1/*
2 * Intel SHA Extensions optimized implementation of a SHA-1 update function
3 *
4 * This file is provided under a dual BSD/GPLv2 license.  When using or
5 * redistributing this file, you may do so under either license.
6 *
7 * GPL LICENSE SUMMARY
8 *
9 * Copyright(c) 2015 Intel Corporation.
10 *
11 * This program is free software; you can redistribute it and/or modify
12 * it under the terms of version 2 of the GNU General Public License as
13 * published by the Free Software Foundation.
14 *
15 * This program is distributed in the hope that it will be useful, but
16 * WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
18 * General Public License for more details.
19 *
20 * Contact Information:
21 * 	Sean Gulley <sean.m.gulley@intel.com>
22 * 	Tim Chen <tim.c.chen@linux.intel.com>
23 *
24 * BSD LICENSE
25 *
26 * Copyright(c) 2015 Intel Corporation.
27 *
28 * Redistribution and use in source and binary forms, with or without
29 * modification, are permitted provided that the following conditions
30 * are met:
31 *
32 * 	* Redistributions of source code must retain the above copyright
33 * 	  notice, this list of conditions and the following disclaimer.
34 * 	* Redistributions in binary form must reproduce the above copyright
35 * 	  notice, this list of conditions and the following disclaimer in
36 * 	  the documentation and/or other materials provided with the
37 * 	  distribution.
38 * 	* Neither the name of Intel Corporation nor the names of its
39 * 	  contributors may be used to endorse or promote products derived
40 * 	  from this software without specific prior written permission.
41 *
42 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
43 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
44 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
45 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
46 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
47 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
48 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
49 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
50 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
51 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
52 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
53 *
54 */
55
56#include <linux/linkage.h>
57
58#define DIGEST_PTR	%rdi	/* 1st arg */
59#define DATA_PTR	%rsi	/* 2nd arg */
60#define NUM_BLKS	%rdx	/* 3rd arg */
61
62/* gcc conversion */
63#define FRAME_SIZE	32	/* space for 2x16 bytes */
64
65#define ABCD		%xmm0
66#define E0		%xmm1	/* Need two E's b/c they ping pong */
67#define E1		%xmm2
68#define MSG0		%xmm3
69#define MSG1		%xmm4
70#define MSG2		%xmm5
71#define MSG3		%xmm6
72#define SHUF_MASK	%xmm7
73
74
75/*
76 * Intel SHA Extensions optimized implementation of a SHA-1 update function
77 *
78 * The function takes a pointer to the current hash values, a pointer to the
79 * input data, and a number of 64 byte blocks to process.  Once all blocks have
80 * been processed, the digest pointer is  updated with the resulting hash value.
81 * The function only processes complete blocks, there is no functionality to
82 * store partial blocks. All message padding and hash value initialization must
83 * be done outside the update function.
84 *
85 * The indented lines in the loop are instructions related to rounds processing.
86 * The non-indented lines are instructions related to the message schedule.
87 *
88 * void sha1_ni_transform(uint32_t *digest, const void *data,
89		uint32_t numBlocks)
90 * digest : pointer to digest
91 * data: pointer to input data
92 * numBlocks: Number of blocks to process
93 */
94.text
95.align 32
96SYM_FUNC_START(sha1_ni_transform)
97	push		%rbp
98	mov		%rsp, %rbp
99	sub		$FRAME_SIZE, %rsp
100	and		$~0xF, %rsp
101
102	shl		$6, NUM_BLKS		/* convert to bytes */
103	jz		.Ldone_hash
104	add		DATA_PTR, NUM_BLKS	/* pointer to end of data */
105
106	/* load initial hash values */
107	pinsrd		$3, 1*16(DIGEST_PTR), E0
108	movdqu		0*16(DIGEST_PTR), ABCD
109	pand		UPPER_WORD_MASK(%rip), E0
110	pshufd		$0x1B, ABCD, ABCD
111
112	movdqa		PSHUFFLE_BYTE_FLIP_MASK(%rip), SHUF_MASK
113
114.Lloop0:
115	/* Save hash values for addition after rounds */
116	movdqa		E0, (0*16)(%rsp)
117	movdqa		ABCD, (1*16)(%rsp)
118
119	/* Rounds 0-3 */
120	movdqu		0*16(DATA_PTR), MSG0
121	pshufb		SHUF_MASK, MSG0
122		paddd		MSG0, E0
123		movdqa		ABCD, E1
124		sha1rnds4	$0, E0, ABCD
125
126	/* Rounds 4-7 */
127	movdqu		1*16(DATA_PTR), MSG1
128	pshufb		SHUF_MASK, MSG1
129		sha1nexte	MSG1, E1
130		movdqa		ABCD, E0
131		sha1rnds4	$0, E1, ABCD
132	sha1msg1	MSG1, MSG0
133
134	/* Rounds 8-11 */
135	movdqu		2*16(DATA_PTR), MSG2
136	pshufb		SHUF_MASK, MSG2
137		sha1nexte	MSG2, E0
138		movdqa		ABCD, E1
139		sha1rnds4	$0, E0, ABCD
140	sha1msg1	MSG2, MSG1
141	pxor		MSG2, MSG0
142
143	/* Rounds 12-15 */
144	movdqu		3*16(DATA_PTR), MSG3
145	pshufb		SHUF_MASK, MSG3
146		sha1nexte	MSG3, E1
147		movdqa		ABCD, E0
148	sha1msg2	MSG3, MSG0
149		sha1rnds4	$0, E1, ABCD
150	sha1msg1	MSG3, MSG2
151	pxor		MSG3, MSG1
152
153	/* Rounds 16-19 */
154		sha1nexte	MSG0, E0
155		movdqa		ABCD, E1
156	sha1msg2	MSG0, MSG1
157		sha1rnds4	$0, E0, ABCD
158	sha1msg1	MSG0, MSG3
159	pxor		MSG0, MSG2
160
161	/* Rounds 20-23 */
162		sha1nexte	MSG1, E1
163		movdqa		ABCD, E0
164	sha1msg2	MSG1, MSG2
165		sha1rnds4	$1, E1, ABCD
166	sha1msg1	MSG1, MSG0
167	pxor		MSG1, MSG3
168
169	/* Rounds 24-27 */
170		sha1nexte	MSG2, E0
171		movdqa		ABCD, E1
172	sha1msg2	MSG2, MSG3
173		sha1rnds4	$1, E0, ABCD
174	sha1msg1	MSG2, MSG1
175	pxor		MSG2, MSG0
176
177	/* Rounds 28-31 */
178		sha1nexte	MSG3, E1
179		movdqa		ABCD, E0
180	sha1msg2	MSG3, MSG0
181		sha1rnds4	$1, E1, ABCD
182	sha1msg1	MSG3, MSG2
183	pxor		MSG3, MSG1
184
185	/* Rounds 32-35 */
186		sha1nexte	MSG0, E0
187		movdqa		ABCD, E1
188	sha1msg2	MSG0, MSG1
189		sha1rnds4	$1, E0, ABCD
190	sha1msg1	MSG0, MSG3
191	pxor		MSG0, MSG2
192
193	/* Rounds 36-39 */
194		sha1nexte	MSG1, E1
195		movdqa		ABCD, E0
196	sha1msg2	MSG1, MSG2
197		sha1rnds4	$1, E1, ABCD
198	sha1msg1	MSG1, MSG0
199	pxor		MSG1, MSG3
200
201	/* Rounds 40-43 */
202		sha1nexte	MSG2, E0
203		movdqa		ABCD, E1
204	sha1msg2	MSG2, MSG3
205		sha1rnds4	$2, E0, ABCD
206	sha1msg1	MSG2, MSG1
207	pxor		MSG2, MSG0
208
209	/* Rounds 44-47 */
210		sha1nexte	MSG3, E1
211		movdqa		ABCD, E0
212	sha1msg2	MSG3, MSG0
213		sha1rnds4	$2, E1, ABCD
214	sha1msg1	MSG3, MSG2
215	pxor		MSG3, MSG1
216
217	/* Rounds 48-51 */
218		sha1nexte	MSG0, E0
219		movdqa		ABCD, E1
220	sha1msg2	MSG0, MSG1
221		sha1rnds4	$2, E0, ABCD
222	sha1msg1	MSG0, MSG3
223	pxor		MSG0, MSG2
224
225	/* Rounds 52-55 */
226		sha1nexte	MSG1, E1
227		movdqa		ABCD, E0
228	sha1msg2	MSG1, MSG2
229		sha1rnds4	$2, E1, ABCD
230	sha1msg1	MSG1, MSG0
231	pxor		MSG1, MSG3
232
233	/* Rounds 56-59 */
234		sha1nexte	MSG2, E0
235		movdqa		ABCD, E1
236	sha1msg2	MSG2, MSG3
237		sha1rnds4	$2, E0, ABCD
238	sha1msg1	MSG2, MSG1
239	pxor		MSG2, MSG0
240
241	/* Rounds 60-63 */
242		sha1nexte	MSG3, E1
243		movdqa		ABCD, E0
244	sha1msg2	MSG3, MSG0
245		sha1rnds4	$3, E1, ABCD
246	sha1msg1	MSG3, MSG2
247	pxor		MSG3, MSG1
248
249	/* Rounds 64-67 */
250		sha1nexte	MSG0, E0
251		movdqa		ABCD, E1
252	sha1msg2	MSG0, MSG1
253		sha1rnds4	$3, E0, ABCD
254	sha1msg1	MSG0, MSG3
255	pxor		MSG0, MSG2
256
257	/* Rounds 68-71 */
258		sha1nexte	MSG1, E1
259		movdqa		ABCD, E0
260	sha1msg2	MSG1, MSG2
261		sha1rnds4	$3, E1, ABCD
262	pxor		MSG1, MSG3
263
264	/* Rounds 72-75 */
265		sha1nexte	MSG2, E0
266		movdqa		ABCD, E1
267	sha1msg2	MSG2, MSG3
268		sha1rnds4	$3, E0, ABCD
269
270	/* Rounds 76-79 */
271		sha1nexte	MSG3, E1
272		movdqa		ABCD, E0
273		sha1rnds4	$3, E1, ABCD
274
275	/* Add current hash values with previously saved */
276	sha1nexte	(0*16)(%rsp), E0
277	paddd		(1*16)(%rsp), ABCD
278
279	/* Increment data pointer and loop if more to process */
280	add		$64, DATA_PTR
281	cmp		NUM_BLKS, DATA_PTR
282	jne		.Lloop0
283
284	/* Write hash values back in the correct order */
285	pshufd		$0x1B, ABCD, ABCD
286	movdqu		ABCD, 0*16(DIGEST_PTR)
287	pextrd		$3, E0, 1*16(DIGEST_PTR)
288
289.Ldone_hash:
290	mov		%rbp, %rsp
291	pop		%rbp
292
293	ret
294SYM_FUNC_END(sha1_ni_transform)
295
296.section	.rodata.cst16.PSHUFFLE_BYTE_FLIP_MASK, "aM", @progbits, 16
297.align 16
298PSHUFFLE_BYTE_FLIP_MASK:
299	.octa 0x000102030405060708090a0b0c0d0e0f
300
301.section	.rodata.cst16.UPPER_WORD_MASK, "aM", @progbits, 16
302.align 16
303UPPER_WORD_MASK:
304	.octa 0xFFFFFFFF000000000000000000000000
305