xref: /titanic_41/usr/src/common/crypto/aes/aes_impl.c (revision bbb1277b6ec1b0daad4e3ed1a2b891d3e2ece2eb)
1 /*
2  * CDDL HEADER START
3  *
4  * The contents of this file are subject to the terms of the
5  * Common Development and Distribution License (the "License").
6  * You may not use this file except in compliance with the License.
7  *
8  * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9  * or http://www.opensolaris.org/os/licensing.
10  * See the License for the specific language governing permissions
11  * and limitations under the License.
12  *
13  * When distributing Covered Code, include this CDDL HEADER in each
14  * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15  * If applicable, add the following below this CDDL HEADER, with the
16  * fields enclosed by brackets "[]" replaced with your own identifying
17  * information: Portions Copyright [yyyy] [name of copyright owner]
18  *
19  * CDDL HEADER END
20  */
21 /*
22  * Copyright 2010 Sun Microsystems, Inc.  All rights reserved.
23  * Use is subject to license terms.
24  */
25 
26 #include <sys/types.h>
27 #include <sys/systm.h>
28 #include <sys/sysmacros.h>
29 #include <netinet/in.h>
30 #include "aes_impl.h"
31 #ifndef	_KERNEL
32 #include <strings.h>
33 #include <stdlib.h>
34 #endif	/* !_KERNEL */
35 
36 #ifdef __amd64
37 
38 #ifdef _KERNEL
39 #include <sys/cpuvar.h>		/* cpu_t, CPU */
40 #include <sys/x86_archext.h>	/* x86_feature, X86_AES */
41 #include <sys/disp.h>		/* kpreempt_disable(), kpreempt_enable */
42 
43 /* Workaround for no XMM kernel thread save/restore */
44 #define	KPREEMPT_DISABLE	kpreempt_disable()
45 #define	KPREEMPT_ENABLE		kpreempt_enable()
46 
47 #else
48 #include <sys/auxv.h>		/* getisax() */
49 #include <sys/auxv_386.h>	/* AV_386_AES bit */
50 #define	KPREEMPT_DISABLE
51 #define	KPREEMPT_ENABLE
52 #endif	/* _KERNEL */
53 #endif  /* __amd64 */
54 
55 
56 /*
57  * This file is derived from the file  rijndael-alg-fst.c  taken from the
58  * "optimized C code v3.0" on the "rijndael home page"
59  * http://www.iaik.tu-graz.ac.at/research/krypto/AES/old/~rijmen/rijndael/
60  * pointed by the NIST web-site http://csrc.nist.gov/archive/aes/
61  *
62  * The following note is from the original file:
63  */
64 
65 /*
66  * rijndael-alg-fst.c
67  *
68  * @version 3.0 (December 2000)
69  *
70  * Optimised ANSI C code for the Rijndael cipher (now AES)
71  *
72  * @author Vincent Rijmen <vincent.rijmen@esat.kuleuven.ac.be>
73  * @author Antoon Bosselaers <antoon.bosselaers@esat.kuleuven.ac.be>
74  * @author Paulo Barreto <paulo.barreto@terra.com.br>
75  *
76  * This code is hereby placed in the public domain.
77  *
78  * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ''AS IS'' AND ANY EXPRESS
79  * OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
80  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
81  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE
82  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
83  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
84  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
85  * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
86  * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
87  * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
88  * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
89  */
90 
91 /* EXPORT DELETE START */
92 
93 #if defined(sun4u)
94 /* External assembly functions: */
95 extern void aes_encrypt_impl(const uint32_t rk[], int Nr, const uint32_t pt[4],
96 	uint32_t ct[4]);
97 extern void aes_decrypt_impl(const uint32_t rk[], int Nr, const uint32_t ct[4],
98 	uint32_t pt[4]);
99 
100 #define	AES_ENCRYPT_IMPL(a, b, c, d, e)	aes_encrypt_impl(a, b, c, d)
101 #define	AES_DECRYPT_IMPL(a, b, c, d, e)	aes_decrypt_impl(a, b, c, d)
102 
103 #elif defined(__amd64)
104 
105 /* These functions are used to execute amd64 instructions for AMD or Intel: */
106 extern int rijndael_key_setup_enc_amd64(uint32_t rk[],
107 	const uint32_t cipherKey[], int keyBits);
108 extern int rijndael_key_setup_dec_amd64(uint32_t rk[],
109 	const uint32_t cipherKey[], int keyBits);
110 extern void aes_encrypt_amd64(const uint32_t rk[], int Nr,
111 	const uint32_t pt[4], uint32_t ct[4]);
112 extern void aes_decrypt_amd64(const uint32_t rk[], int Nr,
113 	const uint32_t ct[4], uint32_t pt[4]);
114 
115 /* These functions are used to execute Intel-specific AES-NI instructions: */
116 extern int rijndael_key_setup_enc_intel(uint32_t rk[],
117 	const uint32_t cipherKey[], uint64_t keyBits);
118 extern int rijndael_key_setup_dec_intel(uint32_t rk[],
119 	const uint32_t cipherKey[], uint64_t keyBits);
120 extern void aes_encrypt_intel(const uint32_t rk[], int Nr,
121 	const uint32_t pt[4], uint32_t ct[4]);
122 extern void aes_decrypt_intel(const uint32_t rk[], int Nr,
123 	const uint32_t ct[4], uint32_t pt[4]);
124 
125 static int intel_aes_instructions_present(void);
126 
127 #define	AES_ENCRYPT_IMPL(a, b, c, d, e) rijndael_encrypt(a, b, c, d, e)
128 #define	AES_DECRYPT_IMPL(a, b, c, d, e) rijndael_decrypt(a, b, c, d, e)
129 
130 #else /* Generic C implementation */
131 
132 #define	AES_ENCRYPT_IMPL(a, b, c, d, e)	rijndael_encrypt(a, b, c, d)
133 #define	AES_DECRYPT_IMPL(a, b, c, d, e)	rijndael_decrypt(a, b, c, d)
134 #define	rijndael_key_setup_enc_raw	rijndael_key_setup_enc
135 #endif	/* sun4u || __amd64 */
136 
137 #if defined(_LITTLE_ENDIAN) && !defined(__amd64)
138 #define	AES_BYTE_SWAP
139 #endif
140 
141 
142 #if !defined(__amd64)
143 /*
144  *  Constant tables
145  */
146 
147 /*
148  * Te0[x] = S [x].[02, 01, 01, 03];
149  * Te1[x] = S [x].[03, 02, 01, 01];
150  * Te2[x] = S [x].[01, 03, 02, 01];
151  * Te3[x] = S [x].[01, 01, 03, 02];
152  * Te4[x] = S [x].[01, 01, 01, 01];
153  *
154  * Td0[x] = Si[x].[0e, 09, 0d, 0b];
155  * Td1[x] = Si[x].[0b, 0e, 09, 0d];
156  * Td2[x] = Si[x].[0d, 0b, 0e, 09];
157  * Td3[x] = Si[x].[09, 0d, 0b, 0e];
158  * Td4[x] = Si[x].[01, 01, 01, 01];
159  */
160 
161 /* Encrypt Sbox constants (for the substitute bytes operation) */
162 
163 #ifndef sun4u
164 
165 static const uint32_t Te0[256] =
166 {
167 	0xc66363a5U, 0xf87c7c84U, 0xee777799U, 0xf67b7b8dU,
168 	0xfff2f20dU, 0xd66b6bbdU, 0xde6f6fb1U, 0x91c5c554U,
169 	0x60303050U, 0x02010103U, 0xce6767a9U, 0x562b2b7dU,
170 	0xe7fefe19U, 0xb5d7d762U, 0x4dababe6U, 0xec76769aU,
171 	0x8fcaca45U, 0x1f82829dU, 0x89c9c940U, 0xfa7d7d87U,
172 	0xeffafa15U, 0xb25959ebU, 0x8e4747c9U, 0xfbf0f00bU,
173 	0x41adadecU, 0xb3d4d467U, 0x5fa2a2fdU, 0x45afafeaU,
174 	0x239c9cbfU, 0x53a4a4f7U, 0xe4727296U, 0x9bc0c05bU,
175 	0x75b7b7c2U, 0xe1fdfd1cU, 0x3d9393aeU, 0x4c26266aU,
176 	0x6c36365aU, 0x7e3f3f41U, 0xf5f7f702U, 0x83cccc4fU,
177 	0x6834345cU, 0x51a5a5f4U, 0xd1e5e534U, 0xf9f1f108U,
178 	0xe2717193U, 0xabd8d873U, 0x62313153U, 0x2a15153fU,
179 	0x0804040cU, 0x95c7c752U, 0x46232365U, 0x9dc3c35eU,
180 	0x30181828U, 0x379696a1U, 0x0a05050fU, 0x2f9a9ab5U,
181 	0x0e070709U, 0x24121236U, 0x1b80809bU, 0xdfe2e23dU,
182 	0xcdebeb26U, 0x4e272769U, 0x7fb2b2cdU, 0xea75759fU,
183 	0x1209091bU, 0x1d83839eU, 0x582c2c74U, 0x341a1a2eU,
184 	0x361b1b2dU, 0xdc6e6eb2U, 0xb45a5aeeU, 0x5ba0a0fbU,
185 	0xa45252f6U, 0x763b3b4dU, 0xb7d6d661U, 0x7db3b3ceU,
186 	0x5229297bU, 0xdde3e33eU, 0x5e2f2f71U, 0x13848497U,
187 	0xa65353f5U, 0xb9d1d168U, 0x00000000U, 0xc1eded2cU,
188 	0x40202060U, 0xe3fcfc1fU, 0x79b1b1c8U, 0xb65b5bedU,
189 	0xd46a6abeU, 0x8dcbcb46U, 0x67bebed9U, 0x7239394bU,
190 	0x944a4adeU, 0x984c4cd4U, 0xb05858e8U, 0x85cfcf4aU,
191 	0xbbd0d06bU, 0xc5efef2aU, 0x4faaaae5U, 0xedfbfb16U,
192 	0x864343c5U, 0x9a4d4dd7U, 0x66333355U, 0x11858594U,
193 	0x8a4545cfU, 0xe9f9f910U, 0x04020206U, 0xfe7f7f81U,
194 	0xa05050f0U, 0x783c3c44U, 0x259f9fbaU, 0x4ba8a8e3U,
195 	0xa25151f3U, 0x5da3a3feU, 0x804040c0U, 0x058f8f8aU,
196 	0x3f9292adU, 0x219d9dbcU, 0x70383848U, 0xf1f5f504U,
197 	0x63bcbcdfU, 0x77b6b6c1U, 0xafdada75U, 0x42212163U,
198 	0x20101030U, 0xe5ffff1aU, 0xfdf3f30eU, 0xbfd2d26dU,
199 	0x81cdcd4cU, 0x180c0c14U, 0x26131335U, 0xc3ecec2fU,
200 	0xbe5f5fe1U, 0x359797a2U, 0x884444ccU, 0x2e171739U,
201 	0x93c4c457U, 0x55a7a7f2U, 0xfc7e7e82U, 0x7a3d3d47U,
202 	0xc86464acU, 0xba5d5de7U, 0x3219192bU, 0xe6737395U,
203 	0xc06060a0U, 0x19818198U, 0x9e4f4fd1U, 0xa3dcdc7fU,
204 	0x44222266U, 0x542a2a7eU, 0x3b9090abU, 0x0b888883U,
205 	0x8c4646caU, 0xc7eeee29U, 0x6bb8b8d3U, 0x2814143cU,
206 	0xa7dede79U, 0xbc5e5ee2U, 0x160b0b1dU, 0xaddbdb76U,
207 	0xdbe0e03bU, 0x64323256U, 0x743a3a4eU, 0x140a0a1eU,
208 	0x924949dbU, 0x0c06060aU, 0x4824246cU, 0xb85c5ce4U,
209 	0x9fc2c25dU, 0xbdd3d36eU, 0x43acacefU, 0xc46262a6U,
210 	0x399191a8U, 0x319595a4U, 0xd3e4e437U, 0xf279798bU,
211 	0xd5e7e732U, 0x8bc8c843U, 0x6e373759U, 0xda6d6db7U,
212 	0x018d8d8cU, 0xb1d5d564U, 0x9c4e4ed2U, 0x49a9a9e0U,
213 	0xd86c6cb4U, 0xac5656faU, 0xf3f4f407U, 0xcfeaea25U,
214 	0xca6565afU, 0xf47a7a8eU, 0x47aeaee9U, 0x10080818U,
215 	0x6fbabad5U, 0xf0787888U, 0x4a25256fU, 0x5c2e2e72U,
216 	0x381c1c24U, 0x57a6a6f1U, 0x73b4b4c7U, 0x97c6c651U,
217 	0xcbe8e823U, 0xa1dddd7cU, 0xe874749cU, 0x3e1f1f21U,
218 	0x964b4bddU, 0x61bdbddcU, 0x0d8b8b86U, 0x0f8a8a85U,
219 	0xe0707090U, 0x7c3e3e42U, 0x71b5b5c4U, 0xcc6666aaU,
220 	0x904848d8U, 0x06030305U, 0xf7f6f601U, 0x1c0e0e12U,
221 	0xc26161a3U, 0x6a35355fU, 0xae5757f9U, 0x69b9b9d0U,
222 	0x17868691U, 0x99c1c158U, 0x3a1d1d27U, 0x279e9eb9U,
223 	0xd9e1e138U, 0xebf8f813U, 0x2b9898b3U, 0x22111133U,
224 	0xd26969bbU, 0xa9d9d970U, 0x078e8e89U, 0x339494a7U,
225 	0x2d9b9bb6U, 0x3c1e1e22U, 0x15878792U, 0xc9e9e920U,
226 	0x87cece49U, 0xaa5555ffU, 0x50282878U, 0xa5dfdf7aU,
227 	0x038c8c8fU, 0x59a1a1f8U, 0x09898980U, 0x1a0d0d17U,
228 	0x65bfbfdaU, 0xd7e6e631U, 0x844242c6U, 0xd06868b8U,
229 	0x824141c3U, 0x299999b0U, 0x5a2d2d77U, 0x1e0f0f11U,
230 	0x7bb0b0cbU, 0xa85454fcU, 0x6dbbbbd6U, 0x2c16163aU
231 };
232 
233 
234 static const uint32_t Te1[256] =
235 {
236 	0xa5c66363U, 0x84f87c7cU, 0x99ee7777U, 0x8df67b7bU,
237 	0x0dfff2f2U, 0xbdd66b6bU, 0xb1de6f6fU, 0x5491c5c5U,
238 	0x50603030U, 0x03020101U, 0xa9ce6767U, 0x7d562b2bU,
239 	0x19e7fefeU, 0x62b5d7d7U, 0xe64dababU, 0x9aec7676U,
240 	0x458fcacaU, 0x9d1f8282U, 0x4089c9c9U, 0x87fa7d7dU,
241 	0x15effafaU, 0xebb25959U, 0xc98e4747U, 0x0bfbf0f0U,
242 	0xec41adadU, 0x67b3d4d4U, 0xfd5fa2a2U, 0xea45afafU,
243 	0xbf239c9cU, 0xf753a4a4U, 0x96e47272U, 0x5b9bc0c0U,
244 	0xc275b7b7U, 0x1ce1fdfdU, 0xae3d9393U, 0x6a4c2626U,
245 	0x5a6c3636U, 0x417e3f3fU, 0x02f5f7f7U, 0x4f83ccccU,
246 	0x5c683434U, 0xf451a5a5U, 0x34d1e5e5U, 0x08f9f1f1U,
247 	0x93e27171U, 0x73abd8d8U, 0x53623131U, 0x3f2a1515U,
248 	0x0c080404U, 0x5295c7c7U, 0x65462323U, 0x5e9dc3c3U,
249 	0x28301818U, 0xa1379696U, 0x0f0a0505U, 0xb52f9a9aU,
250 	0x090e0707U, 0x36241212U, 0x9b1b8080U, 0x3ddfe2e2U,
251 	0x26cdebebU, 0x694e2727U, 0xcd7fb2b2U, 0x9fea7575U,
252 	0x1b120909U, 0x9e1d8383U, 0x74582c2cU, 0x2e341a1aU,
253 	0x2d361b1bU, 0xb2dc6e6eU, 0xeeb45a5aU, 0xfb5ba0a0U,
254 	0xf6a45252U, 0x4d763b3bU, 0x61b7d6d6U, 0xce7db3b3U,
255 	0x7b522929U, 0x3edde3e3U, 0x715e2f2fU, 0x97138484U,
256 	0xf5a65353U, 0x68b9d1d1U, 0x00000000U, 0x2cc1ededU,
257 	0x60402020U, 0x1fe3fcfcU, 0xc879b1b1U, 0xedb65b5bU,
258 	0xbed46a6aU, 0x468dcbcbU, 0xd967bebeU, 0x4b723939U,
259 	0xde944a4aU, 0xd4984c4cU, 0xe8b05858U, 0x4a85cfcfU,
260 	0x6bbbd0d0U, 0x2ac5efefU, 0xe54faaaaU, 0x16edfbfbU,
261 	0xc5864343U, 0xd79a4d4dU, 0x55663333U, 0x94118585U,
262 	0xcf8a4545U, 0x10e9f9f9U, 0x06040202U, 0x81fe7f7fU,
263 	0xf0a05050U, 0x44783c3cU, 0xba259f9fU, 0xe34ba8a8U,
264 	0xf3a25151U, 0xfe5da3a3U, 0xc0804040U, 0x8a058f8fU,
265 	0xad3f9292U, 0xbc219d9dU, 0x48703838U, 0x04f1f5f5U,
266 	0xdf63bcbcU, 0xc177b6b6U, 0x75afdadaU, 0x63422121U,
267 	0x30201010U, 0x1ae5ffffU, 0x0efdf3f3U, 0x6dbfd2d2U,
268 	0x4c81cdcdU, 0x14180c0cU, 0x35261313U, 0x2fc3ececU,
269 	0xe1be5f5fU, 0xa2359797U, 0xcc884444U, 0x392e1717U,
270 	0x5793c4c4U, 0xf255a7a7U, 0x82fc7e7eU, 0x477a3d3dU,
271 	0xacc86464U, 0xe7ba5d5dU, 0x2b321919U, 0x95e67373U,
272 	0xa0c06060U, 0x98198181U, 0xd19e4f4fU, 0x7fa3dcdcU,
273 	0x66442222U, 0x7e542a2aU, 0xab3b9090U, 0x830b8888U,
274 	0xca8c4646U, 0x29c7eeeeU, 0xd36bb8b8U, 0x3c281414U,
275 	0x79a7dedeU, 0xe2bc5e5eU, 0x1d160b0bU, 0x76addbdbU,
276 	0x3bdbe0e0U, 0x56643232U, 0x4e743a3aU, 0x1e140a0aU,
277 	0xdb924949U, 0x0a0c0606U, 0x6c482424U, 0xe4b85c5cU,
278 	0x5d9fc2c2U, 0x6ebdd3d3U, 0xef43acacU, 0xa6c46262U,
279 	0xa8399191U, 0xa4319595U, 0x37d3e4e4U, 0x8bf27979U,
280 	0x32d5e7e7U, 0x438bc8c8U, 0x596e3737U, 0xb7da6d6dU,
281 	0x8c018d8dU, 0x64b1d5d5U, 0xd29c4e4eU, 0xe049a9a9U,
282 	0xb4d86c6cU, 0xfaac5656U, 0x07f3f4f4U, 0x25cfeaeaU,
283 	0xafca6565U, 0x8ef47a7aU, 0xe947aeaeU, 0x18100808U,
284 	0xd56fbabaU, 0x88f07878U, 0x6f4a2525U, 0x725c2e2eU,
285 	0x24381c1cU, 0xf157a6a6U, 0xc773b4b4U, 0x5197c6c6U,
286 	0x23cbe8e8U, 0x7ca1ddddU, 0x9ce87474U, 0x213e1f1fU,
287 	0xdd964b4bU, 0xdc61bdbdU, 0x860d8b8bU, 0x850f8a8aU,
288 	0x90e07070U, 0x427c3e3eU, 0xc471b5b5U, 0xaacc6666U,
289 	0xd8904848U, 0x05060303U, 0x01f7f6f6U, 0x121c0e0eU,
290 	0xa3c26161U, 0x5f6a3535U, 0xf9ae5757U, 0xd069b9b9U,
291 	0x91178686U, 0x5899c1c1U, 0x273a1d1dU, 0xb9279e9eU,
292 	0x38d9e1e1U, 0x13ebf8f8U, 0xb32b9898U, 0x33221111U,
293 	0xbbd26969U, 0x70a9d9d9U, 0x89078e8eU, 0xa7339494U,
294 	0xb62d9b9bU, 0x223c1e1eU, 0x92158787U, 0x20c9e9e9U,
295 	0x4987ceceU, 0xffaa5555U, 0x78502828U, 0x7aa5dfdfU,
296 	0x8f038c8cU, 0xf859a1a1U, 0x80098989U, 0x171a0d0dU,
297 	0xda65bfbfU, 0x31d7e6e6U, 0xc6844242U, 0xb8d06868U,
298 	0xc3824141U, 0xb0299999U, 0x775a2d2dU, 0x111e0f0fU,
299 	0xcb7bb0b0U, 0xfca85454U, 0xd66dbbbbU, 0x3a2c1616U
300 };
301 
302 
303 static const uint32_t Te2[256] =
304 {
305 	0x63a5c663U, 0x7c84f87cU, 0x7799ee77U, 0x7b8df67bU,
306 	0xf20dfff2U, 0x6bbdd66bU, 0x6fb1de6fU, 0xc55491c5U,
307 	0x30506030U, 0x01030201U, 0x67a9ce67U, 0x2b7d562bU,
308 	0xfe19e7feU, 0xd762b5d7U, 0xabe64dabU, 0x769aec76U,
309 	0xca458fcaU, 0x829d1f82U, 0xc94089c9U, 0x7d87fa7dU,
310 	0xfa15effaU, 0x59ebb259U, 0x47c98e47U, 0xf00bfbf0U,
311 	0xadec41adU, 0xd467b3d4U, 0xa2fd5fa2U, 0xafea45afU,
312 	0x9cbf239cU, 0xa4f753a4U, 0x7296e472U, 0xc05b9bc0U,
313 	0xb7c275b7U, 0xfd1ce1fdU, 0x93ae3d93U, 0x266a4c26U,
314 	0x365a6c36U, 0x3f417e3fU, 0xf702f5f7U, 0xcc4f83ccU,
315 	0x345c6834U, 0xa5f451a5U, 0xe534d1e5U, 0xf108f9f1U,
316 	0x7193e271U, 0xd873abd8U, 0x31536231U, 0x153f2a15U,
317 	0x040c0804U, 0xc75295c7U, 0x23654623U, 0xc35e9dc3U,
318 	0x18283018U, 0x96a13796U, 0x050f0a05U, 0x9ab52f9aU,
319 	0x07090e07U, 0x12362412U, 0x809b1b80U, 0xe23ddfe2U,
320 	0xeb26cdebU, 0x27694e27U, 0xb2cd7fb2U, 0x759fea75U,
321 	0x091b1209U, 0x839e1d83U, 0x2c74582cU, 0x1a2e341aU,
322 	0x1b2d361bU, 0x6eb2dc6eU, 0x5aeeb45aU, 0xa0fb5ba0U,
323 	0x52f6a452U, 0x3b4d763bU, 0xd661b7d6U, 0xb3ce7db3U,
324 	0x297b5229U, 0xe33edde3U, 0x2f715e2fU, 0x84971384U,
325 	0x53f5a653U, 0xd168b9d1U, 0x00000000U, 0xed2cc1edU,
326 	0x20604020U, 0xfc1fe3fcU, 0xb1c879b1U, 0x5bedb65bU,
327 	0x6abed46aU, 0xcb468dcbU, 0xbed967beU, 0x394b7239U,
328 	0x4ade944aU, 0x4cd4984cU, 0x58e8b058U, 0xcf4a85cfU,
329 	0xd06bbbd0U, 0xef2ac5efU, 0xaae54faaU, 0xfb16edfbU,
330 	0x43c58643U, 0x4dd79a4dU, 0x33556633U, 0x85941185U,
331 	0x45cf8a45U, 0xf910e9f9U, 0x02060402U, 0x7f81fe7fU,
332 	0x50f0a050U, 0x3c44783cU, 0x9fba259fU, 0xa8e34ba8U,
333 	0x51f3a251U, 0xa3fe5da3U, 0x40c08040U, 0x8f8a058fU,
334 	0x92ad3f92U, 0x9dbc219dU, 0x38487038U, 0xf504f1f5U,
335 	0xbcdf63bcU, 0xb6c177b6U, 0xda75afdaU, 0x21634221U,
336 	0x10302010U, 0xff1ae5ffU, 0xf30efdf3U, 0xd26dbfd2U,
337 	0xcd4c81cdU, 0x0c14180cU, 0x13352613U, 0xec2fc3ecU,
338 	0x5fe1be5fU, 0x97a23597U, 0x44cc8844U, 0x17392e17U,
339 	0xc45793c4U, 0xa7f255a7U, 0x7e82fc7eU, 0x3d477a3dU,
340 	0x64acc864U, 0x5de7ba5dU, 0x192b3219U, 0x7395e673U,
341 	0x60a0c060U, 0x81981981U, 0x4fd19e4fU, 0xdc7fa3dcU,
342 	0x22664422U, 0x2a7e542aU, 0x90ab3b90U, 0x88830b88U,
343 	0x46ca8c46U, 0xee29c7eeU, 0xb8d36bb8U, 0x143c2814U,
344 	0xde79a7deU, 0x5ee2bc5eU, 0x0b1d160bU, 0xdb76addbU,
345 	0xe03bdbe0U, 0x32566432U, 0x3a4e743aU, 0x0a1e140aU,
346 	0x49db9249U, 0x060a0c06U, 0x246c4824U, 0x5ce4b85cU,
347 	0xc25d9fc2U, 0xd36ebdd3U, 0xacef43acU, 0x62a6c462U,
348 	0x91a83991U, 0x95a43195U, 0xe437d3e4U, 0x798bf279U,
349 	0xe732d5e7U, 0xc8438bc8U, 0x37596e37U, 0x6db7da6dU,
350 	0x8d8c018dU, 0xd564b1d5U, 0x4ed29c4eU, 0xa9e049a9U,
351 	0x6cb4d86cU, 0x56faac56U, 0xf407f3f4U, 0xea25cfeaU,
352 	0x65afca65U, 0x7a8ef47aU, 0xaee947aeU, 0x08181008U,
353 	0xbad56fbaU, 0x7888f078U, 0x256f4a25U, 0x2e725c2eU,
354 	0x1c24381cU, 0xa6f157a6U, 0xb4c773b4U, 0xc65197c6U,
355 	0xe823cbe8U, 0xdd7ca1ddU, 0x749ce874U, 0x1f213e1fU,
356 	0x4bdd964bU, 0xbddc61bdU, 0x8b860d8bU, 0x8a850f8aU,
357 	0x7090e070U, 0x3e427c3eU, 0xb5c471b5U, 0x66aacc66U,
358 	0x48d89048U, 0x03050603U, 0xf601f7f6U, 0x0e121c0eU,
359 	0x61a3c261U, 0x355f6a35U, 0x57f9ae57U, 0xb9d069b9U,
360 	0x86911786U, 0xc15899c1U, 0x1d273a1dU, 0x9eb9279eU,
361 	0xe138d9e1U, 0xf813ebf8U, 0x98b32b98U, 0x11332211U,
362 	0x69bbd269U, 0xd970a9d9U, 0x8e89078eU, 0x94a73394U,
363 	0x9bb62d9bU, 0x1e223c1eU, 0x87921587U, 0xe920c9e9U,
364 	0xce4987ceU, 0x55ffaa55U, 0x28785028U, 0xdf7aa5dfU,
365 	0x8c8f038cU, 0xa1f859a1U, 0x89800989U, 0x0d171a0dU,
366 	0xbfda65bfU, 0xe631d7e6U, 0x42c68442U, 0x68b8d068U,
367 	0x41c38241U, 0x99b02999U, 0x2d775a2dU, 0x0f111e0fU,
368 	0xb0cb7bb0U, 0x54fca854U, 0xbbd66dbbU, 0x163a2c16U
369 };
370 
371 
372 static const uint32_t Te3[256] =
373 {
374 	0x6363a5c6U, 0x7c7c84f8U, 0x777799eeU, 0x7b7b8df6U,
375 	0xf2f20dffU, 0x6b6bbdd6U, 0x6f6fb1deU, 0xc5c55491U,
376 	0x30305060U, 0x01010302U, 0x6767a9ceU, 0x2b2b7d56U,
377 	0xfefe19e7U, 0xd7d762b5U, 0xababe64dU, 0x76769aecU,
378 	0xcaca458fU, 0x82829d1fU, 0xc9c94089U, 0x7d7d87faU,
379 	0xfafa15efU, 0x5959ebb2U, 0x4747c98eU, 0xf0f00bfbU,
380 	0xadadec41U, 0xd4d467b3U, 0xa2a2fd5fU, 0xafafea45U,
381 	0x9c9cbf23U, 0xa4a4f753U, 0x727296e4U, 0xc0c05b9bU,
382 	0xb7b7c275U, 0xfdfd1ce1U, 0x9393ae3dU, 0x26266a4cU,
383 	0x36365a6cU, 0x3f3f417eU, 0xf7f702f5U, 0xcccc4f83U,
384 	0x34345c68U, 0xa5a5f451U, 0xe5e534d1U, 0xf1f108f9U,
385 	0x717193e2U, 0xd8d873abU, 0x31315362U, 0x15153f2aU,
386 	0x04040c08U, 0xc7c75295U, 0x23236546U, 0xc3c35e9dU,
387 	0x18182830U, 0x9696a137U, 0x05050f0aU, 0x9a9ab52fU,
388 	0x0707090eU, 0x12123624U, 0x80809b1bU, 0xe2e23ddfU,
389 	0xebeb26cdU, 0x2727694eU, 0xb2b2cd7fU, 0x75759feaU,
390 	0x09091b12U, 0x83839e1dU, 0x2c2c7458U, 0x1a1a2e34U,
391 	0x1b1b2d36U, 0x6e6eb2dcU, 0x5a5aeeb4U, 0xa0a0fb5bU,
392 	0x5252f6a4U, 0x3b3b4d76U, 0xd6d661b7U, 0xb3b3ce7dU,
393 	0x29297b52U, 0xe3e33eddU, 0x2f2f715eU, 0x84849713U,
394 	0x5353f5a6U, 0xd1d168b9U, 0x00000000U, 0xeded2cc1U,
395 	0x20206040U, 0xfcfc1fe3U, 0xb1b1c879U, 0x5b5bedb6U,
396 	0x6a6abed4U, 0xcbcb468dU, 0xbebed967U, 0x39394b72U,
397 	0x4a4ade94U, 0x4c4cd498U, 0x5858e8b0U, 0xcfcf4a85U,
398 	0xd0d06bbbU, 0xefef2ac5U, 0xaaaae54fU, 0xfbfb16edU,
399 	0x4343c586U, 0x4d4dd79aU, 0x33335566U, 0x85859411U,
400 	0x4545cf8aU, 0xf9f910e9U, 0x02020604U, 0x7f7f81feU,
401 	0x5050f0a0U, 0x3c3c4478U, 0x9f9fba25U, 0xa8a8e34bU,
402 	0x5151f3a2U, 0xa3a3fe5dU, 0x4040c080U, 0x8f8f8a05U,
403 	0x9292ad3fU, 0x9d9dbc21U, 0x38384870U, 0xf5f504f1U,
404 	0xbcbcdf63U, 0xb6b6c177U, 0xdada75afU, 0x21216342U,
405 	0x10103020U, 0xffff1ae5U, 0xf3f30efdU, 0xd2d26dbfU,
406 	0xcdcd4c81U, 0x0c0c1418U, 0x13133526U, 0xecec2fc3U,
407 	0x5f5fe1beU, 0x9797a235U, 0x4444cc88U, 0x1717392eU,
408 	0xc4c45793U, 0xa7a7f255U, 0x7e7e82fcU, 0x3d3d477aU,
409 	0x6464acc8U, 0x5d5de7baU, 0x19192b32U, 0x737395e6U,
410 	0x6060a0c0U, 0x81819819U, 0x4f4fd19eU, 0xdcdc7fa3U,
411 	0x22226644U, 0x2a2a7e54U, 0x9090ab3bU, 0x8888830bU,
412 	0x4646ca8cU, 0xeeee29c7U, 0xb8b8d36bU, 0x14143c28U,
413 	0xdede79a7U, 0x5e5ee2bcU, 0x0b0b1d16U, 0xdbdb76adU,
414 	0xe0e03bdbU, 0x32325664U, 0x3a3a4e74U, 0x0a0a1e14U,
415 	0x4949db92U, 0x06060a0cU, 0x24246c48U, 0x5c5ce4b8U,
416 	0xc2c25d9fU, 0xd3d36ebdU, 0xacacef43U, 0x6262a6c4U,
417 	0x9191a839U, 0x9595a431U, 0xe4e437d3U, 0x79798bf2U,
418 	0xe7e732d5U, 0xc8c8438bU, 0x3737596eU, 0x6d6db7daU,
419 	0x8d8d8c01U, 0xd5d564b1U, 0x4e4ed29cU, 0xa9a9e049U,
420 	0x6c6cb4d8U, 0x5656faacU, 0xf4f407f3U, 0xeaea25cfU,
421 	0x6565afcaU, 0x7a7a8ef4U, 0xaeaee947U, 0x08081810U,
422 	0xbabad56fU, 0x787888f0U, 0x25256f4aU, 0x2e2e725cU,
423 	0x1c1c2438U, 0xa6a6f157U, 0xb4b4c773U, 0xc6c65197U,
424 	0xe8e823cbU, 0xdddd7ca1U, 0x74749ce8U, 0x1f1f213eU,
425 	0x4b4bdd96U, 0xbdbddc61U, 0x8b8b860dU, 0x8a8a850fU,
426 	0x707090e0U, 0x3e3e427cU, 0xb5b5c471U, 0x6666aaccU,
427 	0x4848d890U, 0x03030506U, 0xf6f601f7U, 0x0e0e121cU,
428 	0x6161a3c2U, 0x35355f6aU, 0x5757f9aeU, 0xb9b9d069U,
429 	0x86869117U, 0xc1c15899U, 0x1d1d273aU, 0x9e9eb927U,
430 	0xe1e138d9U, 0xf8f813ebU, 0x9898b32bU, 0x11113322U,
431 	0x6969bbd2U, 0xd9d970a9U, 0x8e8e8907U, 0x9494a733U,
432 	0x9b9bb62dU, 0x1e1e223cU, 0x87879215U, 0xe9e920c9U,
433 	0xcece4987U, 0x5555ffaaU, 0x28287850U, 0xdfdf7aa5U,
434 	0x8c8c8f03U, 0xa1a1f859U, 0x89898009U, 0x0d0d171aU,
435 	0xbfbfda65U, 0xe6e631d7U, 0x4242c684U, 0x6868b8d0U,
436 	0x4141c382U, 0x9999b029U, 0x2d2d775aU, 0x0f0f111eU,
437 	0xb0b0cb7bU, 0x5454fca8U, 0xbbbbd66dU, 0x16163a2cU
438 };
439 
440 #endif /* !sun4u */
441 
442 static const uint32_t Te4[256] =
443 {
444 	0x63636363U, 0x7c7c7c7cU, 0x77777777U, 0x7b7b7b7bU,
445 	0xf2f2f2f2U, 0x6b6b6b6bU, 0x6f6f6f6fU, 0xc5c5c5c5U,
446 	0x30303030U, 0x01010101U, 0x67676767U, 0x2b2b2b2bU,
447 	0xfefefefeU, 0xd7d7d7d7U, 0xababababU, 0x76767676U,
448 	0xcacacacaU, 0x82828282U, 0xc9c9c9c9U, 0x7d7d7d7dU,
449 	0xfafafafaU, 0x59595959U, 0x47474747U, 0xf0f0f0f0U,
450 	0xadadadadU, 0xd4d4d4d4U, 0xa2a2a2a2U, 0xafafafafU,
451 	0x9c9c9c9cU, 0xa4a4a4a4U, 0x72727272U, 0xc0c0c0c0U,
452 	0xb7b7b7b7U, 0xfdfdfdfdU, 0x93939393U, 0x26262626U,
453 	0x36363636U, 0x3f3f3f3fU, 0xf7f7f7f7U, 0xccccccccU,
454 	0x34343434U, 0xa5a5a5a5U, 0xe5e5e5e5U, 0xf1f1f1f1U,
455 	0x71717171U, 0xd8d8d8d8U, 0x31313131U, 0x15151515U,
456 	0x04040404U, 0xc7c7c7c7U, 0x23232323U, 0xc3c3c3c3U,
457 	0x18181818U, 0x96969696U, 0x05050505U, 0x9a9a9a9aU,
458 	0x07070707U, 0x12121212U, 0x80808080U, 0xe2e2e2e2U,
459 	0xebebebebU, 0x27272727U, 0xb2b2b2b2U, 0x75757575U,
460 	0x09090909U, 0x83838383U, 0x2c2c2c2cU, 0x1a1a1a1aU,
461 	0x1b1b1b1bU, 0x6e6e6e6eU, 0x5a5a5a5aU, 0xa0a0a0a0U,
462 	0x52525252U, 0x3b3b3b3bU, 0xd6d6d6d6U, 0xb3b3b3b3U,
463 	0x29292929U, 0xe3e3e3e3U, 0x2f2f2f2fU, 0x84848484U,
464 	0x53535353U, 0xd1d1d1d1U, 0x00000000U, 0xededededU,
465 	0x20202020U, 0xfcfcfcfcU, 0xb1b1b1b1U, 0x5b5b5b5bU,
466 	0x6a6a6a6aU, 0xcbcbcbcbU, 0xbebebebeU, 0x39393939U,
467 	0x4a4a4a4aU, 0x4c4c4c4cU, 0x58585858U, 0xcfcfcfcfU,
468 	0xd0d0d0d0U, 0xefefefefU, 0xaaaaaaaaU, 0xfbfbfbfbU,
469 	0x43434343U, 0x4d4d4d4dU, 0x33333333U, 0x85858585U,
470 	0x45454545U, 0xf9f9f9f9U, 0x02020202U, 0x7f7f7f7fU,
471 	0x50505050U, 0x3c3c3c3cU, 0x9f9f9f9fU, 0xa8a8a8a8U,
472 	0x51515151U, 0xa3a3a3a3U, 0x40404040U, 0x8f8f8f8fU,
473 	0x92929292U, 0x9d9d9d9dU, 0x38383838U, 0xf5f5f5f5U,
474 	0xbcbcbcbcU, 0xb6b6b6b6U, 0xdadadadaU, 0x21212121U,
475 	0x10101010U, 0xffffffffU, 0xf3f3f3f3U, 0xd2d2d2d2U,
476 	0xcdcdcdcdU, 0x0c0c0c0cU, 0x13131313U, 0xececececU,
477 	0x5f5f5f5fU, 0x97979797U, 0x44444444U, 0x17171717U,
478 	0xc4c4c4c4U, 0xa7a7a7a7U, 0x7e7e7e7eU, 0x3d3d3d3dU,
479 	0x64646464U, 0x5d5d5d5dU, 0x19191919U, 0x73737373U,
480 	0x60606060U, 0x81818181U, 0x4f4f4f4fU, 0xdcdcdcdcU,
481 	0x22222222U, 0x2a2a2a2aU, 0x90909090U, 0x88888888U,
482 	0x46464646U, 0xeeeeeeeeU, 0xb8b8b8b8U, 0x14141414U,
483 	0xdedededeU, 0x5e5e5e5eU, 0x0b0b0b0bU, 0xdbdbdbdbU,
484 	0xe0e0e0e0U, 0x32323232U, 0x3a3a3a3aU, 0x0a0a0a0aU,
485 	0x49494949U, 0x06060606U, 0x24242424U, 0x5c5c5c5cU,
486 	0xc2c2c2c2U, 0xd3d3d3d3U, 0xacacacacU, 0x62626262U,
487 	0x91919191U, 0x95959595U, 0xe4e4e4e4U, 0x79797979U,
488 	0xe7e7e7e7U, 0xc8c8c8c8U, 0x37373737U, 0x6d6d6d6dU,
489 	0x8d8d8d8dU, 0xd5d5d5d5U, 0x4e4e4e4eU, 0xa9a9a9a9U,
490 	0x6c6c6c6cU, 0x56565656U, 0xf4f4f4f4U, 0xeaeaeaeaU,
491 	0x65656565U, 0x7a7a7a7aU, 0xaeaeaeaeU, 0x08080808U,
492 	0xbabababaU, 0x78787878U, 0x25252525U, 0x2e2e2e2eU,
493 	0x1c1c1c1cU, 0xa6a6a6a6U, 0xb4b4b4b4U, 0xc6c6c6c6U,
494 	0xe8e8e8e8U, 0xddddddddU, 0x74747474U, 0x1f1f1f1fU,
495 	0x4b4b4b4bU, 0xbdbdbdbdU, 0x8b8b8b8bU, 0x8a8a8a8aU,
496 	0x70707070U, 0x3e3e3e3eU, 0xb5b5b5b5U, 0x66666666U,
497 	0x48484848U, 0x03030303U, 0xf6f6f6f6U, 0x0e0e0e0eU,
498 	0x61616161U, 0x35353535U, 0x57575757U, 0xb9b9b9b9U,
499 	0x86868686U, 0xc1c1c1c1U, 0x1d1d1d1dU, 0x9e9e9e9eU,
500 	0xe1e1e1e1U, 0xf8f8f8f8U, 0x98989898U, 0x11111111U,
501 	0x69696969U, 0xd9d9d9d9U, 0x8e8e8e8eU, 0x94949494U,
502 	0x9b9b9b9bU, 0x1e1e1e1eU, 0x87878787U, 0xe9e9e9e9U,
503 	0xcecececeU, 0x55555555U, 0x28282828U, 0xdfdfdfdfU,
504 	0x8c8c8c8cU, 0xa1a1a1a1U, 0x89898989U, 0x0d0d0d0dU,
505 	0xbfbfbfbfU, 0xe6e6e6e6U, 0x42424242U, 0x68686868U,
506 	0x41414141U, 0x99999999U, 0x2d2d2d2dU, 0x0f0f0f0fU,
507 	0xb0b0b0b0U, 0x54545454U, 0xbbbbbbbbU, 0x16161616U
508 };
509 
510 /* Decrypt Sbox constants (for the substitute bytes operation) */
511 
512 static const uint32_t Td0[256] =
513 {
514 	0x51f4a750U, 0x7e416553U, 0x1a17a4c3U, 0x3a275e96U,
515 	0x3bab6bcbU, 0x1f9d45f1U, 0xacfa58abU, 0x4be30393U,
516 	0x2030fa55U, 0xad766df6U, 0x88cc7691U, 0xf5024c25U,
517 	0x4fe5d7fcU, 0xc52acbd7U, 0x26354480U, 0xb562a38fU,
518 	0xdeb15a49U, 0x25ba1b67U, 0x45ea0e98U, 0x5dfec0e1U,
519 	0xc32f7502U, 0x814cf012U, 0x8d4697a3U, 0x6bd3f9c6U,
520 	0x038f5fe7U, 0x15929c95U, 0xbf6d7aebU, 0x955259daU,
521 	0xd4be832dU, 0x587421d3U, 0x49e06929U, 0x8ec9c844U,
522 	0x75c2896aU, 0xf48e7978U, 0x99583e6bU, 0x27b971ddU,
523 	0xbee14fb6U, 0xf088ad17U, 0xc920ac66U, 0x7dce3ab4U,
524 	0x63df4a18U, 0xe51a3182U, 0x97513360U, 0x62537f45U,
525 	0xb16477e0U, 0xbb6bae84U, 0xfe81a01cU, 0xf9082b94U,
526 	0x70486858U, 0x8f45fd19U, 0x94de6c87U, 0x527bf8b7U,
527 	0xab73d323U, 0x724b02e2U, 0xe31f8f57U, 0x6655ab2aU,
528 	0xb2eb2807U, 0x2fb5c203U, 0x86c57b9aU, 0xd33708a5U,
529 	0x302887f2U, 0x23bfa5b2U, 0x02036abaU, 0xed16825cU,
530 	0x8acf1c2bU, 0xa779b492U, 0xf307f2f0U, 0x4e69e2a1U,
531 	0x65daf4cdU, 0x0605bed5U, 0xd134621fU, 0xc4a6fe8aU,
532 	0x342e539dU, 0xa2f355a0U, 0x058ae132U, 0xa4f6eb75U,
533 	0x0b83ec39U, 0x4060efaaU, 0x5e719f06U, 0xbd6e1051U,
534 	0x3e218af9U, 0x96dd063dU, 0xdd3e05aeU, 0x4de6bd46U,
535 	0x91548db5U, 0x71c45d05U, 0x0406d46fU, 0x605015ffU,
536 	0x1998fb24U, 0xd6bde997U, 0x894043ccU, 0x67d99e77U,
537 	0xb0e842bdU, 0x07898b88U, 0xe7195b38U, 0x79c8eedbU,
538 	0xa17c0a47U, 0x7c420fe9U, 0xf8841ec9U, 0x00000000U,
539 	0x09808683U, 0x322bed48U, 0x1e1170acU, 0x6c5a724eU,
540 	0xfd0efffbU, 0x0f853856U, 0x3daed51eU, 0x362d3927U,
541 	0x0a0fd964U, 0x685ca621U, 0x9b5b54d1U, 0x24362e3aU,
542 	0x0c0a67b1U, 0x9357e70fU, 0xb4ee96d2U, 0x1b9b919eU,
543 	0x80c0c54fU, 0x61dc20a2U, 0x5a774b69U, 0x1c121a16U,
544 	0xe293ba0aU, 0xc0a02ae5U, 0x3c22e043U, 0x121b171dU,
545 	0x0e090d0bU, 0xf28bc7adU, 0x2db6a8b9U, 0x141ea9c8U,
546 	0x57f11985U, 0xaf75074cU, 0xee99ddbbU, 0xa37f60fdU,
547 	0xf701269fU, 0x5c72f5bcU, 0x44663bc5U, 0x5bfb7e34U,
548 	0x8b432976U, 0xcb23c6dcU, 0xb6edfc68U, 0xb8e4f163U,
549 	0xd731dccaU, 0x42638510U, 0x13972240U, 0x84c61120U,
550 	0x854a247dU, 0xd2bb3df8U, 0xaef93211U, 0xc729a16dU,
551 	0x1d9e2f4bU, 0xdcb230f3U, 0x0d8652ecU, 0x77c1e3d0U,
552 	0x2bb3166cU, 0xa970b999U, 0x119448faU, 0x47e96422U,
553 	0xa8fc8cc4U, 0xa0f03f1aU, 0x567d2cd8U, 0x223390efU,
554 	0x87494ec7U, 0xd938d1c1U, 0x8ccaa2feU, 0x98d40b36U,
555 	0xa6f581cfU, 0xa57ade28U, 0xdab78e26U, 0x3fadbfa4U,
556 	0x2c3a9de4U, 0x5078920dU, 0x6a5fcc9bU, 0x547e4662U,
557 	0xf68d13c2U, 0x90d8b8e8U, 0x2e39f75eU, 0x82c3aff5U,
558 	0x9f5d80beU, 0x69d0937cU, 0x6fd52da9U, 0xcf2512b3U,
559 	0xc8ac993bU, 0x10187da7U, 0xe89c636eU, 0xdb3bbb7bU,
560 	0xcd267809U, 0x6e5918f4U, 0xec9ab701U, 0x834f9aa8U,
561 	0xe6956e65U, 0xaaffe67eU, 0x21bccf08U, 0xef15e8e6U,
562 	0xbae79bd9U, 0x4a6f36ceU, 0xea9f09d4U, 0x29b07cd6U,
563 	0x31a4b2afU, 0x2a3f2331U, 0xc6a59430U, 0x35a266c0U,
564 	0x744ebc37U, 0xfc82caa6U, 0xe090d0b0U, 0x33a7d815U,
565 	0xf104984aU, 0x41ecdaf7U, 0x7fcd500eU, 0x1791f62fU,
566 	0x764dd68dU, 0x43efb04dU, 0xccaa4d54U, 0xe49604dfU,
567 	0x9ed1b5e3U, 0x4c6a881bU, 0xc12c1fb8U, 0x4665517fU,
568 	0x9d5eea04U, 0x018c355dU, 0xfa877473U, 0xfb0b412eU,
569 	0xb3671d5aU, 0x92dbd252U, 0xe9105633U, 0x6dd64713U,
570 	0x9ad7618cU, 0x37a10c7aU, 0x59f8148eU, 0xeb133c89U,
571 	0xcea927eeU, 0xb761c935U, 0xe11ce5edU, 0x7a47b13cU,
572 	0x9cd2df59U, 0x55f2733fU, 0x1814ce79U, 0x73c737bfU,
573 	0x53f7cdeaU, 0x5ffdaa5bU, 0xdf3d6f14U, 0x7844db86U,
574 	0xcaaff381U, 0xb968c43eU, 0x3824342cU, 0xc2a3405fU,
575 	0x161dc372U, 0xbce2250cU, 0x283c498bU, 0xff0d9541U,
576 	0x39a80171U, 0x080cb3deU, 0xd8b4e49cU, 0x6456c190U,
577 	0x7bcb8461U, 0xd532b670U, 0x486c5c74U, 0xd0b85742U
578 };
579 
580 static const uint32_t Td1[256] =
581 {
582 	0x5051f4a7U, 0x537e4165U, 0xc31a17a4U, 0x963a275eU,
583 	0xcb3bab6bU, 0xf11f9d45U, 0xabacfa58U, 0x934be303U,
584 	0x552030faU, 0xf6ad766dU, 0x9188cc76U, 0x25f5024cU,
585 	0xfc4fe5d7U, 0xd7c52acbU, 0x80263544U, 0x8fb562a3U,
586 	0x49deb15aU, 0x6725ba1bU, 0x9845ea0eU, 0xe15dfec0U,
587 	0x02c32f75U, 0x12814cf0U, 0xa38d4697U, 0xc66bd3f9U,
588 	0xe7038f5fU, 0x9515929cU, 0xebbf6d7aU, 0xda955259U,
589 	0x2dd4be83U, 0xd3587421U, 0x2949e069U, 0x448ec9c8U,
590 	0x6a75c289U, 0x78f48e79U, 0x6b99583eU, 0xdd27b971U,
591 	0xb6bee14fU, 0x17f088adU, 0x66c920acU, 0xb47dce3aU,
592 	0x1863df4aU, 0x82e51a31U, 0x60975133U, 0x4562537fU,
593 	0xe0b16477U, 0x84bb6baeU, 0x1cfe81a0U, 0x94f9082bU,
594 	0x58704868U, 0x198f45fdU, 0x8794de6cU, 0xb7527bf8U,
595 	0x23ab73d3U, 0xe2724b02U, 0x57e31f8fU, 0x2a6655abU,
596 	0x07b2eb28U, 0x032fb5c2U, 0x9a86c57bU, 0xa5d33708U,
597 	0xf2302887U, 0xb223bfa5U, 0xba02036aU, 0x5ced1682U,
598 	0x2b8acf1cU, 0x92a779b4U, 0xf0f307f2U, 0xa14e69e2U,
599 	0xcd65daf4U, 0xd50605beU, 0x1fd13462U, 0x8ac4a6feU,
600 	0x9d342e53U, 0xa0a2f355U, 0x32058ae1U, 0x75a4f6ebU,
601 	0x390b83ecU, 0xaa4060efU, 0x065e719fU, 0x51bd6e10U,
602 	0xf93e218aU, 0x3d96dd06U, 0xaedd3e05U, 0x464de6bdU,
603 	0xb591548dU, 0x0571c45dU, 0x6f0406d4U, 0xff605015U,
604 	0x241998fbU, 0x97d6bde9U, 0xcc894043U, 0x7767d99eU,
605 	0xbdb0e842U, 0x8807898bU, 0x38e7195bU, 0xdb79c8eeU,
606 	0x47a17c0aU, 0xe97c420fU, 0xc9f8841eU, 0x00000000U,
607 	0x83098086U, 0x48322bedU, 0xac1e1170U, 0x4e6c5a72U,
608 	0xfbfd0effU, 0x560f8538U, 0x1e3daed5U, 0x27362d39U,
609 	0x640a0fd9U, 0x21685ca6U, 0xd19b5b54U, 0x3a24362eU,
610 	0xb10c0a67U, 0x0f9357e7U, 0xd2b4ee96U, 0x9e1b9b91U,
611 	0x4f80c0c5U, 0xa261dc20U, 0x695a774bU, 0x161c121aU,
612 	0x0ae293baU, 0xe5c0a02aU, 0x433c22e0U, 0x1d121b17U,
613 	0x0b0e090dU, 0xadf28bc7U, 0xb92db6a8U, 0xc8141ea9U,
614 	0x8557f119U, 0x4caf7507U, 0xbbee99ddU, 0xfda37f60U,
615 	0x9ff70126U, 0xbc5c72f5U, 0xc544663bU, 0x345bfb7eU,
616 	0x768b4329U, 0xdccb23c6U, 0x68b6edfcU, 0x63b8e4f1U,
617 	0xcad731dcU, 0x10426385U, 0x40139722U, 0x2084c611U,
618 	0x7d854a24U, 0xf8d2bb3dU, 0x11aef932U, 0x6dc729a1U,
619 	0x4b1d9e2fU, 0xf3dcb230U, 0xec0d8652U, 0xd077c1e3U,
620 	0x6c2bb316U, 0x99a970b9U, 0xfa119448U, 0x2247e964U,
621 	0xc4a8fc8cU, 0x1aa0f03fU, 0xd8567d2cU, 0xef223390U,
622 	0xc787494eU, 0xc1d938d1U, 0xfe8ccaa2U, 0x3698d40bU,
623 	0xcfa6f581U, 0x28a57adeU, 0x26dab78eU, 0xa43fadbfU,
624 	0xe42c3a9dU, 0x0d507892U, 0x9b6a5fccU, 0x62547e46U,
625 	0xc2f68d13U, 0xe890d8b8U, 0x5e2e39f7U, 0xf582c3afU,
626 	0xbe9f5d80U, 0x7c69d093U, 0xa96fd52dU, 0xb3cf2512U,
627 	0x3bc8ac99U, 0xa710187dU, 0x6ee89c63U, 0x7bdb3bbbU,
628 	0x09cd2678U, 0xf46e5918U, 0x01ec9ab7U, 0xa8834f9aU,
629 	0x65e6956eU, 0x7eaaffe6U, 0x0821bccfU, 0xe6ef15e8U,
630 	0xd9bae79bU, 0xce4a6f36U, 0xd4ea9f09U, 0xd629b07cU,
631 	0xaf31a4b2U, 0x312a3f23U, 0x30c6a594U, 0xc035a266U,
632 	0x37744ebcU, 0xa6fc82caU, 0xb0e090d0U, 0x1533a7d8U,
633 	0x4af10498U, 0xf741ecdaU, 0x0e7fcd50U, 0x2f1791f6U,
634 	0x8d764dd6U, 0x4d43efb0U, 0x54ccaa4dU, 0xdfe49604U,
635 	0xe39ed1b5U, 0x1b4c6a88U, 0xb8c12c1fU, 0x7f466551U,
636 	0x049d5eeaU, 0x5d018c35U, 0x73fa8774U, 0x2efb0b41U,
637 	0x5ab3671dU, 0x5292dbd2U, 0x33e91056U, 0x136dd647U,
638 	0x8c9ad761U, 0x7a37a10cU, 0x8e59f814U, 0x89eb133cU,
639 	0xeecea927U, 0x35b761c9U, 0xede11ce5U, 0x3c7a47b1U,
640 	0x599cd2dfU, 0x3f55f273U, 0x791814ceU, 0xbf73c737U,
641 	0xea53f7cdU, 0x5b5ffdaaU, 0x14df3d6fU, 0x867844dbU,
642 	0x81caaff3U, 0x3eb968c4U, 0x2c382434U, 0x5fc2a340U,
643 	0x72161dc3U, 0x0cbce225U, 0x8b283c49U, 0x41ff0d95U,
644 	0x7139a801U, 0xde080cb3U, 0x9cd8b4e4U, 0x906456c1U,
645 	0x617bcb84U, 0x70d532b6U, 0x74486c5cU, 0x42d0b857U
646 };
647 
648 static const uint32_t Td2[256] =
649 {
650 	0xa75051f4U, 0x65537e41U, 0xa4c31a17U, 0x5e963a27U,
651 	0x6bcb3babU, 0x45f11f9dU, 0x58abacfaU, 0x03934be3U,
652 	0xfa552030U, 0x6df6ad76U, 0x769188ccU, 0x4c25f502U,
653 	0xd7fc4fe5U, 0xcbd7c52aU, 0x44802635U, 0xa38fb562U,
654 	0x5a49deb1U, 0x1b6725baU, 0x0e9845eaU, 0xc0e15dfeU,
655 	0x7502c32fU, 0xf012814cU, 0x97a38d46U, 0xf9c66bd3U,
656 	0x5fe7038fU, 0x9c951592U, 0x7aebbf6dU, 0x59da9552U,
657 	0x832dd4beU, 0x21d35874U, 0x692949e0U, 0xc8448ec9U,
658 	0x896a75c2U, 0x7978f48eU, 0x3e6b9958U, 0x71dd27b9U,
659 	0x4fb6bee1U, 0xad17f088U, 0xac66c920U, 0x3ab47dceU,
660 	0x4a1863dfU, 0x3182e51aU, 0x33609751U, 0x7f456253U,
661 	0x77e0b164U, 0xae84bb6bU, 0xa01cfe81U, 0x2b94f908U,
662 	0x68587048U, 0xfd198f45U, 0x6c8794deU, 0xf8b7527bU,
663 	0xd323ab73U, 0x02e2724bU, 0x8f57e31fU, 0xab2a6655U,
664 	0x2807b2ebU, 0xc2032fb5U, 0x7b9a86c5U, 0x08a5d337U,
665 	0x87f23028U, 0xa5b223bfU, 0x6aba0203U, 0x825ced16U,
666 	0x1c2b8acfU, 0xb492a779U, 0xf2f0f307U, 0xe2a14e69U,
667 	0xf4cd65daU, 0xbed50605U, 0x621fd134U, 0xfe8ac4a6U,
668 	0x539d342eU, 0x55a0a2f3U, 0xe132058aU, 0xeb75a4f6U,
669 	0xec390b83U, 0xefaa4060U, 0x9f065e71U, 0x1051bd6eU,
670 	0x8af93e21U, 0x063d96ddU, 0x05aedd3eU, 0xbd464de6U,
671 	0x8db59154U, 0x5d0571c4U, 0xd46f0406U, 0x15ff6050U,
672 	0xfb241998U, 0xe997d6bdU, 0x43cc8940U, 0x9e7767d9U,
673 	0x42bdb0e8U, 0x8b880789U, 0x5b38e719U, 0xeedb79c8U,
674 	0x0a47a17cU, 0x0fe97c42U, 0x1ec9f884U, 0x00000000U,
675 	0x86830980U, 0xed48322bU, 0x70ac1e11U, 0x724e6c5aU,
676 	0xfffbfd0eU, 0x38560f85U, 0xd51e3daeU, 0x3927362dU,
677 	0xd9640a0fU, 0xa621685cU, 0x54d19b5bU, 0x2e3a2436U,
678 	0x67b10c0aU, 0xe70f9357U, 0x96d2b4eeU, 0x919e1b9bU,
679 	0xc54f80c0U, 0x20a261dcU, 0x4b695a77U, 0x1a161c12U,
680 	0xba0ae293U, 0x2ae5c0a0U, 0xe0433c22U, 0x171d121bU,
681 	0x0d0b0e09U, 0xc7adf28bU, 0xa8b92db6U, 0xa9c8141eU,
682 	0x198557f1U, 0x074caf75U, 0xddbbee99U, 0x60fda37fU,
683 	0x269ff701U, 0xf5bc5c72U, 0x3bc54466U, 0x7e345bfbU,
684 	0x29768b43U, 0xc6dccb23U, 0xfc68b6edU, 0xf163b8e4U,
685 	0xdccad731U, 0x85104263U, 0x22401397U, 0x112084c6U,
686 	0x247d854aU, 0x3df8d2bbU, 0x3211aef9U, 0xa16dc729U,
687 	0x2f4b1d9eU, 0x30f3dcb2U, 0x52ec0d86U, 0xe3d077c1U,
688 	0x166c2bb3U, 0xb999a970U, 0x48fa1194U, 0x642247e9U,
689 	0x8cc4a8fcU, 0x3f1aa0f0U, 0x2cd8567dU, 0x90ef2233U,
690 	0x4ec78749U, 0xd1c1d938U, 0xa2fe8ccaU, 0x0b3698d4U,
691 	0x81cfa6f5U, 0xde28a57aU, 0x8e26dab7U, 0xbfa43fadU,
692 	0x9de42c3aU, 0x920d5078U, 0xcc9b6a5fU, 0x4662547eU,
693 	0x13c2f68dU, 0xb8e890d8U, 0xf75e2e39U, 0xaff582c3U,
694 	0x80be9f5dU, 0x937c69d0U, 0x2da96fd5U, 0x12b3cf25U,
695 	0x993bc8acU, 0x7da71018U, 0x636ee89cU, 0xbb7bdb3bU,
696 	0x7809cd26U, 0x18f46e59U, 0xb701ec9aU, 0x9aa8834fU,
697 	0x6e65e695U, 0xe67eaaffU, 0xcf0821bcU, 0xe8e6ef15U,
698 	0x9bd9bae7U, 0x36ce4a6fU, 0x09d4ea9fU, 0x7cd629b0U,
699 	0xb2af31a4U, 0x23312a3fU, 0x9430c6a5U, 0x66c035a2U,
700 	0xbc37744eU, 0xcaa6fc82U, 0xd0b0e090U, 0xd81533a7U,
701 	0x984af104U, 0xdaf741ecU, 0x500e7fcdU, 0xf62f1791U,
702 	0xd68d764dU, 0xb04d43efU, 0x4d54ccaaU, 0x04dfe496U,
703 	0xb5e39ed1U, 0x881b4c6aU, 0x1fb8c12cU, 0x517f4665U,
704 	0xea049d5eU, 0x355d018cU, 0x7473fa87U, 0x412efb0bU,
705 	0x1d5ab367U, 0xd25292dbU, 0x5633e910U, 0x47136dd6U,
706 	0x618c9ad7U, 0x0c7a37a1U, 0x148e59f8U, 0x3c89eb13U,
707 	0x27eecea9U, 0xc935b761U, 0xe5ede11cU, 0xb13c7a47U,
708 	0xdf599cd2U, 0x733f55f2U, 0xce791814U, 0x37bf73c7U,
709 	0xcdea53f7U, 0xaa5b5ffdU, 0x6f14df3dU, 0xdb867844U,
710 	0xf381caafU, 0xc43eb968U, 0x342c3824U, 0x405fc2a3U,
711 	0xc372161dU, 0x250cbce2U, 0x498b283cU, 0x9541ff0dU,
712 	0x017139a8U, 0xb3de080cU, 0xe49cd8b4U, 0xc1906456U,
713 	0x84617bcbU, 0xb670d532U, 0x5c74486cU, 0x5742d0b8U
714 };
715 
716 static const uint32_t Td3[256] =
717 {
718 	0xf4a75051U, 0x4165537eU, 0x17a4c31aU, 0x275e963aU,
719 	0xab6bcb3bU, 0x9d45f11fU, 0xfa58abacU, 0xe303934bU,
720 	0x30fa5520U, 0x766df6adU, 0xcc769188U, 0x024c25f5U,
721 	0xe5d7fc4fU, 0x2acbd7c5U, 0x35448026U, 0x62a38fb5U,
722 	0xb15a49deU, 0xba1b6725U, 0xea0e9845U, 0xfec0e15dU,
723 	0x2f7502c3U, 0x4cf01281U, 0x4697a38dU, 0xd3f9c66bU,
724 	0x8f5fe703U, 0x929c9515U, 0x6d7aebbfU, 0x5259da95U,
725 	0xbe832dd4U, 0x7421d358U, 0xe0692949U, 0xc9c8448eU,
726 	0xc2896a75U, 0x8e7978f4U, 0x583e6b99U, 0xb971dd27U,
727 	0xe14fb6beU, 0x88ad17f0U, 0x20ac66c9U, 0xce3ab47dU,
728 	0xdf4a1863U, 0x1a3182e5U, 0x51336097U, 0x537f4562U,
729 	0x6477e0b1U, 0x6bae84bbU, 0x81a01cfeU, 0x082b94f9U,
730 	0x48685870U, 0x45fd198fU, 0xde6c8794U, 0x7bf8b752U,
731 	0x73d323abU, 0x4b02e272U, 0x1f8f57e3U, 0x55ab2a66U,
732 	0xeb2807b2U, 0xb5c2032fU, 0xc57b9a86U, 0x3708a5d3U,
733 	0x2887f230U, 0xbfa5b223U, 0x036aba02U, 0x16825cedU,
734 	0xcf1c2b8aU, 0x79b492a7U, 0x07f2f0f3U, 0x69e2a14eU,
735 	0xdaf4cd65U, 0x05bed506U, 0x34621fd1U, 0xa6fe8ac4U,
736 	0x2e539d34U, 0xf355a0a2U, 0x8ae13205U, 0xf6eb75a4U,
737 	0x83ec390bU, 0x60efaa40U, 0x719f065eU, 0x6e1051bdU,
738 	0x218af93eU, 0xdd063d96U, 0x3e05aeddU, 0xe6bd464dU,
739 	0x548db591U, 0xc45d0571U, 0x06d46f04U, 0x5015ff60U,
740 	0x98fb2419U, 0xbde997d6U, 0x4043cc89U, 0xd99e7767U,
741 	0xe842bdb0U, 0x898b8807U, 0x195b38e7U, 0xc8eedb79U,
742 	0x7c0a47a1U, 0x420fe97cU, 0x841ec9f8U, 0x00000000U,
743 	0x80868309U, 0x2bed4832U, 0x1170ac1eU, 0x5a724e6cU,
744 	0x0efffbfdU, 0x8538560fU, 0xaed51e3dU, 0x2d392736U,
745 	0x0fd9640aU, 0x5ca62168U, 0x5b54d19bU, 0x362e3a24U,
746 	0x0a67b10cU, 0x57e70f93U, 0xee96d2b4U, 0x9b919e1bU,
747 	0xc0c54f80U, 0xdc20a261U, 0x774b695aU, 0x121a161cU,
748 	0x93ba0ae2U, 0xa02ae5c0U, 0x22e0433cU, 0x1b171d12U,
749 	0x090d0b0eU, 0x8bc7adf2U, 0xb6a8b92dU, 0x1ea9c814U,
750 	0xf1198557U, 0x75074cafU, 0x99ddbbeeU, 0x7f60fda3U,
751 	0x01269ff7U, 0x72f5bc5cU, 0x663bc544U, 0xfb7e345bU,
752 	0x4329768bU, 0x23c6dccbU, 0xedfc68b6U, 0xe4f163b8U,
753 	0x31dccad7U, 0x63851042U, 0x97224013U, 0xc6112084U,
754 	0x4a247d85U, 0xbb3df8d2U, 0xf93211aeU, 0x29a16dc7U,
755 	0x9e2f4b1dU, 0xb230f3dcU, 0x8652ec0dU, 0xc1e3d077U,
756 	0xb3166c2bU, 0x70b999a9U, 0x9448fa11U, 0xe9642247U,
757 	0xfc8cc4a8U, 0xf03f1aa0U, 0x7d2cd856U, 0x3390ef22U,
758 	0x494ec787U, 0x38d1c1d9U, 0xcaa2fe8cU, 0xd40b3698U,
759 	0xf581cfa6U, 0x7ade28a5U, 0xb78e26daU, 0xadbfa43fU,
760 	0x3a9de42cU, 0x78920d50U, 0x5fcc9b6aU, 0x7e466254U,
761 	0x8d13c2f6U, 0xd8b8e890U, 0x39f75e2eU, 0xc3aff582U,
762 	0x5d80be9fU, 0xd0937c69U, 0xd52da96fU, 0x2512b3cfU,
763 	0xac993bc8U, 0x187da710U, 0x9c636ee8U, 0x3bbb7bdbU,
764 	0x267809cdU, 0x5918f46eU, 0x9ab701ecU, 0x4f9aa883U,
765 	0x956e65e6U, 0xffe67eaaU, 0xbccf0821U, 0x15e8e6efU,
766 	0xe79bd9baU, 0x6f36ce4aU, 0x9f09d4eaU, 0xb07cd629U,
767 	0xa4b2af31U, 0x3f23312aU, 0xa59430c6U, 0xa266c035U,
768 	0x4ebc3774U, 0x82caa6fcU, 0x90d0b0e0U, 0xa7d81533U,
769 	0x04984af1U, 0xecdaf741U, 0xcd500e7fU, 0x91f62f17U,
770 	0x4dd68d76U, 0xefb04d43U, 0xaa4d54ccU, 0x9604dfe4U,
771 	0xd1b5e39eU, 0x6a881b4cU, 0x2c1fb8c1U, 0x65517f46U,
772 	0x5eea049dU, 0x8c355d01U, 0x877473faU, 0x0b412efbU,
773 	0x671d5ab3U, 0xdbd25292U, 0x105633e9U, 0xd647136dU,
774 	0xd7618c9aU, 0xa10c7a37U, 0xf8148e59U, 0x133c89ebU,
775 	0xa927eeceU, 0x61c935b7U, 0x1ce5ede1U, 0x47b13c7aU,
776 	0xd2df599cU, 0xf2733f55U, 0x14ce7918U, 0xc737bf73U,
777 	0xf7cdea53U, 0xfdaa5b5fU, 0x3d6f14dfU, 0x44db8678U,
778 	0xaff381caU, 0x68c43eb9U, 0x24342c38U, 0xa3405fc2U,
779 	0x1dc37216U, 0xe2250cbcU, 0x3c498b28U, 0x0d9541ffU,
780 	0xa8017139U, 0x0cb3de08U, 0xb4e49cd8U, 0x56c19064U,
781 	0xcb84617bU, 0x32b670d5U, 0x6c5c7448U, 0xb85742d0U
782 };
783 
784 #ifndef sun4u
785 
786 static const uint32_t Td4[256] =
787 {
788 	0x52525252U, 0x09090909U, 0x6a6a6a6aU, 0xd5d5d5d5U,
789 	0x30303030U, 0x36363636U, 0xa5a5a5a5U, 0x38383838U,
790 	0xbfbfbfbfU, 0x40404040U, 0xa3a3a3a3U, 0x9e9e9e9eU,
791 	0x81818181U, 0xf3f3f3f3U, 0xd7d7d7d7U, 0xfbfbfbfbU,
792 	0x7c7c7c7cU, 0xe3e3e3e3U, 0x39393939U, 0x82828282U,
793 	0x9b9b9b9bU, 0x2f2f2f2fU, 0xffffffffU, 0x87878787U,
794 	0x34343434U, 0x8e8e8e8eU, 0x43434343U, 0x44444444U,
795 	0xc4c4c4c4U, 0xdedededeU, 0xe9e9e9e9U, 0xcbcbcbcbU,
796 	0x54545454U, 0x7b7b7b7bU, 0x94949494U, 0x32323232U,
797 	0xa6a6a6a6U, 0xc2c2c2c2U, 0x23232323U, 0x3d3d3d3dU,
798 	0xeeeeeeeeU, 0x4c4c4c4cU, 0x95959595U, 0x0b0b0b0bU,
799 	0x42424242U, 0xfafafafaU, 0xc3c3c3c3U, 0x4e4e4e4eU,
800 	0x08080808U, 0x2e2e2e2eU, 0xa1a1a1a1U, 0x66666666U,
801 	0x28282828U, 0xd9d9d9d9U, 0x24242424U, 0xb2b2b2b2U,
802 	0x76767676U, 0x5b5b5b5bU, 0xa2a2a2a2U, 0x49494949U,
803 	0x6d6d6d6dU, 0x8b8b8b8bU, 0xd1d1d1d1U, 0x25252525U,
804 	0x72727272U, 0xf8f8f8f8U, 0xf6f6f6f6U, 0x64646464U,
805 	0x86868686U, 0x68686868U, 0x98989898U, 0x16161616U,
806 	0xd4d4d4d4U, 0xa4a4a4a4U, 0x5c5c5c5cU, 0xccccccccU,
807 	0x5d5d5d5dU, 0x65656565U, 0xb6b6b6b6U, 0x92929292U,
808 	0x6c6c6c6cU, 0x70707070U, 0x48484848U, 0x50505050U,
809 	0xfdfdfdfdU, 0xededededU, 0xb9b9b9b9U, 0xdadadadaU,
810 	0x5e5e5e5eU, 0x15151515U, 0x46464646U, 0x57575757U,
811 	0xa7a7a7a7U, 0x8d8d8d8dU, 0x9d9d9d9dU, 0x84848484U,
812 	0x90909090U, 0xd8d8d8d8U, 0xababababU, 0x00000000U,
813 	0x8c8c8c8cU, 0xbcbcbcbcU, 0xd3d3d3d3U, 0x0a0a0a0aU,
814 	0xf7f7f7f7U, 0xe4e4e4e4U, 0x58585858U, 0x05050505U,
815 	0xb8b8b8b8U, 0xb3b3b3b3U, 0x45454545U, 0x06060606U,
816 	0xd0d0d0d0U, 0x2c2c2c2cU, 0x1e1e1e1eU, 0x8f8f8f8fU,
817 	0xcacacacaU, 0x3f3f3f3fU, 0x0f0f0f0fU, 0x02020202U,
818 	0xc1c1c1c1U, 0xafafafafU, 0xbdbdbdbdU, 0x03030303U,
819 	0x01010101U, 0x13131313U, 0x8a8a8a8aU, 0x6b6b6b6bU,
820 	0x3a3a3a3aU, 0x91919191U, 0x11111111U, 0x41414141U,
821 	0x4f4f4f4fU, 0x67676767U, 0xdcdcdcdcU, 0xeaeaeaeaU,
822 	0x97979797U, 0xf2f2f2f2U, 0xcfcfcfcfU, 0xcecececeU,
823 	0xf0f0f0f0U, 0xb4b4b4b4U, 0xe6e6e6e6U, 0x73737373U,
824 	0x96969696U, 0xacacacacU, 0x74747474U, 0x22222222U,
825 	0xe7e7e7e7U, 0xadadadadU, 0x35353535U, 0x85858585U,
826 	0xe2e2e2e2U, 0xf9f9f9f9U, 0x37373737U, 0xe8e8e8e8U,
827 	0x1c1c1c1cU, 0x75757575U, 0xdfdfdfdfU, 0x6e6e6e6eU,
828 	0x47474747U, 0xf1f1f1f1U, 0x1a1a1a1aU, 0x71717171U,
829 	0x1d1d1d1dU, 0x29292929U, 0xc5c5c5c5U, 0x89898989U,
830 	0x6f6f6f6fU, 0xb7b7b7b7U, 0x62626262U, 0x0e0e0e0eU,
831 	0xaaaaaaaaU, 0x18181818U, 0xbebebebeU, 0x1b1b1b1bU,
832 	0xfcfcfcfcU, 0x56565656U, 0x3e3e3e3eU, 0x4b4b4b4bU,
833 	0xc6c6c6c6U, 0xd2d2d2d2U, 0x79797979U, 0x20202020U,
834 	0x9a9a9a9aU, 0xdbdbdbdbU, 0xc0c0c0c0U, 0xfefefefeU,
835 	0x78787878U, 0xcdcdcdcdU, 0x5a5a5a5aU, 0xf4f4f4f4U,
836 	0x1f1f1f1fU, 0xddddddddU, 0xa8a8a8a8U, 0x33333333U,
837 	0x88888888U, 0x07070707U, 0xc7c7c7c7U, 0x31313131U,
838 	0xb1b1b1b1U, 0x12121212U, 0x10101010U, 0x59595959U,
839 	0x27272727U, 0x80808080U, 0xececececU, 0x5f5f5f5fU,
840 	0x60606060U, 0x51515151U, 0x7f7f7f7fU, 0xa9a9a9a9U,
841 	0x19191919U, 0xb5b5b5b5U, 0x4a4a4a4aU, 0x0d0d0d0dU,
842 	0x2d2d2d2dU, 0xe5e5e5e5U, 0x7a7a7a7aU, 0x9f9f9f9fU,
843 	0x93939393U, 0xc9c9c9c9U, 0x9c9c9c9cU, 0xefefefefU,
844 	0xa0a0a0a0U, 0xe0e0e0e0U, 0x3b3b3b3bU, 0x4d4d4d4dU,
845 	0xaeaeaeaeU, 0x2a2a2a2aU, 0xf5f5f5f5U, 0xb0b0b0b0U,
846 	0xc8c8c8c8U, 0xebebebebU, 0xbbbbbbbbU, 0x3c3c3c3cU,
847 	0x83838383U, 0x53535353U, 0x99999999U, 0x61616161U,
848 	0x17171717U, 0x2b2b2b2bU, 0x04040404U, 0x7e7e7e7eU,
849 	0xbabababaU, 0x77777777U, 0xd6d6d6d6U, 0x26262626U,
850 	0xe1e1e1e1U, 0x69696969U, 0x14141414U, 0x63636363U,
851 	0x55555555U, 0x21212121U, 0x0c0c0c0cU, 0x7d7d7d7dU
852 };
853 
854 #endif /* !sun4u */
855 
856 /* Rcon is Round Constant; used for encryption key expansion */
857 static const uint32_t rcon[RC_LENGTH] =
858 {
859 	/* for 128-bit blocks, Rijndael never uses more than 10 rcon values */
860 	0x01000000, 0x02000000, 0x04000000, 0x08000000,
861 	0x10000000, 0x20000000, 0x40000000, 0x80000000,
862 	0x1B000000, 0x36000000
863 };
864 
865 
866 /*
867  * Expand the cipher key into the encryption key schedule.
868  *
869  * Return the number of rounds for the given cipher key size.
870  * The size of the key schedule depends on the number of rounds
871  * (which can be computed from the size of the key), i.e. 4*(Nr + 1).
872  *
873  * Parameters:
874  * rk		AES key schedule 32-bit array to be initialized
875  * cipherKey	User key
876  * keyBits	AES key size (128, 192, or 256 bits)
877  */
878 static int
879 rijndael_key_setup_enc_raw(uint32_t rk[], const uint32_t cipherKey[],
880     int keyBits)
881 {
882 	int		i = 0;
883 	uint32_t	temp;
884 
885 	rk[0] = cipherKey[0];
886 	rk[1] = cipherKey[1];
887 	rk[2] = cipherKey[2];
888 	rk[3] = cipherKey[3];
889 
890 	if (keyBits == 128) {
891 		for (;;) {
892 			temp  = rk[3];
893 			rk[4] = rk[0] ^
894 			    (Te4[(temp >> 16) & 0xff] & 0xff000000) ^
895 			    (Te4[(temp >> 8) & 0xff] & 0x00ff0000) ^
896 			    (Te4[temp & 0xff] & 0x0000ff00) ^
897 			    (Te4[temp >> 24] & 0x000000ff) ^
898 			    rcon[i];
899 			rk[5] = rk[1] ^ rk[4];
900 			rk[6] = rk[2] ^ rk[5];
901 			rk[7] = rk[3] ^ rk[6];
902 
903 			if (++i == 10) {
904 				return (10);
905 			}
906 			rk += 4;
907 		}
908 	}
909 
910 	rk[4] = cipherKey[4];
911 	rk[5] = cipherKey[5];
912 
913 	if (keyBits == 192) {
914 		for (;;) {
915 			temp = rk[5];
916 			rk[6] = rk[0] ^
917 			    (Te4[(temp >> 16) & 0xff] & 0xff000000) ^
918 			    (Te4[(temp >> 8) & 0xff] & 0x00ff0000) ^
919 			    (Te4[temp & 0xff] & 0x0000ff00) ^
920 			    (Te4[temp >> 24] & 0x000000ff) ^
921 			    rcon[i];
922 			rk[7] = rk[1] ^ rk[6];
923 			rk[8] = rk[2] ^ rk[7];
924 			rk[9] = rk[3] ^ rk[8];
925 
926 			if (++i == 8) {
927 				return (12);
928 			}
929 
930 			rk[10] = rk[4] ^ rk[9];
931 			rk[11] = rk[5] ^ rk[10];
932 			rk += 6;
933 		}
934 	}
935 
936 	rk[6] = cipherKey[6];
937 	rk[7] = cipherKey[7];
938 
939 	if (keyBits == 256) {
940 		for (;;) {
941 			temp = rk[7];
942 			rk[8] = rk[0] ^
943 			    (Te4[(temp >> 16) & 0xff] & 0xff000000) ^
944 			    (Te4[(temp >> 8) & 0xff] & 0x00ff0000) ^
945 			    (Te4[temp & 0xff] & 0x0000ff00) ^
946 			    (Te4[temp >> 24] & 0x000000ff) ^
947 			    rcon[i];
948 			rk[9] = rk[1] ^ rk[8];
949 			rk[10] = rk[2] ^ rk[9];
950 			rk[11] = rk[3] ^ rk[10];
951 
952 			if (++i == 7) {
953 				return (14);
954 			}
955 			temp = rk[11];
956 			rk[12] = rk[4] ^
957 			    (Te4[temp >> 24] & 0xff000000) ^
958 			    (Te4[(temp >> 16) & 0xff] & 0x00ff0000) ^
959 			    (Te4[(temp >> 8) & 0xff] & 0x0000ff00) ^
960 			    (Te4[temp & 0xff] & 0x000000ff);
961 			rk[13] = rk[5] ^ rk[12];
962 			rk[14] = rk[6] ^ rk[13];
963 			rk[15] = rk[7] ^ rk[14];
964 
965 			rk += 8;
966 		}
967 	}
968 
969 	return (0);
970 }
971 #endif	/* !__amd64 */
972 
973 
974 #ifdef	sun4u
975 
976 /*
977  * Expand the cipher key into the encryption key schedule.
978  * by the sun4u optimized assembly implementation.
979  *
980  * Return the number of rounds for the given cipher key size.
981  * The size of the key schedule depends on the number of rounds
982  * (which can be computed from the size of the key), i.e. 4*(Nr + 1).
983  *
984  * Parameters:
985  * rk		AES key schedule 64-bit array to be initialized
986  * cipherKey	User key
987  * keyBits	AES key size (128, 192, or 256 bits)
988  */
989 static int
990 rijndael_key_setup_enc(uint64_t rk[], const uint32_t cipherKey[], int keyBits)
991 {
992 	uint32_t	rk1[4 * (MAX_AES_NR + 1)];
993 	uint64_t	*rk64 = (uint64_t *)rk;
994 	uint32_t	*rkt;
995 	uint64_t	t;
996 	int		i, Nr;
997 
998 	Nr = rijndael_key_setup_enc_raw(rk1, cipherKey, keyBits);
999 
1000 	for (i = 0; i < 4 * Nr; i++) {
1001 		t = (uint64_t)(rk1[i]);
1002 		rk64[i] = ((t & 0xff000000) << 11) |
1003 		    ((t & 0xff0000) << 8) |
1004 		    ((t & 0xffff) << 3);
1005 	}
1006 
1007 	rkt = (uint32_t *)(&(rk64[4 * Nr]));
1008 
1009 	for (i = 0; i < 4; i++) {
1010 		rkt[i] = rk1[4 * Nr+i];
1011 	}
1012 
1013 	return (Nr);
1014 }
1015 
1016 
1017 /*
1018  * Expand the cipher key into the decryption key schedule as used
1019  * by the sun4u optimized assembly implementation.
1020  *
1021  * Return the number of rounds for the given cipher key size.
1022  * The size of the key schedule depends on the number of rounds
1023  * (which can be computed from the size of the key), i.e. 4*(Nr + 1).
1024  *
1025  * Parameters:
1026  * rk		AES key schedule 32-bit array to be initialized
1027  * cipherKey	User key
1028  * keyBits	AES key size (128, 192, or 256 bits)
1029  */
1030 static int
1031 rijndael_key_setup_dec_raw(uint32_t rk[], const uint32_t cipherKey[],
1032     int keyBits)
1033 {
1034 	int		Nr, i;
1035 	uint32_t	temp;
1036 
1037 	/* expand the cipher key: */
1038 	Nr = rijndael_key_setup_enc_raw(rk, cipherKey, keyBits);
1039 
1040 	/* invert the order of the round keys: */
1041 
1042 	for (i = 0; i < 2 * Nr + 2; i++) {
1043 		temp = rk[i];
1044 		rk[i] = rk[4 * Nr - i + 3];
1045 		rk[4 * Nr - i + 3] = temp;
1046 	}
1047 
1048 	/*
1049 	 * apply the inverse MixColumn transform to all
1050 	 * round keys but the first and the last:
1051 	 */
1052 	for (i = 1; i < Nr; i++) {
1053 		rk += 4;
1054 		rk[0] = Td0[Te4[rk[0] >> 24] & 0xff] ^
1055 		    Td1[Te4[(rk[0] >> 16) & 0xff] & 0xff] ^
1056 		    Td2[Te4[(rk[0] >>  8) & 0xff] & 0xff] ^
1057 		    Td3[Te4[rk[0] & 0xff] & 0xff];
1058 		rk[1] = Td0[Te4[rk[1] >> 24] & 0xff] ^
1059 		    Td1[Te4[(rk[1] >> 16) & 0xff] & 0xff] ^
1060 		    Td2[Te4[(rk[1] >> 8) & 0xff] & 0xff] ^
1061 		    Td3[Te4[rk[1] & 0xff] & 0xff];
1062 		rk[2] = Td0[Te4[rk[2] >> 24] & 0xff] ^
1063 		    Td1[Te4[(rk[2] >> 16) & 0xff] & 0xff] ^
1064 		    Td2[Te4[(rk[2] >> 8) & 0xff] & 0xff] ^
1065 		    Td3[Te4[rk[2] & 0xff] & 0xff];
1066 		rk[3] = Td0[Te4[rk[3] >> 24] & 0xff] ^
1067 		    Td1[Te4[(rk[3] >> 16) & 0xff] & 0xff] ^
1068 		    Td2[Te4[(rk[3] >> 8) & 0xff] & 0xff] ^
1069 		    Td3[Te4[rk[3] & 0xff] & 0xff];
1070 	}
1071 
1072 	return (Nr);
1073 }
1074 
1075 
1076 /*
1077  * The size of the key schedule depends on the number of rounds
1078  * (which can be computed from the size of the key), i.e. 4*(Nr + 1).
1079  *
1080  * Parameters:
1081  * rk		AES key schedule 64-bit array to be initialized
1082  * cipherKey	User key
1083  * keyBits	AES key size (128, 192, or 256 bits)
1084  */
1085 static int
1086 rijndael_key_setup_dec(uint64_t rk[], const uint32_t cipherKey[], int keyBits)
1087 {
1088 	uint32_t	rk1[4 * (MAX_AES_NR + 1)];
1089 	uint64_t	*rk64 = (uint64_t *)rk;
1090 	uint32_t	*rkt;
1091 	uint64_t	t;
1092 	int		i, Nr;
1093 
1094 	Nr = rijndael_key_setup_dec_raw(rk1, cipherKey, keyBits);
1095 	for (i = 0; i < 4 * Nr; i++) {
1096 		t = (uint64_t)(rk1[i]);
1097 		rk64[i] = ((t & 0xff000000) << 11) |
1098 		    ((t & 0xff0000) << 8) |
1099 		    ((t & 0xffff) << 3);
1100 	}
1101 
1102 	rkt = (uint32_t *)(&(rk64[4 * Nr]));
1103 
1104 	for (i = 0; i < 4; i++) {
1105 		rkt[i] = rk1[4 * Nr + i];
1106 	}
1107 
1108 	return (Nr);
1109 }
1110 
1111 
1112 /*
1113  * Expand the 64-bit AES cipher key array into the encryption and decryption
1114  * key schedules.
1115  *
1116  * Parameters:
1117  * key		AES key schedule to be initialized
1118  * keyarr32	User key
1119  * keyBits	AES key size (128, 192, or 256 bits)
1120  */
1121 static void
1122 aes_setupkeys(aes_key_t *key, const uint32_t *keyarr32, int keybits)
1123 {
1124 	key->nr = rijndael_key_setup_enc(&(key->encr_ks.ks64[0]), keyarr32,
1125 	    keybits);
1126 	key->nr = rijndael_key_setup_dec(&(key->decr_ks.ks64[0]), keyarr32,
1127 	    keybits);
1128 	key->type = AES_64BIT_KS;
1129 }
1130 
1131 
1132 #elif defined(__amd64)
1133 
1134 /*
1135  * Expand the 32-bit AES cipher key array into the encryption and decryption
1136  * key schedules.
1137  *
1138  * Parameters:
1139  * key		AES key schedule to be initialized
1140  * keyarr32	User key
1141  * keyBits	AES key size (128, 192, or 256 bits)
1142  */
1143 static void
1144 aes_setupkeys(aes_key_t *key, const uint32_t *keyarr32, int keybits)
1145 {
1146 	if (intel_aes_instructions_present()) {
1147 		key->flags = INTEL_AES_NI_CAPABLE;
1148 		KPREEMPT_DISABLE;
1149 		key->nr = rijndael_key_setup_enc_intel(&(key->encr_ks.ks32[0]),
1150 		    keyarr32, keybits);
1151 		key->nr = rijndael_key_setup_dec_intel(&(key->decr_ks.ks32[0]),
1152 		    keyarr32, keybits);
1153 		KPREEMPT_ENABLE;
1154 	} else {
1155 		key->flags = 0;
1156 		key->nr = rijndael_key_setup_enc_amd64(&(key->encr_ks.ks32[0]),
1157 		    keyarr32, keybits);
1158 		key->nr = rijndael_key_setup_dec_amd64(&(key->decr_ks.ks32[0]),
1159 		    keyarr32, keybits);
1160 	}
1161 
1162 	key->type = AES_32BIT_KS;
1163 }
1164 
1165 /*
1166  * Encrypt one block of data. The block is assumed to be an array
1167  * of four uint32_t values, so copy for alignment (and byte-order
1168  * reversal for little endian systems might be necessary on the
1169  * input and output byte streams.
1170  * The size of the key schedule depends on the number of rounds
1171  * (which can be computed from the size of the key), i.e. 4*(Nr + 1).
1172  *
1173  * Parameters:
1174  * rk		Key schedule, of aes_ks_t (60 32-bit integers)
1175  * Nr		Number of rounds
1176  * pt		Input block (plain text)
1177  * ct		Output block (crypto text).  Can overlap with pt
1178  * flags	Indicates whether we're on Intel AES-NI-capable hardware
1179  */
1180 static void
1181 rijndael_encrypt(const uint32_t rk[], int Nr, const uint32_t pt[4],
1182     uint32_t ct[4], int flags) {
1183 	if (flags & INTEL_AES_NI_CAPABLE) {
1184 		KPREEMPT_DISABLE;
1185 		aes_encrypt_intel(rk, Nr, pt, ct);
1186 		KPREEMPT_ENABLE;
1187 	} else {
1188 		aes_encrypt_amd64(rk, Nr, pt, ct);
1189 	}
1190 }
1191 
1192 /*
1193  * Decrypt one block of data. The block is assumed to be an array
1194  * of four uint32_t values, so copy for alignment (and byte-order
1195  * reversal for little endian systems might be necessary on the
1196  * input and output byte streams.
1197  * The size of the key schedule depends on the number of rounds
1198  * (which can be computed from the size of the key), i.e. 4*(Nr + 1).
1199  *
1200  * Parameters:
1201  * rk		Key schedule, of aes_ks_t (60 32-bit integers)
1202  * Nr		Number of rounds
1203  * ct		Input block (crypto text)
1204  * pt		Output block (plain text). Can overlap with pt
1205  * flags	Indicates whether we're on Intel AES-NI-capable hardware
1206  */
1207 static void
1208 rijndael_decrypt(const uint32_t rk[], int Nr, const uint32_t ct[4],
1209     uint32_t pt[4], int flags) {
1210 	if (flags & INTEL_AES_NI_CAPABLE) {
1211 		KPREEMPT_DISABLE;
1212 		aes_decrypt_intel(rk, Nr, ct, pt);
1213 		KPREEMPT_ENABLE;
1214 	} else {
1215 		aes_decrypt_amd64(rk, Nr, ct, pt);
1216 	}
1217 }
1218 
1219 
1220 #else /* generic C implementation */
1221 
1222 /*
1223  *  Expand the cipher key into the decryption key schedule.
1224  *  Return the number of rounds for the given cipher key size.
1225  *  The size of the key schedule depends on the number of rounds
1226  *  (which can be computed from the size of the key), i.e. 4*(Nr + 1).
1227  *
1228  * Parameters:
1229  * rk		AES key schedule 32-bit array to be initialized
1230  * cipherKey	User key
1231  * keyBits	AES key size (128, 192, or 256 bits)
1232  */
1233 static int
1234 rijndael_key_setup_dec(uint32_t rk[], const uint32_t cipherKey[], int keyBits)
1235 {
1236 	int	 Nr, i, j;
1237 	uint32_t temp;
1238 
1239 	/* expand the cipher key: */
1240 	Nr = rijndael_key_setup_enc_raw(rk, cipherKey, keyBits);
1241 
1242 	/* invert the order of the round keys: */
1243 	for (i = 0, j = 4 * Nr; i < j; i += 4, j -= 4) {
1244 		temp = rk[i];
1245 		rk[i] = rk[j];
1246 		rk[j] = temp;
1247 		temp = rk[i + 1];
1248 		rk[i + 1] = rk[j + 1];
1249 		rk[j + 1] = temp;
1250 		temp = rk[i + 2];
1251 		rk[i + 2] = rk[j + 2];
1252 		rk[j + 2] = temp;
1253 		temp = rk[i + 3];
1254 		rk[i + 3] = rk[j + 3];
1255 		rk[j + 3] = temp;
1256 	}
1257 
1258 	/*
1259 	 * apply the inverse MixColumn transform to all
1260 	 * round keys but the first and the last:
1261 	 */
1262 	for (i = 1; i < Nr; i++) {
1263 		rk += 4;
1264 		rk[0] = Td0[Te4[rk[0] >> 24] & 0xff] ^
1265 		    Td1[Te4[(rk[0] >> 16) & 0xff] & 0xff] ^
1266 		    Td2[Te4[(rk[0] >> 8) & 0xff] & 0xff] ^
1267 		    Td3[Te4[rk[0] & 0xff] & 0xff];
1268 		rk[1] = Td0[Te4[rk[1] >> 24] & 0xff] ^
1269 		    Td1[Te4[(rk[1] >> 16) & 0xff] & 0xff] ^
1270 		    Td2[Te4[(rk[1] >> 8) & 0xff] & 0xff] ^
1271 		    Td3[Te4[rk[1] & 0xff] & 0xff];
1272 		rk[2] = Td0[Te4[rk[2] >> 24] & 0xff] ^
1273 		    Td1[Te4[(rk[2] >> 16) & 0xff] & 0xff] ^
1274 		    Td2[Te4[(rk[2] >> 8) & 0xff] & 0xff] ^
1275 		    Td3[Te4[rk[2] & 0xff] & 0xff];
1276 		rk[3] = Td0[Te4[rk[3] >> 24] & 0xff] ^
1277 		    Td1[Te4[(rk[3] >> 16) & 0xff] & 0xff] ^
1278 		    Td2[Te4[(rk[3] >> 8) & 0xff] & 0xff] ^
1279 		    Td3[Te4[rk[3] & 0xff] & 0xff];
1280 	}
1281 
1282 	return (Nr);
1283 }
1284 
1285 
1286 /*
1287  * Expand the 32-bit AES cipher key array into the encryption and decryption
1288  * key schedules.
1289  *
1290  * Parameters:
1291  * key		AES key schedule to be initialized
1292  * keyarr32	User key
1293  * keyBits	AES key size (128, 192, or 256 bits)
1294  */
1295 static void
1296 aes_setupkeys(aes_key_t *key, const uint32_t *keyarr32, int keybits)
1297 {
1298 	key->nr = rijndael_key_setup_enc(&(key->encr_ks.ks32[0]), keyarr32,
1299 	    keybits);
1300 	key->nr = rijndael_key_setup_dec(&(key->decr_ks.ks32[0]), keyarr32,
1301 	    keybits);
1302 	key->type = AES_32BIT_KS;
1303 }
1304 
1305 
1306 /*
1307  * Encrypt one block of data. The block is assumed to be an array
1308  * of four uint32_t values, so copy for alignment (and byte-order
1309  * reversal for little endian systems might be necessary on the
1310  * input and output byte streams.
1311  * The size of the key schedule depends on the number of rounds
1312  * (which can be computed from the size of the key), i.e. 4*(Nr + 1).
1313  *
1314  * Parameters:
1315  * rk	Key schedule, of aes_ks_t (60 32-bit integers)
1316  * Nr	Number of rounds
1317  * pt	Input block (plain text)
1318  * ct	Output block (crypto text).  Can overlap with pt
1319  */
1320 static void
1321 rijndael_encrypt(const uint32_t rk[], int Nr, const uint32_t pt[4],
1322     uint32_t ct[4])
1323 {
1324 	uint32_t	s0, s1, s2, s3, t0, t1, t2, t3;
1325 	int		r;
1326 
1327 	/*
1328 	 * map byte array block to cipher state
1329 	 * and add initial round key:
1330 	 */
1331 
1332 	s0 = pt[0] ^ rk[0];
1333 	s1 = pt[1] ^ rk[1];
1334 	s2 = pt[2] ^ rk[2];
1335 	s3 = pt[3] ^ rk[3];
1336 
1337 	/*
1338 	 * Nr - 1 full rounds:
1339 	 */
1340 
1341 	r = Nr >> 1;
1342 
1343 	for (;;) {
1344 		t0 = Te0[s0 >> 24] ^
1345 		    Te1[(s1 >> 16) & 0xff] ^
1346 		    Te2[(s2 >>  8) & 0xff] ^
1347 		    Te3[s3 & 0xff] ^
1348 		    rk[4];
1349 
1350 		t1 = Te0[s1 >> 24] ^
1351 		    Te1[(s2 >> 16) & 0xff] ^
1352 		    Te2[(s3 >>  8) & 0xff] ^
1353 		    Te3[s0 & 0xff] ^
1354 		    rk[5];
1355 
1356 		t2 = Te0[s2 >> 24] ^
1357 		    Te1[(s3 >> 16) & 0xff] ^
1358 		    Te2[(s0 >>  8) & 0xff] ^
1359 		    Te3[s1 & 0xff] ^
1360 		    rk[6];
1361 
1362 		t3 = Te0[s3 >> 24] ^
1363 		    Te1[(s0 >> 16) & 0xff] ^
1364 		    Te2[(s1 >>  8) & 0xff] ^
1365 		    Te3[s2 & 0xff] ^
1366 		    rk[7];
1367 
1368 		rk += 8;
1369 
1370 		if (--r == 0) {
1371 			break;
1372 		}
1373 
1374 		s0 = Te0[t0 >> 24] ^
1375 		    Te1[(t1 >> 16) & 0xff] ^
1376 		    Te2[(t2 >>  8) & 0xff] ^
1377 		    Te3[t3 & 0xff] ^
1378 		    rk[0];
1379 
1380 		s1 = Te0[t1 >> 24] ^
1381 		    Te1[(t2 >> 16) & 0xff] ^
1382 		    Te2[(t3 >>  8) & 0xff] ^
1383 		    Te3[t0 & 0xff] ^
1384 		    rk[1];
1385 
1386 		s2 = Te0[t2 >> 24] ^
1387 		    Te1[(t3 >> 16) & 0xff] ^
1388 		    Te2[(t0 >>  8) & 0xff] ^
1389 		    Te3[t1 & 0xff] ^
1390 		    rk[2];
1391 
1392 		s3 = Te0[t3 >> 24] ^
1393 		    Te1[(t0 >> 16) & 0xff] ^
1394 		    Te2[(t1 >>  8) & 0xff] ^
1395 		    Te3[t2 & 0xff] ^
1396 		    rk[3];
1397 	}
1398 
1399 	/*
1400 	 * apply last round and
1401 	 * map cipher state to byte array block:
1402 	 */
1403 
1404 	s0 = (Te4[(t0 >> 24)] & 0xff000000) ^
1405 	    (Te4[(t1 >> 16) & 0xff] & 0x00ff0000) ^
1406 	    (Te4[(t2 >>  8) & 0xff] & 0x0000ff00) ^
1407 	    (Te4[t3 & 0xff] & 0x000000ff) ^
1408 	    rk[0];
1409 	ct[0] = s0;
1410 
1411 	s1 = (Te4[(t1 >> 24)] & 0xff000000) ^
1412 	    (Te4[(t2 >> 16) & 0xff] & 0x00ff0000) ^
1413 	    (Te4[(t3 >>  8) & 0xff] & 0x0000ff00) ^
1414 	    (Te4[t0 & 0xff] & 0x000000ff) ^
1415 	    rk[1];
1416 	ct[1] = s1;
1417 
1418 	s2 = (Te4[(t2 >> 24)] & 0xff000000) ^
1419 	    (Te4[(t3 >> 16) & 0xff] & 0x00ff0000) ^
1420 	    (Te4[(t0 >>  8) & 0xff] & 0x0000ff00) ^
1421 	    (Te4[t1 & 0xff] & 0x000000ff) ^
1422 	    rk[2];
1423 	ct[2] = s2;
1424 
1425 	s3 = (Te4[(t3 >> 24)] & 0xff000000) ^
1426 	    (Te4[(t0 >> 16) & 0xff] & 0x00ff0000) ^
1427 	    (Te4[(t1 >>  8) & 0xff] & 0x0000ff00) ^
1428 	    (Te4[t2 & 0xff] & 0x000000ff) ^
1429 	    rk[3];
1430 	ct[3] = s3;
1431 }
1432 
1433 
1434 /*
1435  * Decrypt one block of data. The block is assumed to be an array
1436  * of four uint32_t values, so copy for alignment (and byte-order
1437  * reversal for little endian systems might be necessary on the
1438  * input and output byte streams.
1439  * The size of the key schedule depends on the number of rounds
1440  * (which can be computed from the size of the key), i.e. 4*(Nr + 1).
1441  *
1442  * Parameters:
1443  * rk	Key schedule, of aes_ks_t (60 32-bit integers)
1444  * Nr	Number of rounds
1445  * ct	Input block (crypto text)
1446  * pt	Output block (plain text). Can overlap with pt
1447  */
1448 static void
1449 rijndael_decrypt(const uint32_t rk[], int Nr, const uint32_t ct[4],
1450     uint32_t pt[4])
1451 {
1452 	uint32_t s0, s1, s2, s3, t0, t1, t2, t3;
1453 	int	 r;
1454 
1455 	/*
1456 	 * map byte array block to cipher state
1457 	 * and add initial round key:
1458 	 */
1459 	s0 = ct[0] ^ rk[0];
1460 	s1 = ct[1] ^ rk[1];
1461 	s2 = ct[2] ^ rk[2];
1462 	s3 = ct[3] ^ rk[3];
1463 
1464 	/*
1465 	 * Nr - 1 full rounds:
1466 	 */
1467 
1468 	r = Nr >> 1;
1469 
1470 	for (;;) {
1471 		t0 = Td0[s0 >> 24] ^
1472 		    Td1[(s3 >> 16) & 0xff] ^
1473 		    Td2[(s2 >> 8) & 0xff] ^
1474 		    Td3[s1 & 0xff] ^
1475 		    rk[4];
1476 
1477 		t1 = Td0[s1 >> 24] ^
1478 		    Td1[(s0 >> 16) & 0xff] ^
1479 		    Td2[(s3 >>  8) & 0xff] ^
1480 		    Td3[s2 & 0xff] ^
1481 		    rk[5];
1482 
1483 		t2 = Td0[s2 >> 24] ^
1484 		    Td1[(s1 >> 16) & 0xff] ^
1485 		    Td2[(s0 >>  8) & 0xff] ^
1486 		    Td3[s3 & 0xff] ^
1487 		    rk[6];
1488 
1489 		t3 = Td0[s3 >> 24] ^
1490 		    Td1[(s2 >> 16) & 0xff] ^
1491 		    Td2[(s1 >> 8) & 0xff] ^
1492 		    Td3[s0 & 0xff] ^
1493 		    rk[7];
1494 
1495 		rk += 8;
1496 
1497 		if (--r == 0) {
1498 			break;
1499 		}
1500 
1501 		s0 = Td0[t0 >> 24] ^
1502 		    Td1[(t3 >> 16) & 0xff] ^
1503 		    Td2[(t2 >> 8) & 0xff] ^
1504 		    Td3[t1 & 0xff] ^
1505 		    rk[0];
1506 
1507 		s1 = Td0[t1 >> 24] ^
1508 		    Td1[(t0 >> 16) & 0xff] ^
1509 		    Td2[(t3 >> 8) & 0xff] ^
1510 		    Td3[t2 & 0xff] ^
1511 		    rk[1];
1512 
1513 		s2 = Td0[t2 >> 24] ^
1514 		    Td1[(t1 >> 16) & 0xff] ^
1515 		    Td2[(t0 >> 8) & 0xff] ^
1516 		    Td3[t3 & 0xff] ^
1517 		    rk[2];
1518 
1519 		s3 = Td0[t3 >> 24] ^
1520 		    Td1[(t2 >> 16) & 0xff] ^
1521 		    Td2[(t1 >> 8) & 0xff] ^
1522 		    Td3[t0 & 0xff] ^
1523 		    rk[3];
1524 	}
1525 
1526 	/*
1527 	 * apply last round and
1528 	 * map cipher state to byte array block:
1529 	 */
1530 
1531 	s0 = (Td4[t0 >> 24] & 0xff000000) ^
1532 	    (Td4[(t3 >> 16) & 0xff] & 0x00ff0000) ^
1533 	    (Td4[(t2 >> 8) & 0xff] & 0x0000ff00) ^
1534 	    (Td4[t1 & 0xff] & 0x000000ff) ^
1535 	    rk[0];
1536 	pt[0] = s0;
1537 
1538 	s1 = (Td4[t1 >> 24] & 0xff000000) ^
1539 	    (Td4[(t0 >> 16) & 0xff] & 0x00ff0000) ^
1540 	    (Td4[(t3 >>  8) & 0xff] & 0x0000ff00) ^
1541 	    (Td4[t2 & 0xff] & 0x000000ff) ^
1542 	    rk[1];
1543 	pt[1] = s1;
1544 
1545 	s2 = (Td4[t2 >> 24] & 0xff000000) ^
1546 	    (Td4[(t1 >> 16) & 0xff] & 0x00ff0000) ^
1547 	    (Td4[(t0 >> 8) & 0xff] & 0x0000ff00) ^
1548 	    (Td4[t3 & 0xff] & 0x000000ff) ^
1549 	    rk[2];
1550 	pt[2] = s2;
1551 
1552 	s3 = (Td4[t3 >> 24] & 0xff000000) ^
1553 	    (Td4[(t2 >> 16) & 0xff] & 0x00ff0000) ^
1554 	    (Td4[(t1 >>  8) & 0xff] & 0x0000ff00) ^
1555 	    (Td4[t0 & 0xff] & 0x000000ff) ^
1556 	    rk[3];
1557 	pt[3] = s3;
1558 }
1559 #endif	/* sun4u, __amd64 */
1560 /* EXPORT DELETE END */
1561 
1562 
1563 /*
1564  * Initialize AES encryption and decryption key schedules.
1565  *
1566  * Parameters:
1567  * cipherKey	User key
1568  * keyBits	AES key size (128, 192, or 256 bits)
1569  * keysched	AES key schedule to be initialized, of type aes_key_t.
1570  *		Allocated by aes_alloc_keysched().
1571  */
1572 void
1573 aes_init_keysched(const uint8_t *cipherKey, uint_t keyBits, void *keysched)
1574 {
1575 /* EXPORT DELETE START */
1576 	aes_key_t	*newbie = keysched;
1577 	uint_t		keysize, i, j;
1578 	union {
1579 		uint64_t	ka64[4];
1580 		uint32_t	ka32[8];
1581 		} keyarr;
1582 
1583 	switch (keyBits) {
1584 	case 128:
1585 		newbie->nr = 10;
1586 		break;
1587 
1588 	case 192:
1589 		newbie->nr = 12;
1590 		break;
1591 
1592 	case 256:
1593 		newbie->nr = 14;
1594 		break;
1595 
1596 	default:
1597 		/* should never get here */
1598 		return;
1599 	}
1600 	keysize = CRYPTO_BITS2BYTES(keyBits);
1601 
1602 	/*
1603 	 * For _LITTLE_ENDIAN machines (except AMD64), reverse every
1604 	 * 4 bytes in the key.  On _BIG_ENDIAN and AMD64, copy the key
1605 	 * without reversing bytes.
1606 	 * For AMD64, do not byte swap for aes_setupkeys().
1607 	 *
1608 	 * SPARCv8/v9 uses a key schedule array with 64-bit elements.
1609 	 * X86/AMD64  uses a key schedule array with 32-bit elements.
1610 	 */
1611 #ifndef	AES_BYTE_SWAP
1612 	if (IS_P2ALIGNED(cipherKey, sizeof (uint64_t))) {
1613 		for (i = 0, j = 0; j < keysize; i++, j += 8) {
1614 			/* LINTED: pointer alignment */
1615 			keyarr.ka64[i] = *((uint64_t *)&cipherKey[j]);
1616 		}
1617 	} else {
1618 		bcopy(cipherKey, keyarr.ka32, keysize);
1619 	}
1620 
1621 #else	/* byte swap */
1622 	for (i = 0, j = 0; j < keysize; i++, j += 4) {
1623 		keyarr.ka32[i] = htonl(*(uint32_t *)(void *)&cipherKey[j]);
1624 	}
1625 #endif
1626 
1627 	aes_setupkeys(newbie, keyarr.ka32, keyBits);
1628 /* EXPORT DELETE END */
1629 }
1630 
1631 
1632 /*
1633  * Encrypt one block using AES.
1634  * Align if needed and (for x86 32-bit only) byte-swap.
1635  *
1636  * Parameters:
1637  * ks	Key schedule, of type aes_key_t
1638  * pt	Input block (plain text)
1639  * ct	Output block (crypto text).  Can overlap with pt
1640  */
1641 int
1642 aes_encrypt_block(const void *ks, const uint8_t *pt, uint8_t *ct)
1643 {
1644 /* EXPORT DELETE START */
1645 	aes_key_t	*ksch = (aes_key_t *)ks;
1646 
1647 #ifndef	AES_BYTE_SWAP
1648 	if (IS_P2ALIGNED2(pt, ct, sizeof (uint32_t))) {
1649 		/* LINTED:  pointer alignment */
1650 		AES_ENCRYPT_IMPL(&ksch->encr_ks.ks32[0], ksch->nr,
1651 		    /* LINTED:  pointer alignment */
1652 		    (uint32_t *)pt, (uint32_t *)ct, ksch->flags);
1653 	} else {
1654 #endif
1655 		uint32_t buffer[AES_BLOCK_LEN / sizeof (uint32_t)];
1656 
1657 		/* Copy input block into buffer */
1658 #ifndef	AES_BYTE_SWAP
1659 		bcopy(pt, &buffer, AES_BLOCK_LEN);
1660 
1661 #else	/* byte swap */
1662 		buffer[0] = htonl(*(uint32_t *)(void *)&pt[0]);
1663 		buffer[1] = htonl(*(uint32_t *)(void *)&pt[4]);
1664 		buffer[2] = htonl(*(uint32_t *)(void *)&pt[8]);
1665 		buffer[3] = htonl(*(uint32_t *)(void *)&pt[12]);
1666 #endif
1667 
1668 		AES_ENCRYPT_IMPL(&ksch->encr_ks.ks32[0], ksch->nr,
1669 		    buffer, buffer, ksch->flags);
1670 
1671 		/* Copy result from buffer to output block */
1672 #ifndef	AES_BYTE_SWAP
1673 		bcopy(&buffer, ct, AES_BLOCK_LEN);
1674 	}
1675 
1676 #else	/* byte swap */
1677 		*(uint32_t *)(void *)&ct[0] = htonl(buffer[0]);
1678 		*(uint32_t *)(void *)&ct[4] = htonl(buffer[1]);
1679 		*(uint32_t *)(void *)&ct[8] = htonl(buffer[2]);
1680 		*(uint32_t *)(void *)&ct[12] = htonl(buffer[3]);
1681 #endif
1682 /* EXPORT DELETE END */
1683 	return (CRYPTO_SUCCESS);
1684 }
1685 
1686 
1687 /*
1688  * Decrypt one block using AES.
1689  * Align and byte-swap if needed.
1690  *
1691  * Parameters:
1692  * ks	Key schedule, of type aes_key_t
1693  * ct	Input block (crypto text)
1694  * pt	Output block (plain text). Can overlap with pt
1695  */
1696 int
1697 aes_decrypt_block(const void *ks, const uint8_t *ct, uint8_t *pt)
1698 {
1699 /* EXPORT DELETE START */
1700 	aes_key_t	*ksch = (aes_key_t *)ks;
1701 
1702 #ifndef	AES_BYTE_SWAP
1703 	if (IS_P2ALIGNED2(ct, pt, sizeof (uint32_t))) {
1704 		/* LINTED:  pointer alignment */
1705 		AES_DECRYPT_IMPL(&ksch->decr_ks.ks32[0], ksch->nr,
1706 		    /* LINTED:  pointer alignment */
1707 		    (uint32_t *)ct, (uint32_t *)pt, ksch->flags);
1708 	} else {
1709 #endif
1710 		uint32_t buffer[AES_BLOCK_LEN / sizeof (uint32_t)];
1711 
1712 		/* Copy input block into buffer */
1713 #ifndef	AES_BYTE_SWAP
1714 		bcopy(ct, &buffer, AES_BLOCK_LEN);
1715 
1716 #else	/* byte swap */
1717 		buffer[0] = htonl(*(uint32_t *)(void *)&ct[0]);
1718 		buffer[1] = htonl(*(uint32_t *)(void *)&ct[4]);
1719 		buffer[2] = htonl(*(uint32_t *)(void *)&ct[8]);
1720 		buffer[3] = htonl(*(uint32_t *)(void *)&ct[12]);
1721 #endif
1722 
1723 		AES_DECRYPT_IMPL(&ksch->decr_ks.ks32[0], ksch->nr,
1724 		    buffer, buffer, ksch->flags);
1725 
1726 		/* Copy result from buffer to output block */
1727 #ifndef	AES_BYTE_SWAP
1728 		bcopy(&buffer, pt, AES_BLOCK_LEN);
1729 	}
1730 
1731 #else	/* byte swap */
1732 	*(uint32_t *)(void *)&pt[0] = htonl(buffer[0]);
1733 	*(uint32_t *)(void *)&pt[4] = htonl(buffer[1]);
1734 	*(uint32_t *)(void *)&pt[8] = htonl(buffer[2]);
1735 	*(uint32_t *)(void *)&pt[12] = htonl(buffer[3]);
1736 #endif
1737 
1738 /* EXPORT DELETE END */
1739 	return (CRYPTO_SUCCESS);
1740 }
1741 
1742 
1743 /*
1744  * Allocate key schedule for AES.
1745  *
1746  * Return the pointer and set size to the number of bytes allocated.
1747  * Memory allocated must be freed by the caller when done.
1748  *
1749  * Parameters:
1750  * size		Size of key schedule allocated, in bytes
1751  * kmflag	Flag passed to kmem_alloc(9F); ignored in userland.
1752  */
1753 /* ARGSUSED */
1754 void *
1755 aes_alloc_keysched(size_t *size, int kmflag)
1756 {
1757 /* EXPORT DELETE START */
1758 	aes_key_t *keysched;
1759 
1760 #ifdef	_KERNEL
1761 	keysched = (aes_key_t *)kmem_alloc(sizeof (aes_key_t), kmflag);
1762 #else	/* !_KERNEL */
1763 	keysched = (aes_key_t *)malloc(sizeof (aes_key_t));
1764 #endif	/* _KERNEL */
1765 
1766 	if (keysched != NULL) {
1767 		*size = sizeof (aes_key_t);
1768 		return (keysched);
1769 	}
1770 /* EXPORT DELETE END */
1771 	return (NULL);
1772 }
1773 
1774 
1775 #ifdef __amd64
1776 /*
1777  * Return 1 if executing on Intel with AES-NI instructions,
1778  * otherwise 0 (i.e., Intel without AES-NI or AMD64).
1779  * Cache the result, as the CPU can't change.
1780  *
1781  * Note: the userland version uses getisax().  The kernel version uses
1782  * global variable x86_feature.
1783  */
1784 static int
1785 intel_aes_instructions_present(void)
1786 {
1787 	static int	cached_result = -1;
1788 
1789 	if (cached_result == -1) { /* first time */
1790 #ifdef _KERNEL
1791 		cached_result = (x86_feature & X86_AES) != 0;
1792 #else
1793 		uint_t		ui = 0;
1794 
1795 		(void) getisax(&ui, 1);
1796 		cached_result = (ui & AV_386_AES) != 0;
1797 #endif	/* _KERNEL */
1798 	}
1799 
1800 	return (cached_result);
1801 }
1802 #endif	/* __amd64 */
1803