xref: /linux/arch/x86/crypto/sha256_ni_asm.S (revision 621cde16e49b3ecf7d59a8106a20aaebfb4a59a9)
1/*
2 * Intel SHA Extensions optimized implementation of a SHA-256 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#include <linux/cfi_types.h>
58
59#define DIGEST_PTR	%rdi	/* 1st arg */
60#define DATA_PTR	%rsi	/* 2nd arg */
61#define NUM_BLKS	%rdx	/* 3rd arg */
62
63#define SHA256CONSTANTS	%rax
64
65#define MSG		%xmm0  /* sha256rnds2 implicit operand */
66#define STATE0		%xmm1
67#define STATE1		%xmm2
68#define MSG0		%xmm3
69#define MSG1		%xmm4
70#define MSG2		%xmm5
71#define MSG3		%xmm6
72#define TMP		%xmm7
73
74#define SHUF_MASK	%xmm8
75
76#define ABEF_SAVE	%xmm9
77#define CDGH_SAVE	%xmm10
78
79.macro do_4rounds	i, m0, m1, m2, m3
80.if \i < 16
81	movdqu		\i*4(DATA_PTR), \m0
82	pshufb		SHUF_MASK, \m0
83.endif
84	movdqa		(\i-32)*4(SHA256CONSTANTS), MSG
85	paddd		\m0, MSG
86	sha256rnds2	STATE0, STATE1
87.if \i >= 12 && \i < 60
88	movdqa		\m0, TMP
89	palignr		$4, \m3, TMP
90	paddd		TMP, \m1
91	sha256msg2	\m0, \m1
92.endif
93	punpckhqdq	MSG, MSG
94	sha256rnds2	STATE1, STATE0
95.if \i >= 4 && \i < 52
96	sha256msg1	\m0, \m3
97.endif
98.endm
99
100/*
101 * Intel SHA Extensions optimized implementation of a SHA-256 update function
102 *
103 * The function takes a pointer to the current hash values, a pointer to the
104 * input data, and a number of 64 byte blocks to process.  Once all blocks have
105 * been processed, the digest pointer is  updated with the resulting hash value.
106 * The function only processes complete blocks, there is no functionality to
107 * store partial blocks.  All message padding and hash value initialization must
108 * be done outside the update function.
109 *
110 * void sha256_ni_transform(uint32_t *digest, const void *data,
111		uint32_t numBlocks);
112 * digest : pointer to digest
113 * data: pointer to input data
114 * numBlocks: Number of blocks to process
115 */
116
117.text
118SYM_TYPED_FUNC_START(sha256_ni_transform)
119
120	shl		$6, NUM_BLKS		/*  convert to bytes */
121	jz		.Ldone_hash
122	add		DATA_PTR, NUM_BLKS	/* pointer to end of data */
123
124	/*
125	 * load initial hash values
126	 * Need to reorder these appropriately
127	 * DCBA, HGFE -> ABEF, CDGH
128	 */
129	movdqu		0*16(DIGEST_PTR), STATE0	/* DCBA */
130	movdqu		1*16(DIGEST_PTR), STATE1	/* HGFE */
131
132	movdqa		STATE0, TMP
133	punpcklqdq	STATE1, STATE0			/* FEBA */
134	punpckhqdq	TMP, STATE1			/* DCHG */
135	pshufd		$0x1B, STATE0, STATE0		/* ABEF */
136	pshufd		$0xB1, STATE1, STATE1		/* CDGH */
137
138	movdqa		PSHUFFLE_BYTE_FLIP_MASK(%rip), SHUF_MASK
139	lea		K256+32*4(%rip), SHA256CONSTANTS
140
141.Lloop0:
142	/* Save hash values for addition after rounds */
143	movdqa		STATE0, ABEF_SAVE
144	movdqa		STATE1, CDGH_SAVE
145
146.irp i, 0, 16, 32, 48
147	do_4rounds	(\i + 0),  MSG0, MSG1, MSG2, MSG3
148	do_4rounds	(\i + 4),  MSG1, MSG2, MSG3, MSG0
149	do_4rounds	(\i + 8),  MSG2, MSG3, MSG0, MSG1
150	do_4rounds	(\i + 12), MSG3, MSG0, MSG1, MSG2
151.endr
152
153	/* Add current hash values with previously saved */
154	paddd		ABEF_SAVE, STATE0
155	paddd		CDGH_SAVE, STATE1
156
157	/* Increment data pointer and loop if more to process */
158	add		$64, DATA_PTR
159	cmp		NUM_BLKS, DATA_PTR
160	jne		.Lloop0
161
162	/* Write hash values back in the correct order */
163	movdqa		STATE0, TMP
164	punpcklqdq	STATE1, STATE0			/* GHEF */
165	punpckhqdq	TMP, STATE1			/* ABCD */
166	pshufd		$0xB1, STATE0, STATE0		/* HGFE */
167	pshufd		$0x1B, STATE1, STATE1		/* DCBA */
168
169	movdqu		STATE1, 0*16(DIGEST_PTR)
170	movdqu		STATE0, 1*16(DIGEST_PTR)
171
172.Ldone_hash:
173
174	RET
175SYM_FUNC_END(sha256_ni_transform)
176
177.section	.rodata.cst256.K256, "aM", @progbits, 256
178.align 64
179K256:
180	.long	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
181	.long	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
182	.long	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
183	.long	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
184	.long	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
185	.long	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
186	.long	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
187	.long	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
188	.long	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
189	.long	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
190	.long	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
191	.long	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
192	.long	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
193	.long	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
194	.long	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
195	.long	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
196
197.section	.rodata.cst16.PSHUFFLE_BYTE_FLIP_MASK, "aM", @progbits, 16
198.align 16
199PSHUFFLE_BYTE_FLIP_MASK:
200	.octa 0x0c0d0e0f08090a0b0405060700010203
201