--- /dev/null
+// Copyright 2011 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Package twofish implements Bruce Schneier's Twofish encryption algorithm.
+package twofish // import "golang.org/x/crypto/twofish"
+
+// Twofish is defined in https://www.schneier.com/paper-twofish-paper.pdf [TWOFISH]
+
+// This code is a port of the LibTom C implementation.
+// See http://libtom.org/?page=features&newsitems=5&whatfile=crypt.
+// LibTomCrypt is free for all purposes under the public domain.
+// It was heavily inspired by the go blowfish package.
+
+import "strconv"
+
+// BlockSize is the constant block size of Twofish.
+const BlockSize = 16
+
+const mdsPolynomial = 0x169 // x^8 + x^6 + x^5 + x^3 + 1, see [TWOFISH] 4.2
+const rsPolynomial = 0x14d // x^8 + x^6 + x^3 + x^2 + 1, see [TWOFISH] 4.3
+
+// A Cipher is an instance of Twofish encryption using a particular key.
+type Cipher struct {
+ s [4][256]uint32
+ k [40]uint32
+}
+
+type KeySizeError int
+
+func (k KeySizeError) Error() string {
+ return "crypto/twofish: invalid key size " + strconv.Itoa(int(k))
+}
+
+// NewCipher creates and returns a Cipher.
+// The key argument should be the Twofish key, 16, 24 or 32 bytes.
+func NewCipher(key []byte) (*Cipher, error) {
+ keylen := len(key)
+
+ if keylen != 16 && keylen != 24 && keylen != 32 {
+ return nil, KeySizeError(keylen)
+ }
+
+ // k is the number of 64 bit words in key
+ k := keylen / 8
+
+ // Create the S[..] words
+ var S [4 * 4]byte
+ for i := 0; i < k; i++ {
+ // Computes [y0 y1 y2 y3] = rs . [x0 x1 x2 x3 x4 x5 x6 x7]
+ for j, rsRow := range rs {
+ for k, rsVal := range rsRow {
+ S[4*i+j] ^= gfMult(key[8*i+k], rsVal, rsPolynomial)
+ }
+ }
+ }
+
+ // Calculate subkeys
+ c := new(Cipher)
+ var tmp [4]byte
+ for i := byte(0); i < 20; i++ {
+ // A = h(p * 2x, Me)
+ for j := range tmp {
+ tmp[j] = 2 * i
+ }
+ A := h(tmp[:], key, 0)
+
+ // B = rolc(h(p * (2x + 1), Mo), 8)
+ for j := range tmp {
+ tmp[j] = 2*i + 1
+ }
+ B := h(tmp[:], key, 1)
+ B = rol(B, 8)
+
+ c.k[2*i] = A + B
+
+ // K[2i+1] = (A + 2B) <<< 9
+ c.k[2*i+1] = rol(2*B+A, 9)
+ }
+
+ // Calculate sboxes
+ switch k {
+ case 2:
+ for i := range c.s[0] {
+ c.s[0][i] = mdsColumnMult(sbox[1][sbox[0][sbox[0][byte(i)]^S[0]]^S[4]], 0)
+ c.s[1][i] = mdsColumnMult(sbox[0][sbox[0][sbox[1][byte(i)]^S[1]]^S[5]], 1)
+ c.s[2][i] = mdsColumnMult(sbox[1][sbox[1][sbox[0][byte(i)]^S[2]]^S[6]], 2)
+ c.s[3][i] = mdsColumnMult(sbox[0][sbox[1][sbox[1][byte(i)]^S[3]]^S[7]], 3)
+ }
+ case 3:
+ for i := range c.s[0] {
+ c.s[0][i] = mdsColumnMult(sbox[1][sbox[0][sbox[0][sbox[1][byte(i)]^S[0]]^S[4]]^S[8]], 0)
+ c.s[1][i] = mdsColumnMult(sbox[0][sbox[0][sbox[1][sbox[1][byte(i)]^S[1]]^S[5]]^S[9]], 1)
+ c.s[2][i] = mdsColumnMult(sbox[1][sbox[1][sbox[0][sbox[0][byte(i)]^S[2]]^S[6]]^S[10]], 2)
+ c.s[3][i] = mdsColumnMult(sbox[0][sbox[1][sbox[1][sbox[0][byte(i)]^S[3]]^S[7]]^S[11]], 3)
+ }
+ default:
+ for i := range c.s[0] {
+ c.s[0][i] = mdsColumnMult(sbox[1][sbox[0][sbox[0][sbox[1][sbox[1][byte(i)]^S[0]]^S[4]]^S[8]]^S[12]], 0)
+ c.s[1][i] = mdsColumnMult(sbox[0][sbox[0][sbox[1][sbox[1][sbox[0][byte(i)]^S[1]]^S[5]]^S[9]]^S[13]], 1)
+ c.s[2][i] = mdsColumnMult(sbox[1][sbox[1][sbox[0][sbox[0][sbox[0][byte(i)]^S[2]]^S[6]]^S[10]]^S[14]], 2)
+ c.s[3][i] = mdsColumnMult(sbox[0][sbox[1][sbox[1][sbox[0][sbox[1][byte(i)]^S[3]]^S[7]]^S[11]]^S[15]], 3)
+ }
+ }
+
+ return c, nil
+}
+
+// BlockSize returns the Twofish block size, 16 bytes.
+func (c *Cipher) BlockSize() int { return BlockSize }
+
+// store32l stores src in dst in little-endian form.
+func store32l(dst []byte, src uint32) {
+ dst[0] = byte(src)
+ dst[1] = byte(src >> 8)
+ dst[2] = byte(src >> 16)
+ dst[3] = byte(src >> 24)
+ return
+}
+
+// load32l reads a little-endian uint32 from src.
+func load32l(src []byte) uint32 {
+ return uint32(src[0]) | uint32(src[1])<<8 | uint32(src[2])<<16 | uint32(src[3])<<24
+}
+
+// rol returns x after a left circular rotation of y bits.
+func rol(x, y uint32) uint32 {
+ return (x << (y & 31)) | (x >> (32 - (y & 31)))
+}
+
+// ror returns x after a right circular rotation of y bits.
+func ror(x, y uint32) uint32 {
+ return (x >> (y & 31)) | (x << (32 - (y & 31)))
+}
+
+// The RS matrix. See [TWOFISH] 4.3
+var rs = [4][8]byte{
+ {0x01, 0xA4, 0x55, 0x87, 0x5A, 0x58, 0xDB, 0x9E},
+ {0xA4, 0x56, 0x82, 0xF3, 0x1E, 0xC6, 0x68, 0xE5},
+ {0x02, 0xA1, 0xFC, 0xC1, 0x47, 0xAE, 0x3D, 0x19},
+ {0xA4, 0x55, 0x87, 0x5A, 0x58, 0xDB, 0x9E, 0x03},
+}
+
+// sbox tables
+var sbox = [2][256]byte{
+ {
+ 0xa9, 0x67, 0xb3, 0xe8, 0x04, 0xfd, 0xa3, 0x76, 0x9a, 0x92, 0x80, 0x78, 0xe4, 0xdd, 0xd1, 0x38,
+ 0x0d, 0xc6, 0x35, 0x98, 0x18, 0xf7, 0xec, 0x6c, 0x43, 0x75, 0x37, 0x26, 0xfa, 0x13, 0x94, 0x48,
+ 0xf2, 0xd0, 0x8b, 0x30, 0x84, 0x54, 0xdf, 0x23, 0x19, 0x5b, 0x3d, 0x59, 0xf3, 0xae, 0xa2, 0x82,
+ 0x63, 0x01, 0x83, 0x2e, 0xd9, 0x51, 0x9b, 0x7c, 0xa6, 0xeb, 0xa5, 0xbe, 0x16, 0x0c, 0xe3, 0x61,
+ 0xc0, 0x8c, 0x3a, 0xf5, 0x73, 0x2c, 0x25, 0x0b, 0xbb, 0x4e, 0x89, 0x6b, 0x53, 0x6a, 0xb4, 0xf1,
+ 0xe1, 0xe6, 0xbd, 0x45, 0xe2, 0xf4, 0xb6, 0x66, 0xcc, 0x95, 0x03, 0x56, 0xd4, 0x1c, 0x1e, 0xd7,
+ 0xfb, 0xc3, 0x8e, 0xb5, 0xe9, 0xcf, 0xbf, 0xba, 0xea, 0x77, 0x39, 0xaf, 0x33, 0xc9, 0x62, 0x71,
+ 0x81, 0x79, 0x09, 0xad, 0x24, 0xcd, 0xf9, 0xd8, 0xe5, 0xc5, 0xb9, 0x4d, 0x44, 0x08, 0x86, 0xe7,
+ 0xa1, 0x1d, 0xaa, 0xed, 0x06, 0x70, 0xb2, 0xd2, 0x41, 0x7b, 0xa0, 0x11, 0x31, 0xc2, 0x27, 0x90,
+ 0x20, 0xf6, 0x60, 0xff, 0x96, 0x5c, 0xb1, 0xab, 0x9e, 0x9c, 0x52, 0x1b, 0x5f, 0x93, 0x0a, 0xef,
+ 0x91, 0x85, 0x49, 0xee, 0x2d, 0x4f, 0x8f, 0x3b, 0x47, 0x87, 0x6d, 0x46, 0xd6, 0x3e, 0x69, 0x64,
+ 0x2a, 0xce, 0xcb, 0x2f, 0xfc, 0x97, 0x05, 0x7a, 0xac, 0x7f, 0xd5, 0x1a, 0x4b, 0x0e, 0xa7, 0x5a,
+ 0x28, 0x14, 0x3f, 0x29, 0x88, 0x3c, 0x4c, 0x02, 0xb8, 0xda, 0xb0, 0x17, 0x55, 0x1f, 0x8a, 0x7d,
+ 0x57, 0xc7, 0x8d, 0x74, 0xb7, 0xc4, 0x9f, 0x72, 0x7e, 0x15, 0x22, 0x12, 0x58, 0x07, 0x99, 0x34,
+ 0x6e, 0x50, 0xde, 0x68, 0x65, 0xbc, 0xdb, 0xf8, 0xc8, 0xa8, 0x2b, 0x40, 0xdc, 0xfe, 0x32, 0xa4,
+ 0xca, 0x10, 0x21, 0xf0, 0xd3, 0x5d, 0x0f, 0x00, 0x6f, 0x9d, 0x36, 0x42, 0x4a, 0x5e, 0xc1, 0xe0,
+ },
+ {
+ 0x75, 0xf3, 0xc6, 0xf4, 0xdb, 0x7b, 0xfb, 0xc8, 0x4a, 0xd3, 0xe6, 0x6b, 0x45, 0x7d, 0xe8, 0x4b,
+ 0xd6, 0x32, 0xd8, 0xfd, 0x37, 0x71, 0xf1, 0xe1, 0x30, 0x0f, 0xf8, 0x1b, 0x87, 0xfa, 0x06, 0x3f,
+ 0x5e, 0xba, 0xae, 0x5b, 0x8a, 0x00, 0xbc, 0x9d, 0x6d, 0xc1, 0xb1, 0x0e, 0x80, 0x5d, 0xd2, 0xd5,
+ 0xa0, 0x84, 0x07, 0x14, 0xb5, 0x90, 0x2c, 0xa3, 0xb2, 0x73, 0x4c, 0x54, 0x92, 0x74, 0x36, 0x51,
+ 0x38, 0xb0, 0xbd, 0x5a, 0xfc, 0x60, 0x62, 0x96, 0x6c, 0x42, 0xf7, 0x10, 0x7c, 0x28, 0x27, 0x8c,
+ 0x13, 0x95, 0x9c, 0xc7, 0x24, 0x46, 0x3b, 0x70, 0xca, 0xe3, 0x85, 0xcb, 0x11, 0xd0, 0x93, 0xb8,
+ 0xa6, 0x83, 0x20, 0xff, 0x9f, 0x77, 0xc3, 0xcc, 0x03, 0x6f, 0x08, 0xbf, 0x40, 0xe7, 0x2b, 0xe2,
+ 0x79, 0x0c, 0xaa, 0x82, 0x41, 0x3a, 0xea, 0xb9, 0xe4, 0x9a, 0xa4, 0x97, 0x7e, 0xda, 0x7a, 0x17,
+ 0x66, 0x94, 0xa1, 0x1d, 0x3d, 0xf0, 0xde, 0xb3, 0x0b, 0x72, 0xa7, 0x1c, 0xef, 0xd1, 0x53, 0x3e,
+ 0x8f, 0x33, 0x26, 0x5f, 0xec, 0x76, 0x2a, 0x49, 0x81, 0x88, 0xee, 0x21, 0xc4, 0x1a, 0xeb, 0xd9,
+ 0xc5, 0x39, 0x99, 0xcd, 0xad, 0x31, 0x8b, 0x01, 0x18, 0x23, 0xdd, 0x1f, 0x4e, 0x2d, 0xf9, 0x48,
+ 0x4f, 0xf2, 0x65, 0x8e, 0x78, 0x5c, 0x58, 0x19, 0x8d, 0xe5, 0x98, 0x57, 0x67, 0x7f, 0x05, 0x64,
+ 0xaf, 0x63, 0xb6, 0xfe, 0xf5, 0xb7, 0x3c, 0xa5, 0xce, 0xe9, 0x68, 0x44, 0xe0, 0x4d, 0x43, 0x69,
+ 0x29, 0x2e, 0xac, 0x15, 0x59, 0xa8, 0x0a, 0x9e, 0x6e, 0x47, 0xdf, 0x34, 0x35, 0x6a, 0xcf, 0xdc,
+ 0x22, 0xc9, 0xc0, 0x9b, 0x89, 0xd4, 0xed, 0xab, 0x12, 0xa2, 0x0d, 0x52, 0xbb, 0x02, 0x2f, 0xa9,
+ 0xd7, 0x61, 0x1e, 0xb4, 0x50, 0x04, 0xf6, 0xc2, 0x16, 0x25, 0x86, 0x56, 0x55, 0x09, 0xbe, 0x91,
+ },
+}
+
+// gfMult returns a·b in GF(2^8)/p
+func gfMult(a, b byte, p uint32) byte {
+ B := [2]uint32{0, uint32(b)}
+ P := [2]uint32{0, p}
+ var result uint32
+
+ // branchless GF multiplier
+ for i := 0; i < 7; i++ {
+ result ^= B[a&1]
+ a >>= 1
+ B[1] = P[B[1]>>7] ^ (B[1] << 1)
+ }
+ result ^= B[a&1]
+ return byte(result)
+}
+
+// mdsColumnMult calculates y{col} where [y0 y1 y2 y3] = MDS · [x0]
+func mdsColumnMult(in byte, col int) uint32 {
+ mul01 := in
+ mul5B := gfMult(in, 0x5B, mdsPolynomial)
+ mulEF := gfMult(in, 0xEF, mdsPolynomial)
+
+ switch col {
+ case 0:
+ return uint32(mul01) | uint32(mul5B)<<8 | uint32(mulEF)<<16 | uint32(mulEF)<<24
+ case 1:
+ return uint32(mulEF) | uint32(mulEF)<<8 | uint32(mul5B)<<16 | uint32(mul01)<<24
+ case 2:
+ return uint32(mul5B) | uint32(mulEF)<<8 | uint32(mul01)<<16 | uint32(mulEF)<<24
+ case 3:
+ return uint32(mul5B) | uint32(mul01)<<8 | uint32(mulEF)<<16 | uint32(mul5B)<<24
+ }
+
+ panic("unreachable")
+}
+
+// h implements the S-box generation function. See [TWOFISH] 4.3.5
+func h(in, key []byte, offset int) uint32 {
+ var y [4]byte
+ for x := range y {
+ y[x] = in[x]
+ }
+ switch len(key) / 8 {
+ case 4:
+ y[0] = sbox[1][y[0]] ^ key[4*(6+offset)+0]
+ y[1] = sbox[0][y[1]] ^ key[4*(6+offset)+1]
+ y[2] = sbox[0][y[2]] ^ key[4*(6+offset)+2]
+ y[3] = sbox[1][y[3]] ^ key[4*(6+offset)+3]
+ fallthrough
+ case 3:
+ y[0] = sbox[1][y[0]] ^ key[4*(4+offset)+0]
+ y[1] = sbox[1][y[1]] ^ key[4*(4+offset)+1]
+ y[2] = sbox[0][y[2]] ^ key[4*(4+offset)+2]
+ y[3] = sbox[0][y[3]] ^ key[4*(4+offset)+3]
+ fallthrough
+ case 2:
+ y[0] = sbox[1][sbox[0][sbox[0][y[0]]^key[4*(2+offset)+0]]^key[4*(0+offset)+0]]
+ y[1] = sbox[0][sbox[0][sbox[1][y[1]]^key[4*(2+offset)+1]]^key[4*(0+offset)+1]]
+ y[2] = sbox[1][sbox[1][sbox[0][y[2]]^key[4*(2+offset)+2]]^key[4*(0+offset)+2]]
+ y[3] = sbox[0][sbox[1][sbox[1][y[3]]^key[4*(2+offset)+3]]^key[4*(0+offset)+3]]
+ }
+ // [y0 y1 y2 y3] = MDS . [x0 x1 x2 x3]
+ var mdsMult uint32
+ for i := range y {
+ mdsMult ^= mdsColumnMult(y[i], i)
+ }
+ return mdsMult
+}
+
+// Encrypt encrypts a 16-byte block from src to dst, which may overlap.
+// Note that for amounts of data larger than a block,
+// it is not safe to just call Encrypt on successive blocks;
+// instead, use an encryption mode like CBC (see crypto/cipher/cbc.go).
+func (c *Cipher) Encrypt(dst, src []byte) {
+ S1 := c.s[0]
+ S2 := c.s[1]
+ S3 := c.s[2]
+ S4 := c.s[3]
+
+ // Load input
+ ia := load32l(src[0:4])
+ ib := load32l(src[4:8])
+ ic := load32l(src[8:12])
+ id := load32l(src[12:16])
+
+ // Pre-whitening
+ ia ^= c.k[0]
+ ib ^= c.k[1]
+ ic ^= c.k[2]
+ id ^= c.k[3]
+
+ for i := 0; i < 8; i++ {
+ k := c.k[8+i*4 : 12+i*4]
+ t2 := S2[byte(ib)] ^ S3[byte(ib>>8)] ^ S4[byte(ib>>16)] ^ S1[byte(ib>>24)]
+ t1 := S1[byte(ia)] ^ S2[byte(ia>>8)] ^ S3[byte(ia>>16)] ^ S4[byte(ia>>24)] + t2
+ ic = ror(ic^(t1+k[0]), 1)
+ id = rol(id, 1) ^ (t2 + t1 + k[1])
+
+ t2 = S2[byte(id)] ^ S3[byte(id>>8)] ^ S4[byte(id>>16)] ^ S1[byte(id>>24)]
+ t1 = S1[byte(ic)] ^ S2[byte(ic>>8)] ^ S3[byte(ic>>16)] ^ S4[byte(ic>>24)] + t2
+ ia = ror(ia^(t1+k[2]), 1)
+ ib = rol(ib, 1) ^ (t2 + t1 + k[3])
+ }
+
+ // Output with "undo last swap"
+ ta := ic ^ c.k[4]
+ tb := id ^ c.k[5]
+ tc := ia ^ c.k[6]
+ td := ib ^ c.k[7]
+
+ store32l(dst[0:4], ta)
+ store32l(dst[4:8], tb)
+ store32l(dst[8:12], tc)
+ store32l(dst[12:16], td)
+}
+
+// Decrypt decrypts a 16-byte block from src to dst, which may overlap.
+func (c *Cipher) Decrypt(dst, src []byte) {
+ S1 := c.s[0]
+ S2 := c.s[1]
+ S3 := c.s[2]
+ S4 := c.s[3]
+
+ // Load input
+ ta := load32l(src[0:4])
+ tb := load32l(src[4:8])
+ tc := load32l(src[8:12])
+ td := load32l(src[12:16])
+
+ // Undo undo final swap
+ ia := tc ^ c.k[6]
+ ib := td ^ c.k[7]
+ ic := ta ^ c.k[4]
+ id := tb ^ c.k[5]
+
+ for i := 8; i > 0; i-- {
+ k := c.k[4+i*4 : 8+i*4]
+ t2 := S2[byte(id)] ^ S3[byte(id>>8)] ^ S4[byte(id>>16)] ^ S1[byte(id>>24)]
+ t1 := S1[byte(ic)] ^ S2[byte(ic>>8)] ^ S3[byte(ic>>16)] ^ S4[byte(ic>>24)] + t2
+ ia = rol(ia, 1) ^ (t1 + k[2])
+ ib = ror(ib^(t2+t1+k[3]), 1)
+
+ t2 = S2[byte(ib)] ^ S3[byte(ib>>8)] ^ S4[byte(ib>>16)] ^ S1[byte(ib>>24)]
+ t1 = S1[byte(ia)] ^ S2[byte(ia>>8)] ^ S3[byte(ia>>16)] ^ S4[byte(ia>>24)] + t2
+ ic = rol(ic, 1) ^ (t1 + k[0])
+ id = ror(id^(t2+t1+k[1]), 1)
+ }
+
+ // Undo pre-whitening
+ ia ^= c.k[0]
+ ib ^= c.k[1]
+ ic ^= c.k[2]
+ id ^= c.k[3]
+
+ store32l(dst[0:4], ia)
+ store32l(dst[4:8], ib)
+ store32l(dst[8:12], ic)
+ store32l(dst[12:16], id)
+}
OSDN Git Service
