diff options
-rw-r--r-- | arch/arm/crypto/Kconfig | 9 | ||||
-rw-r--r-- | arch/arm/crypto/Makefile | 2 | ||||
-rw-r--r-- | arch/arm/crypto/sha2-ce-core.S | 134 | ||||
-rw-r--r-- | arch/arm/crypto/sha2-ce-glue.c | 203 |
4 files changed, 348 insertions, 0 deletions
diff --git a/arch/arm/crypto/Kconfig b/arch/arm/crypto/Kconfig index d7bc10beb8ac..9c1478e55a40 100644 --- a/arch/arm/crypto/Kconfig +++ b/arch/arm/crypto/Kconfig | |||
@@ -37,6 +37,15 @@ config CRYPTO_SHA1_ARM_CE | |||
37 | SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2) implemented | 37 | SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2) implemented |
38 | using special ARMv8 Crypto Extensions. | 38 | using special ARMv8 Crypto Extensions. |
39 | 39 | ||
40 | config CRYPTO_SHA2_ARM_CE | ||
41 | tristate "SHA-224/256 digest algorithm (ARM v8 Crypto Extensions)" | ||
42 | depends on KERNEL_MODE_NEON | ||
43 | select CRYPTO_SHA256 | ||
44 | select CRYPTO_HASH | ||
45 | help | ||
46 | SHA-256 secure hash standard (DFIPS 180-2) implemented | ||
47 | using special ARMv8 Crypto Extensions. | ||
48 | |||
40 | config CRYPTO_SHA512_ARM_NEON | 49 | config CRYPTO_SHA512_ARM_NEON |
41 | tristate "SHA384 and SHA512 digest algorithm (ARM NEON)" | 50 | tristate "SHA384 and SHA512 digest algorithm (ARM NEON)" |
42 | depends on KERNEL_MODE_NEON | 51 | depends on KERNEL_MODE_NEON |
diff --git a/arch/arm/crypto/Makefile b/arch/arm/crypto/Makefile index d92d05ba646e..4ea9f96c2782 100644 --- a/arch/arm/crypto/Makefile +++ b/arch/arm/crypto/Makefile | |||
@@ -8,6 +8,7 @@ obj-$(CONFIG_CRYPTO_SHA1_ARM) += sha1-arm.o | |||
8 | obj-$(CONFIG_CRYPTO_SHA1_ARM_NEON) += sha1-arm-neon.o | 8 | obj-$(CONFIG_CRYPTO_SHA1_ARM_NEON) += sha1-arm-neon.o |
9 | obj-$(CONFIG_CRYPTO_SHA512_ARM_NEON) += sha512-arm-neon.o | 9 | obj-$(CONFIG_CRYPTO_SHA512_ARM_NEON) += sha512-arm-neon.o |
10 | obj-$(CONFIG_CRYPTO_SHA1_ARM_CE) += sha1-arm-ce.o | 10 | obj-$(CONFIG_CRYPTO_SHA1_ARM_CE) += sha1-arm-ce.o |
11 | obj-$(CONFIG_CRYPTO_SHA2_ARM_CE) += sha2-arm-ce.o | ||
11 | 12 | ||
12 | aes-arm-y := aes-armv4.o aes_glue.o | 13 | aes-arm-y := aes-armv4.o aes_glue.o |
13 | aes-arm-bs-y := aesbs-core.o aesbs-glue.o | 14 | aes-arm-bs-y := aesbs-core.o aesbs-glue.o |
@@ -15,6 +16,7 @@ sha1-arm-y := sha1-armv4-large.o sha1_glue.o | |||
15 | sha1-arm-neon-y := sha1-armv7-neon.o sha1_neon_glue.o | 16 | sha1-arm-neon-y := sha1-armv7-neon.o sha1_neon_glue.o |
16 | sha512-arm-neon-y := sha512-armv7-neon.o sha512_neon_glue.o | 17 | sha512-arm-neon-y := sha512-armv7-neon.o sha512_neon_glue.o |
17 | sha1-arm-ce-y := sha1-ce-core.o sha1-ce-glue.o | 18 | sha1-arm-ce-y := sha1-ce-core.o sha1-ce-glue.o |
19 | sha2-arm-ce-y := sha2-ce-core.o sha2-ce-glue.o | ||
18 | 20 | ||
19 | quiet_cmd_perl = PERL $@ | 21 | quiet_cmd_perl = PERL $@ |
20 | cmd_perl = $(PERL) $(<) > $(@) | 22 | cmd_perl = $(PERL) $(<) > $(@) |
diff --git a/arch/arm/crypto/sha2-ce-core.S b/arch/arm/crypto/sha2-ce-core.S new file mode 100644 index 000000000000..96af09fe957b --- /dev/null +++ b/arch/arm/crypto/sha2-ce-core.S | |||
@@ -0,0 +1,134 @@ | |||
1 | /* | ||
2 | * sha2-ce-core.S - SHA-224/256 secure hash using ARMv8 Crypto Extensions | ||
3 | * | ||
4 | * Copyright (C) 2015 Linaro Ltd. | ||
5 | * Author: Ard Biesheuvel <ard.biesheuvel@linaro.org> | ||
6 | * | ||
7 | * This program is free software; you can redistribute it and/or modify | ||
8 | * it under the terms of the GNU General Public License version 2 as | ||
9 | * published by the Free Software Foundation. | ||
10 | */ | ||
11 | |||
12 | #include <linux/linkage.h> | ||
13 | #include <asm/assembler.h> | ||
14 | |||
15 | .text | ||
16 | .fpu crypto-neon-fp-armv8 | ||
17 | |||
18 | k0 .req q7 | ||
19 | k1 .req q8 | ||
20 | rk .req r3 | ||
21 | |||
22 | ta0 .req q9 | ||
23 | ta1 .req q10 | ||
24 | tb0 .req q10 | ||
25 | tb1 .req q9 | ||
26 | |||
27 | dga .req q11 | ||
28 | dgb .req q12 | ||
29 | |||
30 | dg0 .req q13 | ||
31 | dg1 .req q14 | ||
32 | dg2 .req q15 | ||
33 | |||
34 | .macro add_only, ev, s0 | ||
35 | vmov dg2, dg0 | ||
36 | .ifnb \s0 | ||
37 | vld1.32 {k\ev}, [rk, :128]! | ||
38 | .endif | ||
39 | sha256h.32 dg0, dg1, tb\ev | ||
40 | sha256h2.32 dg1, dg2, tb\ev | ||
41 | .ifnb \s0 | ||
42 | vadd.u32 ta\ev, q\s0, k\ev | ||
43 | .endif | ||
44 | .endm | ||
45 | |||
46 | .macro add_update, ev, s0, s1, s2, s3 | ||
47 | sha256su0.32 q\s0, q\s1 | ||
48 | add_only \ev, \s1 | ||
49 | sha256su1.32 q\s0, q\s2, q\s3 | ||
50 | .endm | ||
51 | |||
52 | .align 6 | ||
53 | .Lsha256_rcon: | ||
54 | .word 0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5 | ||
55 | .word 0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5 | ||
56 | .word 0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3 | ||
57 | .word 0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174 | ||
58 | .word 0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc | ||
59 | .word 0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da | ||
60 | .word 0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7 | ||
61 | .word 0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967 | ||
62 | .word 0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13 | ||
63 | .word 0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85 | ||
64 | .word 0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3 | ||
65 | .word 0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070 | ||
66 | .word 0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5 | ||
67 | .word 0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3 | ||
68 | .word 0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208 | ||
69 | .word 0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2 | ||
70 | |||
71 | /* | ||
72 | * void sha2_ce_transform(int blocks, u8 const *src, u32 *state, | ||
73 | * u8 *head); | ||
74 | */ | ||
75 | ENTRY(sha2_ce_transform) | ||
76 | /* load state */ | ||
77 | vld1.32 {dga-dgb}, [r2] | ||
78 | |||
79 | /* load partial input (if supplied) */ | ||
80 | teq r3, #0 | ||
81 | beq 0f | ||
82 | vld1.32 {q0-q1}, [r3]! | ||
83 | vld1.32 {q2-q3}, [r3] | ||
84 | teq r0, #0 | ||
85 | b 1f | ||
86 | |||
87 | /* load input */ | ||
88 | 0: vld1.32 {q0-q1}, [r1]! | ||
89 | vld1.32 {q2-q3}, [r1]! | ||
90 | subs r0, r0, #1 | ||
91 | |||
92 | 1: | ||
93 | #ifndef CONFIG_CPU_BIG_ENDIAN | ||
94 | vrev32.8 q0, q0 | ||
95 | vrev32.8 q1, q1 | ||
96 | vrev32.8 q2, q2 | ||
97 | vrev32.8 q3, q3 | ||
98 | #endif | ||
99 | |||
100 | /* load first round constant */ | ||
101 | adr rk, .Lsha256_rcon | ||
102 | vld1.32 {k0}, [rk, :128]! | ||
103 | |||
104 | vadd.u32 ta0, q0, k0 | ||
105 | vmov dg0, dga | ||
106 | vmov dg1, dgb | ||
107 | |||
108 | add_update 1, 0, 1, 2, 3 | ||
109 | add_update 0, 1, 2, 3, 0 | ||
110 | add_update 1, 2, 3, 0, 1 | ||
111 | add_update 0, 3, 0, 1, 2 | ||
112 | add_update 1, 0, 1, 2, 3 | ||
113 | add_update 0, 1, 2, 3, 0 | ||
114 | add_update 1, 2, 3, 0, 1 | ||
115 | add_update 0, 3, 0, 1, 2 | ||
116 | add_update 1, 0, 1, 2, 3 | ||
117 | add_update 0, 1, 2, 3, 0 | ||
118 | add_update 1, 2, 3, 0, 1 | ||
119 | add_update 0, 3, 0, 1, 2 | ||
120 | |||
121 | add_only 1, 1 | ||
122 | add_only 0, 2 | ||
123 | add_only 1, 3 | ||
124 | add_only 0 | ||
125 | |||
126 | /* update state */ | ||
127 | vadd.u32 dga, dga, dg0 | ||
128 | vadd.u32 dgb, dgb, dg1 | ||
129 | bne 0b | ||
130 | |||
131 | /* store new state */ | ||
132 | vst1.32 {dga-dgb}, [r2] | ||
133 | bx lr | ||
134 | ENDPROC(sha2_ce_transform) | ||
diff --git a/arch/arm/crypto/sha2-ce-glue.c b/arch/arm/crypto/sha2-ce-glue.c new file mode 100644 index 000000000000..9ffe8ad27402 --- /dev/null +++ b/arch/arm/crypto/sha2-ce-glue.c | |||
@@ -0,0 +1,203 @@ | |||
1 | /* | ||
2 | * sha2-ce-glue.c - SHA-224/SHA-256 using ARMv8 Crypto Extensions | ||
3 | * | ||
4 | * Copyright (C) 2015 Linaro Ltd <ard.biesheuvel@linaro.org> | ||
5 | * | ||
6 | * This program is free software; you can redistribute it and/or modify | ||
7 | * it under the terms of the GNU General Public License version 2 as | ||
8 | * published by the Free Software Foundation. | ||
9 | */ | ||
10 | |||
11 | #include <crypto/internal/hash.h> | ||
12 | #include <crypto/sha.h> | ||
13 | #include <linux/crypto.h> | ||
14 | #include <linux/module.h> | ||
15 | |||
16 | #include <asm/hwcap.h> | ||
17 | #include <asm/simd.h> | ||
18 | #include <asm/neon.h> | ||
19 | #include <asm/unaligned.h> | ||
20 | |||
21 | MODULE_DESCRIPTION("SHA-224/SHA-256 secure hash using ARMv8 Crypto Extensions"); | ||
22 | MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>"); | ||
23 | MODULE_LICENSE("GPL v2"); | ||
24 | |||
25 | asmlinkage void sha2_ce_transform(int blocks, u8 const *src, u32 *state, | ||
26 | u8 *head); | ||
27 | |||
28 | static int sha224_init(struct shash_desc *desc) | ||
29 | { | ||
30 | struct sha256_state *sctx = shash_desc_ctx(desc); | ||
31 | |||
32 | *sctx = (struct sha256_state){ | ||
33 | .state = { | ||
34 | SHA224_H0, SHA224_H1, SHA224_H2, SHA224_H3, | ||
35 | SHA224_H4, SHA224_H5, SHA224_H6, SHA224_H7, | ||
36 | } | ||
37 | }; | ||
38 | return 0; | ||
39 | } | ||
40 | |||
41 | static int sha256_init(struct shash_desc *desc) | ||
42 | { | ||
43 | struct sha256_state *sctx = shash_desc_ctx(desc); | ||
44 | |||
45 | *sctx = (struct sha256_state){ | ||
46 | .state = { | ||
47 | SHA256_H0, SHA256_H1, SHA256_H2, SHA256_H3, | ||
48 | SHA256_H4, SHA256_H5, SHA256_H6, SHA256_H7, | ||
49 | } | ||
50 | }; | ||
51 | return 0; | ||
52 | } | ||
53 | |||
54 | static int sha2_update(struct shash_desc *desc, const u8 *data, | ||
55 | unsigned int len) | ||
56 | { | ||
57 | struct sha256_state *sctx = shash_desc_ctx(desc); | ||
58 | unsigned int partial; | ||
59 | |||
60 | if (!may_use_simd()) | ||
61 | return crypto_sha256_update(desc, data, len); | ||
62 | |||
63 | partial = sctx->count % SHA256_BLOCK_SIZE; | ||
64 | sctx->count += len; | ||
65 | |||
66 | if ((partial + len) >= SHA256_BLOCK_SIZE) { | ||
67 | int blocks; | ||
68 | |||
69 | if (partial) { | ||
70 | int p = SHA256_BLOCK_SIZE - partial; | ||
71 | |||
72 | memcpy(sctx->buf + partial, data, p); | ||
73 | data += p; | ||
74 | len -= p; | ||
75 | } | ||
76 | |||
77 | blocks = len / SHA256_BLOCK_SIZE; | ||
78 | len %= SHA256_BLOCK_SIZE; | ||
79 | |||
80 | kernel_neon_begin(); | ||
81 | sha2_ce_transform(blocks, data, sctx->state, | ||
82 | partial ? sctx->buf : NULL); | ||
83 | kernel_neon_end(); | ||
84 | |||
85 | data += blocks * SHA256_BLOCK_SIZE; | ||
86 | partial = 0; | ||
87 | } | ||
88 | if (len) | ||
89 | memcpy(sctx->buf + partial, data, len); | ||
90 | return 0; | ||
91 | } | ||
92 | |||
93 | static void sha2_final(struct shash_desc *desc) | ||
94 | { | ||
95 | static const u8 padding[SHA256_BLOCK_SIZE] = { 0x80, }; | ||
96 | |||
97 | struct sha256_state *sctx = shash_desc_ctx(desc); | ||
98 | __be64 bits = cpu_to_be64(sctx->count << 3); | ||
99 | u32 padlen = SHA256_BLOCK_SIZE | ||
100 | - ((sctx->count + sizeof(bits)) % SHA256_BLOCK_SIZE); | ||
101 | |||
102 | sha2_update(desc, padding, padlen); | ||
103 | sha2_update(desc, (const u8 *)&bits, sizeof(bits)); | ||
104 | } | ||
105 | |||
106 | static int sha224_final(struct shash_desc *desc, u8 *out) | ||
107 | { | ||
108 | struct sha256_state *sctx = shash_desc_ctx(desc); | ||
109 | __be32 *dst = (__be32 *)out; | ||
110 | int i; | ||
111 | |||
112 | sha2_final(desc); | ||
113 | |||
114 | for (i = 0; i < SHA224_DIGEST_SIZE / sizeof(__be32); i++) | ||
115 | put_unaligned_be32(sctx->state[i], dst++); | ||
116 | |||
117 | *sctx = (struct sha256_state){}; | ||
118 | return 0; | ||
119 | } | ||
120 | |||
121 | static int sha256_final(struct shash_desc *desc, u8 *out) | ||
122 | { | ||
123 | struct sha256_state *sctx = shash_desc_ctx(desc); | ||
124 | __be32 *dst = (__be32 *)out; | ||
125 | int i; | ||
126 | |||
127 | sha2_final(desc); | ||
128 | |||
129 | for (i = 0; i < SHA256_DIGEST_SIZE / sizeof(__be32); i++) | ||
130 | put_unaligned_be32(sctx->state[i], dst++); | ||
131 | |||
132 | *sctx = (struct sha256_state){}; | ||
133 | return 0; | ||
134 | } | ||
135 | |||
136 | static int sha2_export(struct shash_desc *desc, void *out) | ||
137 | { | ||
138 | struct sha256_state *sctx = shash_desc_ctx(desc); | ||
139 | struct sha256_state *dst = out; | ||
140 | |||
141 | *dst = *sctx; | ||
142 | return 0; | ||
143 | } | ||
144 | |||
145 | static int sha2_import(struct shash_desc *desc, const void *in) | ||
146 | { | ||
147 | struct sha256_state *sctx = shash_desc_ctx(desc); | ||
148 | struct sha256_state const *src = in; | ||
149 | |||
150 | *sctx = *src; | ||
151 | return 0; | ||
152 | } | ||
153 | |||
154 | static struct shash_alg algs[] = { { | ||
155 | .init = sha224_init, | ||
156 | .update = sha2_update, | ||
157 | .final = sha224_final, | ||
158 | .export = sha2_export, | ||
159 | .import = sha2_import, | ||
160 | .descsize = sizeof(struct sha256_state), | ||
161 | .digestsize = SHA224_DIGEST_SIZE, | ||
162 | .statesize = sizeof(struct sha256_state), | ||
163 | .base = { | ||
164 | .cra_name = "sha224", | ||
165 | .cra_driver_name = "sha224-ce", | ||
166 | .cra_priority = 200, | ||
167 | .cra_flags = CRYPTO_ALG_TYPE_SHASH, | ||
168 | .cra_blocksize = SHA256_BLOCK_SIZE, | ||
169 | .cra_module = THIS_MODULE, | ||
170 | } | ||
171 | }, { | ||
172 | .init = sha256_init, | ||
173 | .update = sha2_update, | ||
174 | .final = sha256_final, | ||
175 | .export = sha2_export, | ||
176 | .import = sha2_import, | ||
177 | .descsize = sizeof(struct sha256_state), | ||
178 | .digestsize = SHA256_DIGEST_SIZE, | ||
179 | .statesize = sizeof(struct sha256_state), | ||
180 | .base = { | ||
181 | .cra_name = "sha256", | ||
182 | .cra_driver_name = "sha256-ce", | ||
183 | .cra_priority = 200, | ||
184 | .cra_flags = CRYPTO_ALG_TYPE_SHASH, | ||
185 | .cra_blocksize = SHA256_BLOCK_SIZE, | ||
186 | .cra_module = THIS_MODULE, | ||
187 | } | ||
188 | } }; | ||
189 | |||
190 | static int __init sha2_ce_mod_init(void) | ||
191 | { | ||
192 | if (!(elf_hwcap2 & HWCAP2_SHA2)) | ||
193 | return -ENODEV; | ||
194 | return crypto_register_shashes(algs, ARRAY_SIZE(algs)); | ||
195 | } | ||
196 | |||
197 | static void __exit sha2_ce_mod_fini(void) | ||
198 | { | ||
199 | crypto_unregister_shashes(algs, ARRAY_SIZE(algs)); | ||
200 | } | ||
201 | |||
202 | module_init(sha2_ce_mod_init); | ||
203 | module_exit(sha2_ce_mod_fini); | ||