1/*-
2 * Copyright (C) 2001-2003 by NBMK Encryption Technologies.
3 * All rights reserved.
4 *
5 * NBMK Encryption Technologies provides no support of any kind for
6 * this software.  Questions or concerns about it may be addressed to
7 * the members of the relevant open-source community at
8 * <tech-crypto@netbsd.org>.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions are
12 * met:
13 *
14 * 1. Redistributions of source code must retain the above copyright
15 *    notice, this list of conditions and the following disclaimer.
16 *
17 * 2. Redistributions in binary form must reproduce the above
18 *    copyright notice, this list of conditions and the following
19 *    disclaimer in the documentation and/or other materials provided
20 *    with the distribution.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
23 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
24 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
25 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
26 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
27 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
28 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
29 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
30 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
31 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
32 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
33 */
34
35static char const n8_id[] = "$Id: n8_precomp_md5.c,v 1.1 2008/10/30 12:02:15 darran Exp $";
36/*****************************************************************************/
37/** @file n8_precomp_md5.c
38 *  @brief Private version of md5 for n8_precompute_*.
39 *
40 * A more detailed description of the file.
41 *
42 *****************************************************************************/
43
44/*****************************************************************************
45 * Revision history:
46 * 03/06/02 brr   Removed openssl includes.
47 * 02/18/02 spm   Removed #includes of usr headers.   Converted printf's to
48 *                DBG's.
49 * 01/22/02 dws   Original version.
50 ****************************************************************************/
51/** @defgroup subsystem_name Subsystem Title (not used for a header file)
52 */
53
54#define NO_ASM
55
56#include "n8_precomp_md5.h"
57#include "n8_precomp_md5_locl.h"
58
59#define INIT_DATA_A (unsigned long)0x67452301L
60#define INIT_DATA_B (unsigned long)0xefcdab89L
61#define INIT_DATA_C (unsigned long)0x98badcfeL
62#define INIT_DATA_D (unsigned long)0x10325476L
63
64void n8_precomp_MD5_Init(N8_PRECOMP_MD5_CTX *c)
65{
66   c->A=INIT_DATA_A;
67   c->B=INIT_DATA_B;
68   c->C=INIT_DATA_C;
69   c->D=INIT_DATA_D;
70   c->Nl=0;
71   c->Nh=0;
72   c->num=0;
73}
74
75/*************************************************/
76/* This stuff implements n8_precomp_MD5_Update() */
77/*************************************************/
78#ifndef md5_block_host_order
79void n8_precomp_md5_block_host_order (N8_PRECOMP_MD5_CTX *c, const void *data, int num)
80        {
81        const MD5_LONG *X=data;
82        register unsigned long A,B,C,D;
83        /*
84         * In case you wonder why A-D are declared as long and not
85         * as MD5_LONG. Doing so results in slight performance
86         * boost on LP64 architectures. The catch is we don't
87         * really care if 32 MSBs of a 64-bit register get polluted
88         * with eventual overflows as we *save* only 32 LSBs in
89         * *either* case. Now declaring 'em long excuses the compiler
90         * from keeping 32 MSBs zeroed resulting in 13% performance
91         * improvement under SPARC Solaris7/64 and 5% under AlphaLinux.
92         * Well, to be honest it should say that this *prevents*
93         * performance degradation.
94         *
95         *                              <appro@fy.chalmers.se>
96         */
97
98#ifdef HMAC_DEBUG
99        DBG(("*********************************************************\n"));
100        DBG(("Entering md5_block_host_order ***************************\n"));
101#endif
102
103        A=c->A;
104        B=c->B;
105        C=c->C;
106        D=c->D;
107
108        for (;num--;X+=HASH_LBLOCK)
109                {
110        /* Round 0 */
111        R0(A,B,C,D,X[ 0], 7,0xd76aa478L);
112        R0(D,A,B,C,X[ 1],12,0xe8c7b756L);
113        R0(C,D,A,B,X[ 2],17,0x242070dbL);
114        R0(B,C,D,A,X[ 3],22,0xc1bdceeeL);
115        R0(A,B,C,D,X[ 4], 7,0xf57c0fafL);
116        R0(D,A,B,C,X[ 5],12,0x4787c62aL);
117        R0(C,D,A,B,X[ 6],17,0xa8304613L);
118        R0(B,C,D,A,X[ 7],22,0xfd469501L);
119        R0(A,B,C,D,X[ 8], 7,0x698098d8L);
120        R0(D,A,B,C,X[ 9],12,0x8b44f7afL);
121        R0(C,D,A,B,X[10],17,0xffff5bb1L);
122        R0(B,C,D,A,X[11],22,0x895cd7beL);
123        R0(A,B,C,D,X[12], 7,0x6b901122L);
124        R0(D,A,B,C,X[13],12,0xfd987193L);
125        R0(C,D,A,B,X[14],17,0xa679438eL);
126        R0(B,C,D,A,X[15],22,0x49b40821L);
127        /* Round 1 */
128        R1(A,B,C,D,X[ 1], 5,0xf61e2562L);
129        R1(D,A,B,C,X[ 6], 9,0xc040b340L);
130        R1(C,D,A,B,X[11],14,0x265e5a51L);
131        R1(B,C,D,A,X[ 0],20,0xe9b6c7aaL);
132        R1(A,B,C,D,X[ 5], 5,0xd62f105dL);
133        R1(D,A,B,C,X[10], 9,0x02441453L);
134        R1(C,D,A,B,X[15],14,0xd8a1e681L);
135        R1(B,C,D,A,X[ 4],20,0xe7d3fbc8L);
136        R1(A,B,C,D,X[ 9], 5,0x21e1cde6L);
137        R1(D,A,B,C,X[14], 9,0xc33707d6L);
138        R1(C,D,A,B,X[ 3],14,0xf4d50d87L);
139        R1(B,C,D,A,X[ 8],20,0x455a14edL);
140        R1(A,B,C,D,X[13], 5,0xa9e3e905L);
141        R1(D,A,B,C,X[ 2], 9,0xfcefa3f8L);
142        R1(C,D,A,B,X[ 7],14,0x676f02d9L);
143        R1(B,C,D,A,X[12],20,0x8d2a4c8aL);
144        /* Round 2 */
145        R2(A,B,C,D,X[ 5], 4,0xfffa3942L);
146        R2(D,A,B,C,X[ 8],11,0x8771f681L);
147        R2(C,D,A,B,X[11],16,0x6d9d6122L);
148        R2(B,C,D,A,X[14],23,0xfde5380cL);
149        R2(A,B,C,D,X[ 1], 4,0xa4beea44L);
150        R2(D,A,B,C,X[ 4],11,0x4bdecfa9L);
151        R2(C,D,A,B,X[ 7],16,0xf6bb4b60L);
152        R2(B,C,D,A,X[10],23,0xbebfbc70L);
153        R2(A,B,C,D,X[13], 4,0x289b7ec6L);
154        R2(D,A,B,C,X[ 0],11,0xeaa127faL);
155        R2(C,D,A,B,X[ 3],16,0xd4ef3085L);
156        R2(B,C,D,A,X[ 6],23,0x04881d05L);
157        R2(A,B,C,D,X[ 9], 4,0xd9d4d039L);
158        R2(D,A,B,C,X[12],11,0xe6db99e5L);
159        R2(C,D,A,B,X[15],16,0x1fa27cf8L);
160        R2(B,C,D,A,X[ 2],23,0xc4ac5665L);
161        /* Round 3 */
162        R3(A,B,C,D,X[ 0], 6,0xf4292244L);
163        R3(D,A,B,C,X[ 7],10,0x432aff97L);
164        R3(C,D,A,B,X[14],15,0xab9423a7L);
165        R3(B,C,D,A,X[ 5],21,0xfc93a039L);
166        R3(A,B,C,D,X[12], 6,0x655b59c3L);
167        R3(D,A,B,C,X[ 3],10,0x8f0ccc92L);
168        R3(C,D,A,B,X[10],15,0xffeff47dL);
169        R3(B,C,D,A,X[ 1],21,0x85845dd1L);
170        R3(A,B,C,D,X[ 8], 6,0x6fa87e4fL);
171        R3(D,A,B,C,X[15],10,0xfe2ce6e0L);
172        R3(C,D,A,B,X[ 6],15,0xa3014314L);
173        R3(B,C,D,A,X[13],21,0x4e0811a1L);
174        R3(A,B,C,D,X[ 4], 6,0xf7537e82L);
175        R3(D,A,B,C,X[11],10,0xbd3af235L);
176        R3(C,D,A,B,X[ 2],15,0x2ad7d2bbL);
177        R3(B,C,D,A,X[ 9],21,0xeb86d391L);
178
179        A = c->A += A;
180        B = c->B += B;
181        C = c->C += C;
182        D = c->D += D;
183#ifdef HMAC_DEBUG
184        DBG(("Final      {A=%08lx B=%08lx C=%08lx D=%08lx}\n", A, B, C, D));
185#endif
186                }
187#ifdef HMAC_DEBUG
188        DBG(("Leaving md5_block_host_order ****************************\n"));
189        DBG(("*********************************************************\n"));
190#endif
191        }
192#endif
193
194#ifndef md5_block_data_order
195#ifdef X
196#undef X
197#endif
198void n8_precomp_md5_block_data_order (N8_PRECOMP_MD5_CTX *c, const void *data_, int num)
199        {
200        const unsigned char *data=data_;
201        register unsigned long A,B,C,D,l;
202        /*
203         * In case you wonder why A-D are declared as long and not
204         * as MD5_LONG. Doing so results in slight performance
205         * boost on LP64 architectures. The catch is we don't
206         * really care if 32 MSBs of a 64-bit register get polluted
207         * with eventual overflows as we *save* only 32 LSBs in
208         * *either* case. Now declaring 'em long excuses the compiler
209         * from keeping 32 MSBs zeroed resulting in 13% performance
210         * improvement under SPARC Solaris7/64 and 5% under AlphaLinux.
211         * Well, to be honest it should say that this *prevents*
212         * performance degradation.
213         *
214         *                              <appro@fy.chalmers.se>
215         */
216#ifndef MD32_XARRAY
217        /* See comment in crypto/sha/sha_locl.h for details. */
218        unsigned long   XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
219                        XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
220# define X(i)   XX##i
221#else
222        MD5_LONG XX[MD5_LBLOCK];
223# define X(i)   XX[i]
224#endif
225
226#ifdef HMAC_DEBUG
227        DBG(("*********************************************************\n"));
228        DBG(("Entering md5_block_data_order ***************************\n"));
229#endif
230
231        A=c->A;
232        B=c->B;
233        C=c->C;
234        D=c->D;
235
236        for (;num--;)
237                {
238        HOST_c2l(data,l); X( 0)=l;              HOST_c2l(data,l); X( 1)=l;
239        /* Round 0 */
240        R0(A,B,C,D,X( 0), 7,0xd76aa478L);       HOST_c2l(data,l); X( 2)=l;
241        R0(D,A,B,C,X( 1),12,0xe8c7b756L);       HOST_c2l(data,l); X( 3)=l;
242        R0(C,D,A,B,X( 2),17,0x242070dbL);       HOST_c2l(data,l); X( 4)=l;
243        R0(B,C,D,A,X( 3),22,0xc1bdceeeL);       HOST_c2l(data,l); X( 5)=l;
244        R0(A,B,C,D,X( 4), 7,0xf57c0fafL);       HOST_c2l(data,l); X( 6)=l;
245        R0(D,A,B,C,X( 5),12,0x4787c62aL);       HOST_c2l(data,l); X( 7)=l;
246        R0(C,D,A,B,X( 6),17,0xa8304613L);       HOST_c2l(data,l); X( 8)=l;
247        R0(B,C,D,A,X( 7),22,0xfd469501L);       HOST_c2l(data,l); X( 9)=l;
248        R0(A,B,C,D,X( 8), 7,0x698098d8L);       HOST_c2l(data,l); X(10)=l;
249        R0(D,A,B,C,X( 9),12,0x8b44f7afL);       HOST_c2l(data,l); X(11)=l;
250        R0(C,D,A,B,X(10),17,0xffff5bb1L);       HOST_c2l(data,l); X(12)=l;
251        R0(B,C,D,A,X(11),22,0x895cd7beL);       HOST_c2l(data,l); X(13)=l;
252        R0(A,B,C,D,X(12), 7,0x6b901122L);       HOST_c2l(data,l); X(14)=l;
253        R0(D,A,B,C,X(13),12,0xfd987193L);       HOST_c2l(data,l); X(15)=l;
254        R0(C,D,A,B,X(14),17,0xa679438eL);
255        R0(B,C,D,A,X(15),22,0x49b40821L);
256        /* Round 1 */
257        R1(A,B,C,D,X( 1), 5,0xf61e2562L);
258        R1(D,A,B,C,X( 6), 9,0xc040b340L);
259        R1(C,D,A,B,X(11),14,0x265e5a51L);
260        R1(B,C,D,A,X( 0),20,0xe9b6c7aaL);
261        R1(A,B,C,D,X( 5), 5,0xd62f105dL);
262        R1(D,A,B,C,X(10), 9,0x02441453L);
263        R1(C,D,A,B,X(15),14,0xd8a1e681L);
264        R1(B,C,D,A,X( 4),20,0xe7d3fbc8L);
265        R1(A,B,C,D,X( 9), 5,0x21e1cde6L);
266        R1(D,A,B,C,X(14), 9,0xc33707d6L);
267        R1(C,D,A,B,X( 3),14,0xf4d50d87L);
268        R1(B,C,D,A,X( 8),20,0x455a14edL);
269        R1(A,B,C,D,X(13), 5,0xa9e3e905L);
270        R1(D,A,B,C,X( 2), 9,0xfcefa3f8L);
271        R1(C,D,A,B,X( 7),14,0x676f02d9L);
272        R1(B,C,D,A,X(12),20,0x8d2a4c8aL);
273        /* Round 2 */
274        R2(A,B,C,D,X( 5), 4,0xfffa3942L);
275        R2(D,A,B,C,X( 8),11,0x8771f681L);
276        R2(C,D,A,B,X(11),16,0x6d9d6122L);
277        R2(B,C,D,A,X(14),23,0xfde5380cL);
278        R2(A,B,C,D,X( 1), 4,0xa4beea44L);
279        R2(D,A,B,C,X( 4),11,0x4bdecfa9L);
280        R2(C,D,A,B,X( 7),16,0xf6bb4b60L);
281        R2(B,C,D,A,X(10),23,0xbebfbc70L);
282        R2(A,B,C,D,X(13), 4,0x289b7ec6L);
283        R2(D,A,B,C,X( 0),11,0xeaa127faL);
284        R2(C,D,A,B,X( 3),16,0xd4ef3085L);
285        R2(B,C,D,A,X( 6),23,0x04881d05L);
286        R2(A,B,C,D,X( 9), 4,0xd9d4d039L);
287        R2(D,A,B,C,X(12),11,0xe6db99e5L);
288        R2(C,D,A,B,X(15),16,0x1fa27cf8L);
289        R2(B,C,D,A,X( 2),23,0xc4ac5665L);
290        /* Round 3 */
291        R3(A,B,C,D,X( 0), 6,0xf4292244L);
292        R3(D,A,B,C,X( 7),10,0x432aff97L);
293        R3(C,D,A,B,X(14),15,0xab9423a7L);
294        R3(B,C,D,A,X( 5),21,0xfc93a039L);
295        R3(A,B,C,D,X(12), 6,0x655b59c3L);
296        R3(D,A,B,C,X( 3),10,0x8f0ccc92L);
297        R3(C,D,A,B,X(10),15,0xffeff47dL);
298        R3(B,C,D,A,X( 1),21,0x85845dd1L);
299        R3(A,B,C,D,X( 8), 6,0x6fa87e4fL);
300        R3(D,A,B,C,X(15),10,0xfe2ce6e0L);
301        R3(C,D,A,B,X( 6),15,0xa3014314L);
302        R3(B,C,D,A,X(13),21,0x4e0811a1L);
303        R3(A,B,C,D,X( 4), 6,0xf7537e82L);
304        R3(D,A,B,C,X(11),10,0xbd3af235L);
305        R3(C,D,A,B,X( 2),15,0x2ad7d2bbL);
306        R3(B,C,D,A,X( 9),21,0xeb86d391L);
307
308        A = c->A += A;
309        B = c->B += B;
310        C = c->C += C;
311        D = c->D += D;
312#ifdef HMAC_DEBUG
313        DBG(("Final      {A=%08lx B=%08lx C=%08lx D=%08lx}\n", A, B, C, D));
314#endif
315                }
316#ifdef HMAC_DEBUG
317        DBG(("Leaving md5_block_data_order ****************************\n"));
318        DBG(("*********************************************************\n"));
319#endif
320        }
321#endif
322
323
324
325