1// Copyright 2009 The Go Authors. All rights reserved.
2// Use of this source code is governed by a BSD-style
3// license that can be found in the LICENSE file.
4
5// This Go implementation is derived in part from the reference
6// ANSI C implementation, which carries the following notice:
7//
8//	rijndael-alg-fst.c
9//
10//	@version 3.0 (December 2000)
11//
12//	Optimised ANSI C code for the Rijndael cipher (now AES)
13//
14//	@author Vincent Rijmen <[email protected]>
15//	@author Antoon Bosselaers <[email protected]>
16//	@author Paulo Barreto <[email protected]>
17//
18//	This code is hereby placed in the public domain.
19//
20//	THIS SOFTWARE IS PROVIDED BY THE AUTHORS ''AS IS'' AND ANY EXPRESS
21//	OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
22//	WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23//	ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE
24//	LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
25//	CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
26//	SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
27//	BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
28//	WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
29//	OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
30//	EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31//
32// See FIPS 197 for specification, and see Daemen and Rijmen's Rijndael submission
33// for implementation details.
34//	https://csrc.nist.gov/csrc/media/publications/fips/197/final/documents/fips-197.pdf
35//	https://csrc.nist.gov/archive/aes/rijndael/Rijndael-ammended.pdf
36
37package aes
38
39import "internal/byteorder"
40
41// Encrypt one block from src into dst, using the expanded key xk.
42func encryptBlockGo(xk []uint32, dst, src []byte) {
43	_ = src[15] // early bounds check
44	s0 := byteorder.BeUint32(src[0:4])
45	s1 := byteorder.BeUint32(src[4:8])
46	s2 := byteorder.BeUint32(src[8:12])
47	s3 := byteorder.BeUint32(src[12:16])
48
49	// First round just XORs input with key.
50	s0 ^= xk[0]
51	s1 ^= xk[1]
52	s2 ^= xk[2]
53	s3 ^= xk[3]
54
55	// Middle rounds shuffle using tables.
56	// Number of rounds is set by length of expanded key.
57	nr := len(xk)/4 - 2 // - 2: one above, one more below
58	k := 4
59	var t0, t1, t2, t3 uint32
60	for r := 0; r < nr; r++ {
61		t0 = xk[k+0] ^ te0[uint8(s0>>24)] ^ te1[uint8(s1>>16)] ^ te2[uint8(s2>>8)] ^ te3[uint8(s3)]
62		t1 = xk[k+1] ^ te0[uint8(s1>>24)] ^ te1[uint8(s2>>16)] ^ te2[uint8(s3>>8)] ^ te3[uint8(s0)]
63		t2 = xk[k+2] ^ te0[uint8(s2>>24)] ^ te1[uint8(s3>>16)] ^ te2[uint8(s0>>8)] ^ te3[uint8(s1)]
64		t3 = xk[k+3] ^ te0[uint8(s3>>24)] ^ te1[uint8(s0>>16)] ^ te2[uint8(s1>>8)] ^ te3[uint8(s2)]
65		k += 4
66		s0, s1, s2, s3 = t0, t1, t2, t3
67	}
68
69	// Last round uses s-box directly and XORs to produce output.
70	s0 = uint32(sbox0[t0>>24])<<24 | uint32(sbox0[t1>>16&0xff])<<16 | uint32(sbox0[t2>>8&0xff])<<8 | uint32(sbox0[t3&0xff])
71	s1 = uint32(sbox0[t1>>24])<<24 | uint32(sbox0[t2>>16&0xff])<<16 | uint32(sbox0[t3>>8&0xff])<<8 | uint32(sbox0[t0&0xff])
72	s2 = uint32(sbox0[t2>>24])<<24 | uint32(sbox0[t3>>16&0xff])<<16 | uint32(sbox0[t0>>8&0xff])<<8 | uint32(sbox0[t1&0xff])
73	s3 = uint32(sbox0[t3>>24])<<24 | uint32(sbox0[t0>>16&0xff])<<16 | uint32(sbox0[t1>>8&0xff])<<8 | uint32(sbox0[t2&0xff])
74
75	s0 ^= xk[k+0]
76	s1 ^= xk[k+1]
77	s2 ^= xk[k+2]
78	s3 ^= xk[k+3]
79
80	_ = dst[15] // early bounds check
81	byteorder.BePutUint32(dst[0:4], s0)
82	byteorder.BePutUint32(dst[4:8], s1)
83	byteorder.BePutUint32(dst[8:12], s2)
84	byteorder.BePutUint32(dst[12:16], s3)
85}
86
87// Decrypt one block from src into dst, using the expanded key xk.
88func decryptBlockGo(xk []uint32, dst, src []byte) {
89	_ = src[15] // early bounds check
90	s0 := byteorder.BeUint32(src[0:4])
91	s1 := byteorder.BeUint32(src[4:8])
92	s2 := byteorder.BeUint32(src[8:12])
93	s3 := byteorder.BeUint32(src[12:16])
94
95	// First round just XORs input with key.
96	s0 ^= xk[0]
97	s1 ^= xk[1]
98	s2 ^= xk[2]
99	s3 ^= xk[3]
100
101	// Middle rounds shuffle using tables.
102	// Number of rounds is set by length of expanded key.
103	nr := len(xk)/4 - 2 // - 2: one above, one more below
104	k := 4
105	var t0, t1, t2, t3 uint32
106	for r := 0; r < nr; r++ {
107		t0 = xk[k+0] ^ td0[uint8(s0>>24)] ^ td1[uint8(s3>>16)] ^ td2[uint8(s2>>8)] ^ td3[uint8(s1)]
108		t1 = xk[k+1] ^ td0[uint8(s1>>24)] ^ td1[uint8(s0>>16)] ^ td2[uint8(s3>>8)] ^ td3[uint8(s2)]
109		t2 = xk[k+2] ^ td0[uint8(s2>>24)] ^ td1[uint8(s1>>16)] ^ td2[uint8(s0>>8)] ^ td3[uint8(s3)]
110		t3 = xk[k+3] ^ td0[uint8(s3>>24)] ^ td1[uint8(s2>>16)] ^ td2[uint8(s1>>8)] ^ td3[uint8(s0)]
111		k += 4
112		s0, s1, s2, s3 = t0, t1, t2, t3
113	}
114
115	// Last round uses s-box directly and XORs to produce output.
116	s0 = uint32(sbox1[t0>>24])<<24 | uint32(sbox1[t3>>16&0xff])<<16 | uint32(sbox1[t2>>8&0xff])<<8 | uint32(sbox1[t1&0xff])
117	s1 = uint32(sbox1[t1>>24])<<24 | uint32(sbox1[t0>>16&0xff])<<16 | uint32(sbox1[t3>>8&0xff])<<8 | uint32(sbox1[t2&0xff])
118	s2 = uint32(sbox1[t2>>24])<<24 | uint32(sbox1[t1>>16&0xff])<<16 | uint32(sbox1[t0>>8&0xff])<<8 | uint32(sbox1[t3&0xff])
119	s3 = uint32(sbox1[t3>>24])<<24 | uint32(sbox1[t2>>16&0xff])<<16 | uint32(sbox1[t1>>8&0xff])<<8 | uint32(sbox1[t0&0xff])
120
121	s0 ^= xk[k+0]
122	s1 ^= xk[k+1]
123	s2 ^= xk[k+2]
124	s3 ^= xk[k+3]
125
126	_ = dst[15] // early bounds check
127	byteorder.BePutUint32(dst[0:4], s0)
128	byteorder.BePutUint32(dst[4:8], s1)
129	byteorder.BePutUint32(dst[8:12], s2)
130	byteorder.BePutUint32(dst[12:16], s3)
131}
132
133// Apply sbox0 to each byte in w.
134func subw(w uint32) uint32 {
135	return uint32(sbox0[w>>24])<<24 |
136		uint32(sbox0[w>>16&0xff])<<16 |
137		uint32(sbox0[w>>8&0xff])<<8 |
138		uint32(sbox0[w&0xff])
139}
140
141// Rotate
142func rotw(w uint32) uint32 { return w<<8 | w>>24 }
143
144// Key expansion algorithm. See FIPS-197, Figure 11.
145// Their rcon[i] is our powx[i-1] << 24.
146func expandKeyGo(key []byte, enc, dec []uint32) {
147	// Encryption key setup.
148	var i int
149	nk := len(key) / 4
150	for i = 0; i < nk; i++ {
151		enc[i] = byteorder.BeUint32(key[4*i:])
152	}
153	for ; i < len(enc); i++ {
154		t := enc[i-1]
155		if i%nk == 0 {
156			t = subw(rotw(t)) ^ (uint32(powx[i/nk-1]) << 24)
157		} else if nk > 6 && i%nk == 4 {
158			t = subw(t)
159		}
160		enc[i] = enc[i-nk] ^ t
161	}
162
163	// Derive decryption key from encryption key.
164	// Reverse the 4-word round key sets from enc to produce dec.
165	// All sets but the first and last get the MixColumn transform applied.
166	if dec == nil {
167		return
168	}
169	n := len(enc)
170	for i := 0; i < n; i += 4 {
171		ei := n - i - 4
172		for j := 0; j < 4; j++ {
173			x := enc[ei+j]
174			if i > 0 && i+4 < n {
175				x = td0[sbox0[x>>24]] ^ td1[sbox0[x>>16&0xff]] ^ td2[sbox0[x>>8&0xff]] ^ td3[sbox0[x&0xff]]
176			}
177			dec[i+j] = x
178		}
179	}
180}
181