -0500 Subject: [PATCH] A64: Implement CRC32 --- src/backend_x64/emit_x64_crc32.cpp | 109 +++++++++++++++++- src/frontend/A64/decoder/a64.inc | 2 +- .../translate/impl/data_processing_crc32.cpp | 33 ++++++ src/frontend/ir/ir_emitter.cpp | 16 +++ src/frontend/ir/ir_emitter.h | 4 + src/frontend/ir/opcodes.inc | 4 + 6 files changed, 161 insertions(+), 7 deletions(-) diff --git a/src/backend_x64/emit_x64_crc32.cpp b/src/backend_x64/emit_x64_crc32.cpp index 3c44b56f..1a35a34d 100644 --- a/src/backend_x64/emit_x64_crc32.cpp +++ b/src/backend_x64/emit_x64_crc32.cpp @@ -17,8 +17,10 @@ namespace Dynarmic::BackendX64 { using namespace Xbyak::util; +using CRC32Table = std::array; + // CRC32 algorithm that uses polynomial 0x1EDC6F41 -constexpr std::array castagnoli_table{{ +constexpr CRC32Table castagnoli_table{{ 0x00000000, 0xF26B8303, 0xE13B70F7, 0x1350F3F4, 0xC79A971F, 0x35F1141C, 0x26A1E7E8, 0xD4CA64EB, 0x8AD958CF, 0x78B2DBCC, 0x6BE22838, 0x9989AB3B, @@ -85,7 +87,75 @@ constexpr std::array castagnoli_table{{ 0xBE2DA0A5, 0x4C4623A6, 0x5F16D052, 0xAD7D5351 }}; -static u32 ComputeCRC32(const std::array& table, u32 crc, const u64 value, int length) { +// CRC32 algorithm that uses polynomial 0x04C11DB7 +constexpr CRC32Table iso_table{{ + 0x00000000, 0x77073096, 0xEE0E612C, 0x990951BA, + 0x076DC419, 0x706AF48F, 0xE963A535, 0x9E6495A3, + 0x0EDB8832, 0x79DCB8A4, 0xE0D5E91E, 0x97D2D988, + 0x09B64C2B, 0x7EB17CBD, 0xE7B82D07, 0x90BF1D91, + 0x1DB71064, 0x6AB020F2, 0xF3B97148, 0x84BE41DE, + 0x1ADAD47D, 0x6DDDE4EB, 0xF4D4B551, 0x83D385C7, + 0x136C9856, 0x646BA8C0, 0xFD62F97A, 0x8A65C9EC, + 0x14015C4F, 0x63066CD9, 0xFA0F3D63, 0x8D080DF5, + 0x3B6E20C8, 0x4C69105E, 0xD56041E4, 0xA2677172, + 0x3C03E4D1, 0x4B04D447, 0xD20D85FD, 0xA50AB56B, + 0x35B5A8FA, 0x42B2986C, 0xDBBBC9D6, 0xACBCF940, + 0x32D86CE3, 0x45DF5C75, 0xDCD60DCF, 0xABD13D59, + 0x26D930AC, 0x51DE003A, 0xC8D75180, 0xBFD06116, + 0x21B4F4B5, 0x56B3C423, 0xCFBA9599, 0xB8BDA50F, + 0x2802B89E, 0x5F058808, 0xC60CD9B2, 0xB10BE924, + 0x2F6F7C87, 0x58684C11, 0xC1611DAB, 0xB6662D3D, + 0x76DC4190, 0x01DB7106, 0x98D220BC, 0xEFD5102A, + 0x71B18589, 0x06B6B51F, 0x9FBFE4A5, 0xE8B8D433, + 0x7807C9A2, 0x0F00F934, 0x9609A88E, 0xE10E9818, + 0x7F6A0DBB, 0x086D3D2D, 0x91646C97, 0xE6635C01, + 0x6B6B51F4, 0x1C6C6162, 0x856530D8, 0xF262004E, + 0x6C0695ED, 0x1B01A57B, 0x8208F4C1, 0xF50FC457, + 0x65B0D9C6, 0x12B7E950, 0x8BBEB8EA, 0xFCB9887C, + 0x62DD1DDF, 0x15DA2D49, 0x8CD37CF3, 0xFBD44C65, + 0x4DB26158, 0x3AB551CE, 0xA3BC0074, 0xD4BB30E2, + 0x4ADFA541, 0x3DD895D7, 0xA4D1C46D, 0xD3D6F4FB, + 0x4369E96A, 0x346ED9FC, 0xAD678846, 0xDA60B8D0, + 0x44042D73, 0x33031DE5, 0xAA0A4C5F, 0xDD0D7CC9, + 0x5005713C, 0x270241AA, 0xBE0B1010, 0xC90C2086, + 0x5768B525, 0x206F85B3, 0xB966D409, 0xCE61E49F, + 0x5EDEF90E, 0x29D9C998, 0xB0D09822, 0xC7D7A8B4, + 0x59B33D17, 0x2EB40D81, 0xB7BD5C3B, 0xC0BA6CAD, + 0xEDB88320, 0x9ABFB3B6, 0x03B6E20C, 0x74B1D29A, + 0xEAD54739, 0x9DD277AF, 0x04DB2615, 0x73DC1683, + 0xE3630B12, 0x94643B84, 0x0D6D6A3E, 0x7A6A5AA8, + 0xE40ECF0B, 0x9309FF9D, 0x0A00AE27, 0x7D079EB1, + 0xF00F9344, 0x8708A3D2, 0x1E01F268, 0x6906C2FE, + 0xF762575D, 0x806567CB, 0x196C3671, 0x6E6B06E7, + 0xFED41B76, 0x89D32BE0, 0x10DA7A5A, 0x67DD4ACC, + 0xF9B9DF6F, 0x8EBEEFF9, 0x17B7BE43, 0x60B08ED5, + 0xD6D6A3E8, 0xA1D1937E, 0x38D8C2C4, 0x4FDFF252, + 0xD1BB67F1, 0xA6BC5767, 0x3FB506DD, 0x48B2364B, + 0xD80D2BDA, 0xAF0A1B4C, 0x36034AF6, 0x41047A60, + 0xDF60EFC3, 0xA867DF55, 0x316E8EEF, 0x4669BE79, + 0xCB61B38C, 0xBC66831A, 0x256FD2A0, 0x5268E236, + 0xCC0C7795, 0xBB0B4703, 0x220216B9, 0x5505262F, + 0xC5BA3BBE, 0xB2BD0B28, 0x2BB45A92, 0x5CB36A04, + 0xC2D7FFA7, 0xB5D0CF31, 0x2CD99E8B, 0x5BDEAE1D, + 0x9B64C2B0, 0xEC63F226, 0x756AA39C, 0x026D930A, + 0x9C0906A9, 0xEB0E363F, 0x72076785, 0x05005713, + 0x95BF4A82, 0xE2B87A14, 0x7BB12BAE, 0x0CB61B38, + 0x92D28E9B, 0xE5D5BE0D, 0x7CDCEFB7, 0x0BDBDF21, + 0x86D3D2D4, 0xF1D4E242, 0x68DDB3F8, 0x1FDA836E, + 0x81BE16CD, 0xF6B9265B, 0x6FB077E1, 0x18B74777, + 0x88085AE6, 0xFF0F6A70, 0x66063BCA, 0x11010B5C, + 0x8F659EFF, 0xF862AE69, 0x616BFFD3, 0x166CCF45, + 0xA00AE278, 0xD70DD2EE, 0x4E048354, 0x3903B3C2, + 0xA7672661, 0xD06016F7, 0x4969474D, 0x3E6E77DB, + 0xAED16A4A, 0xD9D65ADC, 0x40DF0B66, 0x37D83BF0, + 0xA9BCAE53, 0xDEBB9EC5, 0x47B2CF7F, 0x30B5FFE9, + 0xBDBDF21C, 0xCABAC28A, 0x53B39330, 0x24B4A3A6, + 0xBAD03605, 0xCDD70693, 0x54DE5729, 0x23D967BF, + 0xB3667A2E, 0xC4614AB8, 0x5D681B02, 0x2A6F2B94, + 0xB40BBE37, 0xC30C8EA1, 0x5A05DF1B, 0x2D02EF8D +}}; + +static u32 ComputeCRC32(const CRC32Table& table, u32 crc, const u64 value, int length) { const auto* data = reinterpret_cast(&value); while (length-- != 0) { @@ -95,6 +165,14 @@ static u32 ComputeCRC32(const std::array& table, u32 crc, const u64 va return crc; } +static void CRC32Fallback(std::array args, BlockOfCode& code, EmitContext& ctx, + IR::Inst* inst, const CRC32Table& table, const int data_size) { + ctx.reg_alloc.HostCall(inst, {}, args[0], args[1], {}); + code.mov(code.ABI_PARAM1, reinterpret_cast(&table)); + code.mov(code.ABI_PARAM4, data_size / CHAR_BIT); + code.CallFunction(&ComputeCRC32); +} + static void EmitCRC32Castagnoli(BlockOfCode& code, EmitContext& ctx, IR::Inst* inst, const int data_size) { auto args = ctx.reg_alloc.GetArgumentInfo(inst); @@ -104,13 +182,16 @@ static void EmitCRC32Castagnoli(BlockOfCode& code, EmitContext& ctx, IR::Inst* i code.crc32(crc, value); ctx.reg_alloc.DefineValue(inst, crc); } else { - ctx.reg_alloc.HostCall(inst, {}, args[0], args[1], {}); - code.mov(code.ABI_PARAM1, reinterpret_cast(&castagnoli_table)); - code.mov(code.ABI_PARAM4, data_size / CHAR_BIT); - code.CallFunction(&ComputeCRC32); + CRC32Fallback(args, code, ctx, inst, castagnoli_table, data_size); } } +static void EmitCRC32ISO(BlockOfCode& code, EmitContext& ctx, IR::Inst* inst, const int data_size) { + auto args = ctx.reg_alloc.GetArgumentInfo(inst); + + CRC32Fallback(args, code, ctx, inst, iso_table, data_size); +} + void EmitX64::EmitCRC32Castagnoli8(EmitContext& ctx, IR::Inst* inst) { EmitCRC32Castagnoli(*code, ctx, inst, 8); } @@ -127,4 +208,20 @@ void EmitX64::EmitCRC32Castagnoli64(EmitContext& ctx, IR::Inst* inst) { EmitCRC32Castagnoli(*code, ctx, inst, 64); } +void EmitX64::EmitCRC32ISO8(EmitContext& ctx, IR::Inst* inst) { + EmitCRC32ISO(*code, ctx, inst, 8); +} + +void EmitX64::EmitCRC32ISO16(EmitContext& ctx, IR::Inst* inst) { + EmitCRC32ISO(*code, ctx, inst, 16); +} + +void EmitX64::EmitCRC32ISO32(EmitContext& ctx, IR::Inst* inst) { + EmitCRC32ISO(*code, ctx, inst, 32); +} + +void EmitX64::EmitCRC32ISO64(EmitContext& ctx, IR::Inst* inst) { + EmitCRC32ISO(*code, ctx, inst, 64); +} + } // namespace Dynarmic::BackendX64 diff --git a/src/frontend/A64/decoder/a64.inc b/src/frontend/A64/decoder/a64.inc index 2820413d..0190eee7 100644 --- a/src/frontend/A64/decoder/a64.inc +++ b/src/frontend/A64/decoder/a64.inc @@ -274,7 +274,7 @@ INST(LSLV, "LSLV", "z0011 INST(LSRV, "LSRV", "z0011010110mmmmm001001nnnnnddddd") INST(ASRV, "ASRV", "z0011010110mmmmm001010nnnnnddddd") INST(RORV, "RORV", "z0011010110mmmmm001011nnnnnddddd") -//INST(CRC32, "CRC32B, CRC32H, CRC32W, CRC32X", "z0011010110mmmmm0100zznnnnnddddd") +INST(CRC32, "CRC32B, CRC32H, CRC32W, CRC32X", "z0011010110mmmmm0100zznnnnnddddd") INST(CRC32C, "CRC32CB, CRC32CH, CRC32CW, CRC32CX", "z0011010110mmmmm0101zznnnnnddddd") //INST(PACGA, "PACGA", "10011010110mmmmm001100nnnnnddddd") diff --git a/src/frontend/A64/translate/impl/data_processing_crc32.cpp b/src/frontend/A64/translate/impl/data_processing_crc32.cpp index 66430a8b..6378ad8f 100644 --- a/src/frontend/A64/translate/impl/data_processing_crc32.cpp +++ b/src/frontend/A64/translate/impl/data_processing_crc32.cpp @@ -8,6 +8,39 @@ namespace Dynarmic::A64 { +bool TranslatorVisitor::CRC32(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.CRC32ISO8(accumulator, data); + case 0b01: + return ir.CRC32ISO16(accumulator, data); + case 0b10: + return ir.CRC32ISO32(accumulator, data); + case 0b11: + default: + return ir.CRC32ISO64(accumulator, data); + } + }(); + + X(32, Rd, result); + return true; +} + bool TranslatorVisitor::CRC32C(bool sf, Reg Rm, Imm<2> sz, Reg Rn, Reg Rd) { const u32 integral_size = sz.ZeroExtend(); diff --git a/src/frontend/ir/ir_emitter.cpp b/src/frontend/ir/ir_emitter.cpp index efd59076..44e41efe 100644 --- a/src/frontend/ir/ir_emitter.cpp +++ b/src/frontend/ir/ir_emitter.cpp @@ -699,6 +699,22 @@ U32 IREmitter::CRC32Castagnoli64(const U32& a, const U64& b) { return Inst(Opcode::CRC32Castagnoli64, a, b); } +U32 IREmitter::CRC32ISO8(const U32& a, const U32& b) { + return Inst(Opcode::CRC32ISO8, a, b); +} + +U32 IREmitter::CRC32ISO16(const U32& a, const U32& b) { + return Inst(Opcode::CRC32ISO16, a, b); +} + +U32 IREmitter::CRC32ISO32(const U32& a, const U32& b) { + return Inst(Opcode::CRC32ISO32, a, b); +} + +U32 IREmitter::CRC32ISO64(const U32& a, const U64& b) { + return Inst(Opcode::CRC32ISO64, a, b); +} + UAny IREmitter::VectorGetElement(size_t esize, const U128& a, size_t index) { ASSERT_MSG(esize * index < 128, "Invalid index"); switch (esize) { diff --git a/src/frontend/ir/ir_emitter.h b/src/frontend/ir/ir_emitter.h index b97f1d18..7551c897 100644 --- a/src/frontend/ir/ir_emitter.h +++ b/src/frontend/ir/ir_emitter.h @@ -190,6 +190,10 @@ public: U32 CRC32Castagnoli16(const U32& a, const U32& b); U32 CRC32Castagnoli32(const U32& a, const U32& b); U32 CRC32Castagnoli64(const U32& a, const U64& b); + U32 CRC32ISO8(const U32& a, const U32& b); + U32 CRC32ISO16(const U32& a, const U32& b); + U32 CRC32ISO32(const U32& a, const U32& b); + U32 CRC32ISO64(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); diff --git a/src/frontend/ir/opcodes.inc b/src/frontend/ir/opcodes.inc index 81b958f2..572fdf63 100644 --- a/src/frontend/ir/opcodes.inc +++ b/src/frontend/ir/opcodes.inc @@ -173,6 +173,10 @@ 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 ) +OPCODE(CRC32ISO8, T::U32, T::U32, T::U32 ) +OPCODE(CRC32ISO16, T::U32, T::U32, T::U32 ) +OPCODE(CRC32ISO32, T::U32, T::U32, T::U32 ) +OPCODE(CRC32ISO64, T::U32, T::U32, T::U64 ) // Vector instructions OPCODE(VectorGetElement8, T::U8, T::U128, T::U8 )