aboutsummaryrefslogtreecommitdiffstats
path: root/arch/x86_64
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86_64')
-rw-r--r--arch/x86_64/Makefile4
-rw-r--r--arch/x86_64/crypto/Makefile9
-rw-r--r--arch/x86_64/crypto/aes-x86_64-asm.S186
-rw-r--r--arch/x86_64/crypto/aes.c324
4 files changed, 522 insertions, 1 deletions
diff --git a/arch/x86_64/Makefile b/arch/x86_64/Makefile
index 8a73794f9b90..428915697675 100644
--- a/arch/x86_64/Makefile
+++ b/arch/x86_64/Makefile
@@ -65,7 +65,9 @@ CFLAGS += $(call cc-option,-mno-sse -mno-mmx -mno-sse2 -mno-3dnow,)
65head-y := arch/x86_64/kernel/head.o arch/x86_64/kernel/head64.o arch/x86_64/kernel/init_task.o 65head-y := arch/x86_64/kernel/head.o arch/x86_64/kernel/head64.o arch/x86_64/kernel/init_task.o
66 66
67libs-y += arch/x86_64/lib/ 67libs-y += arch/x86_64/lib/
68core-y += arch/x86_64/kernel/ arch/x86_64/mm/ 68core-y += arch/x86_64/kernel/ \
69 arch/x86_64/mm/ \
70 arch/x86_64/crypto/
69core-$(CONFIG_IA32_EMULATION) += arch/x86_64/ia32/ 71core-$(CONFIG_IA32_EMULATION) += arch/x86_64/ia32/
70drivers-$(CONFIG_PCI) += arch/x86_64/pci/ 72drivers-$(CONFIG_PCI) += arch/x86_64/pci/
71drivers-$(CONFIG_OPROFILE) += arch/x86_64/oprofile/ 73drivers-$(CONFIG_OPROFILE) += arch/x86_64/oprofile/
diff --git a/arch/x86_64/crypto/Makefile b/arch/x86_64/crypto/Makefile
new file mode 100644
index 000000000000..426d20f4b72e
--- /dev/null
+++ b/arch/x86_64/crypto/Makefile
@@ -0,0 +1,9 @@
1#
2# x86_64/crypto/Makefile
3#
4# Arch-specific CryptoAPI modules.
5#
6
7obj-$(CONFIG_CRYPTO_AES_X86_64) += aes-x86_64.o
8
9aes-x86_64-y := aes-x86_64-asm.o aes.o
diff --git a/arch/x86_64/crypto/aes-x86_64-asm.S b/arch/x86_64/crypto/aes-x86_64-asm.S
new file mode 100644
index 000000000000..483cbb23ab8d
--- /dev/null
+++ b/arch/x86_64/crypto/aes-x86_64-asm.S
@@ -0,0 +1,186 @@
1/* AES (Rijndael) implementation (FIPS PUB 197) for x86_64
2 *
3 * Copyright (C) 2005 Andreas Steinmetz, <ast@domdv.de>
4 *
5 * License:
6 * This code can be distributed under the terms of the GNU General Public
7 * License (GPL) Version 2 provided that the above header down to and
8 * including this sentence is retained in full.
9 */
10
11.extern aes_ft_tab
12.extern aes_it_tab
13.extern aes_fl_tab
14.extern aes_il_tab
15
16.text
17
18#define R1 %rax
19#define R1E %eax
20#define R1X %ax
21#define R1H %ah
22#define R1L %al
23#define R2 %rbx
24#define R2E %ebx
25#define R2X %bx
26#define R2H %bh
27#define R2L %bl
28#define R3 %rcx
29#define R3E %ecx
30#define R3X %cx
31#define R3H %ch
32#define R3L %cl
33#define R4 %rdx
34#define R4E %edx
35#define R4X %dx
36#define R4H %dh
37#define R4L %dl
38#define R5 %rsi
39#define R5E %esi
40#define R6 %rdi
41#define R6E %edi
42#define R7 %rbp
43#define R7E %ebp
44#define R8 %r8
45#define R9 %r9
46#define R10 %r10
47#define R11 %r11
48
49#define prologue(FUNC,BASE,B128,B192,r1,r2,r3,r4,r5,r6,r7,r8,r9,r10,r11) \
50 .global FUNC; \
51 .type FUNC,@function; \
52 .align 8; \
53FUNC: movq r1,r2; \
54 movq r3,r4; \
55 leaq BASE+52(r8),r9; \
56 movq r10,r11; \
57 movl (r7),r5 ## E; \
58 movl 4(r7),r1 ## E; \
59 movl 8(r7),r6 ## E; \
60 movl 12(r7),r7 ## E; \
61 movl (r8),r10 ## E; \
62 xorl -48(r9),r5 ## E; \
63 xorl -44(r9),r1 ## E; \
64 xorl -40(r9),r6 ## E; \
65 xorl -36(r9),r7 ## E; \
66 cmpl $24,r10 ## E; \
67 jb B128; \
68 leaq 32(r9),r9; \
69 je B192; \
70 leaq 32(r9),r9;
71
72#define epilogue(r1,r2,r3,r4,r5,r6,r7,r8,r9) \
73 movq r1,r2; \
74 movq r3,r4; \
75 movl r5 ## E,(r9); \
76 movl r6 ## E,4(r9); \
77 movl r7 ## E,8(r9); \
78 movl r8 ## E,12(r9); \
79 ret;
80
81#define round(TAB,OFFSET,r1,r2,r3,r4,r5,r6,r7,r8,ra,rb,rc,rd) \
82 movzbl r2 ## H,r5 ## E; \
83 movzbl r2 ## L,r6 ## E; \
84 movl TAB+1024(,r5,4),r5 ## E;\
85 movw r4 ## X,r2 ## X; \
86 movl TAB(,r6,4),r6 ## E; \
87 roll $16,r2 ## E; \
88 shrl $16,r4 ## E; \
89 movzbl r4 ## H,r7 ## E; \
90 movzbl r4 ## L,r4 ## E; \
91 xorl OFFSET(r8),ra ## E; \
92 xorl OFFSET+4(r8),rb ## E; \
93 xorl TAB+3072(,r7,4),r5 ## E;\
94 xorl TAB+2048(,r4,4),r6 ## E;\
95 movzbl r1 ## L,r7 ## E; \
96 movzbl r1 ## H,r4 ## E; \
97 movl TAB+1024(,r4,4),r4 ## E;\
98 movw r3 ## X,r1 ## X; \
99 roll $16,r1 ## E; \
100 shrl $16,r3 ## E; \
101 xorl TAB(,r7,4),r5 ## E; \
102 movzbl r3 ## H,r7 ## E; \
103 movzbl r3 ## L,r3 ## E; \
104 xorl TAB+3072(,r7,4),r4 ## E;\
105 xorl TAB+2048(,r3,4),r5 ## E;\
106 movzbl r1 ## H,r7 ## E; \
107 movzbl r1 ## L,r3 ## E; \
108 shrl $16,r1 ## E; \
109 xorl TAB+3072(,r7,4),r6 ## E;\
110 movl TAB+2048(,r3,4),r3 ## E;\
111 movzbl r1 ## H,r7 ## E; \
112 movzbl r1 ## L,r1 ## E; \
113 xorl TAB+1024(,r7,4),r6 ## E;\
114 xorl TAB(,r1,4),r3 ## E; \
115 movzbl r2 ## H,r1 ## E; \
116 movzbl r2 ## L,r7 ## E; \
117 shrl $16,r2 ## E; \
118 xorl TAB+3072(,r1,4),r3 ## E;\
119 xorl TAB+2048(,r7,4),r4 ## E;\
120 movzbl r2 ## H,r1 ## E; \
121 movzbl r2 ## L,r2 ## E; \
122 xorl OFFSET+8(r8),rc ## E; \
123 xorl OFFSET+12(r8),rd ## E; \
124 xorl TAB+1024(,r1,4),r3 ## E;\
125 xorl TAB(,r2,4),r4 ## E;
126
127#define move_regs(r1,r2,r3,r4) \
128 movl r3 ## E,r1 ## E; \
129 movl r4 ## E,r2 ## E;
130
131#define entry(FUNC,BASE,B128,B192) \
132 prologue(FUNC,BASE,B128,B192,R2,R8,R7,R9,R1,R3,R4,R6,R10,R5,R11)
133
134#define return epilogue(R8,R2,R9,R7,R5,R6,R3,R4,R11)
135
136#define encrypt_round(TAB,OFFSET) \
137 round(TAB,OFFSET,R1,R2,R3,R4,R5,R6,R7,R10,R5,R6,R3,R4) \
138 move_regs(R1,R2,R5,R6)
139
140#define encrypt_final(TAB,OFFSET) \
141 round(TAB,OFFSET,R1,R2,R3,R4,R5,R6,R7,R10,R5,R6,R3,R4)
142
143#define decrypt_round(TAB,OFFSET) \
144 round(TAB,OFFSET,R2,R1,R4,R3,R6,R5,R7,R10,R5,R6,R3,R4) \
145 move_regs(R1,R2,R5,R6)
146
147#define decrypt_final(TAB,OFFSET) \
148 round(TAB,OFFSET,R2,R1,R4,R3,R6,R5,R7,R10,R5,R6,R3,R4)
149
150/* void aes_encrypt(void *ctx, u8 *out, const u8 *in) */
151
152 entry(aes_encrypt,0,enc128,enc192)
153 encrypt_round(aes_ft_tab,-96)
154 encrypt_round(aes_ft_tab,-80)
155enc192: encrypt_round(aes_ft_tab,-64)
156 encrypt_round(aes_ft_tab,-48)
157enc128: encrypt_round(aes_ft_tab,-32)
158 encrypt_round(aes_ft_tab,-16)
159 encrypt_round(aes_ft_tab, 0)
160 encrypt_round(aes_ft_tab, 16)
161 encrypt_round(aes_ft_tab, 32)
162 encrypt_round(aes_ft_tab, 48)
163 encrypt_round(aes_ft_tab, 64)
164 encrypt_round(aes_ft_tab, 80)
165 encrypt_round(aes_ft_tab, 96)
166 encrypt_final(aes_fl_tab,112)
167 return
168
169/* void aes_decrypt(void *ctx, u8 *out, const u8 *in) */
170
171 entry(aes_decrypt,240,dec128,dec192)
172 decrypt_round(aes_it_tab,-96)
173 decrypt_round(aes_it_tab,-80)
174dec192: decrypt_round(aes_it_tab,-64)
175 decrypt_round(aes_it_tab,-48)
176dec128: decrypt_round(aes_it_tab,-32)
177 decrypt_round(aes_it_tab,-16)
178 decrypt_round(aes_it_tab, 0)
179 decrypt_round(aes_it_tab, 16)
180 decrypt_round(aes_it_tab, 32)
181 decrypt_round(aes_it_tab, 48)
182 decrypt_round(aes_it_tab, 64)
183 decrypt_round(aes_it_tab, 80)
184 decrypt_round(aes_it_tab, 96)
185 decrypt_final(aes_il_tab,112)
186 return
diff --git a/arch/x86_64/crypto/aes.c b/arch/x86_64/crypto/aes.c
new file mode 100644
index 000000000000..2b5c4010ce38
--- /dev/null
+++ b/arch/x86_64/crypto/aes.c
@@ -0,0 +1,324 @@
1/*
2 * Cryptographic API.
3 *
4 * AES Cipher Algorithm.
5 *
6 * Based on Brian Gladman's code.
7 *
8 * Linux developers:
9 * Alexander Kjeldaas <astor@fast.no>
10 * Herbert Valerio Riedel <hvr@hvrlab.org>
11 * Kyle McMartin <kyle@debian.org>
12 * Adam J. Richter <adam@yggdrasil.com> (conversion to 2.5 API).
13 * Andreas Steinmetz <ast@domdv.de> (adapted to x86_64 assembler)
14 *
15 * This program is free software; you can redistribute it and/or modify
16 * it under the terms of the GNU General Public License as published by
17 * the Free Software Foundation; either version 2 of the License, or
18 * (at your option) any later version.
19 *
20 * ---------------------------------------------------------------------------
21 * Copyright (c) 2002, Dr Brian Gladman <brg@gladman.me.uk>, Worcester, UK.
22 * All rights reserved.
23 *
24 * LICENSE TERMS
25 *
26 * The free distribution and use of this software in both source and binary
27 * form is allowed (with or without changes) provided that:
28 *
29 * 1. distributions of this source code include the above copyright
30 * notice, this list of conditions and the following disclaimer;
31 *
32 * 2. distributions in binary form include the above copyright
33 * notice, this list of conditions and the following disclaimer
34 * in the documentation and/or other associated materials;
35 *
36 * 3. the copyright holder's name is not used to endorse products
37 * built using this software without specific written permission.
38 *
39 * ALTERNATIVELY, provided that this notice is retained in full, this product
40 * may be distributed under the terms of the GNU General Public License (GPL),
41 * in which case the provisions of the GPL apply INSTEAD OF those given above.
42 *
43 * DISCLAIMER
44 *
45 * This software is provided 'as is' with no explicit or implied warranties
46 * in respect of its properties, including, but not limited to, correctness
47 * and/or fitness for purpose.
48 * ---------------------------------------------------------------------------
49 */
50
51/* Some changes from the Gladman version:
52 s/RIJNDAEL(e_key)/E_KEY/g
53 s/RIJNDAEL(d_key)/D_KEY/g
54*/
55
56#include <asm/byteorder.h>
57#include <linux/bitops.h>
58#include <linux/crypto.h>
59#include <linux/errno.h>
60#include <linux/init.h>
61#include <linux/module.h>
62#include <linux/types.h>
63
64#define AES_MIN_KEY_SIZE 16
65#define AES_MAX_KEY_SIZE 32
66
67#define AES_BLOCK_SIZE 16
68
69/*
70 * #define byte(x, nr) ((unsigned char)((x) >> (nr*8)))
71 */
72static inline u8 byte(const u32 x, const unsigned n)
73{
74 return x >> (n << 3);
75}
76
77#define u32_in(x) le32_to_cpu(*(const __le32 *)(x))
78
79struct aes_ctx
80{
81 u32 key_length;
82 u32 E[60];
83 u32 D[60];
84};
85
86#define E_KEY ctx->E
87#define D_KEY ctx->D
88
89static u8 pow_tab[256] __initdata;
90static u8 log_tab[256] __initdata;
91static u8 sbx_tab[256] __initdata;
92static u8 isb_tab[256] __initdata;
93static u32 rco_tab[10];
94u32 aes_ft_tab[4][256];
95u32 aes_it_tab[4][256];
96
97u32 aes_fl_tab[4][256];
98u32 aes_il_tab[4][256];
99
100static inline u8 f_mult(u8 a, u8 b)
101{
102 u8 aa = log_tab[a], cc = aa + log_tab[b];
103
104 return pow_tab[cc + (cc < aa ? 1 : 0)];
105}
106
107#define ff_mult(a, b) (a && b ? f_mult(a, b) : 0)
108
109#define ls_box(x) \
110 (aes_fl_tab[0][byte(x, 0)] ^ \
111 aes_fl_tab[1][byte(x, 1)] ^ \
112 aes_fl_tab[2][byte(x, 2)] ^ \
113 aes_fl_tab[3][byte(x, 3)])
114
115static void __init gen_tabs(void)
116{
117 u32 i, t;
118 u8 p, q;
119
120 /* log and power tables for GF(2**8) finite field with
121 0x011b as modular polynomial - the simplest primitive
122 root is 0x03, used here to generate the tables */
123
124 for (i = 0, p = 1; i < 256; ++i) {
125 pow_tab[i] = (u8)p;
126 log_tab[p] = (u8)i;
127
128 p ^= (p << 1) ^ (p & 0x80 ? 0x01b : 0);
129 }
130
131 log_tab[1] = 0;
132
133 for (i = 0, p = 1; i < 10; ++i) {
134 rco_tab[i] = p;
135
136 p = (p << 1) ^ (p & 0x80 ? 0x01b : 0);
137 }
138
139 for (i = 0; i < 256; ++i) {
140 p = (i ? pow_tab[255 - log_tab[i]] : 0);
141 q = ((p >> 7) | (p << 1)) ^ ((p >> 6) | (p << 2));
142 p ^= 0x63 ^ q ^ ((q >> 6) | (q << 2));
143 sbx_tab[i] = p;
144 isb_tab[p] = (u8)i;
145 }
146
147 for (i = 0; i < 256; ++i) {
148 p = sbx_tab[i];
149
150 t = p;
151 aes_fl_tab[0][i] = t;
152 aes_fl_tab[1][i] = rol32(t, 8);
153 aes_fl_tab[2][i] = rol32(t, 16);
154 aes_fl_tab[3][i] = rol32(t, 24);
155
156 t = ((u32)ff_mult(2, p)) |
157 ((u32)p << 8) |
158 ((u32)p << 16) | ((u32)ff_mult(3, p) << 24);
159
160 aes_ft_tab[0][i] = t;
161 aes_ft_tab[1][i] = rol32(t, 8);
162 aes_ft_tab[2][i] = rol32(t, 16);
163 aes_ft_tab[3][i] = rol32(t, 24);
164
165 p = isb_tab[i];
166
167 t = p;
168 aes_il_tab[0][i] = t;
169 aes_il_tab[1][i] = rol32(t, 8);
170 aes_il_tab[2][i] = rol32(t, 16);
171 aes_il_tab[3][i] = rol32(t, 24);
172
173 t = ((u32)ff_mult(14, p)) |
174 ((u32)ff_mult(9, p) << 8) |
175 ((u32)ff_mult(13, p) << 16) |
176 ((u32)ff_mult(11, p) << 24);
177
178 aes_it_tab[0][i] = t;
179 aes_it_tab[1][i] = rol32(t, 8);
180 aes_it_tab[2][i] = rol32(t, 16);
181 aes_it_tab[3][i] = rol32(t, 24);
182 }
183}
184
185#define star_x(x) (((x) & 0x7f7f7f7f) << 1) ^ ((((x) & 0x80808080) >> 7) * 0x1b)
186
187#define imix_col(y, x) \
188 u = star_x(x); \
189 v = star_x(u); \
190 w = star_x(v); \
191 t = w ^ (x); \
192 (y) = u ^ v ^ w; \
193 (y) ^= ror32(u ^ t, 8) ^ \
194 ror32(v ^ t, 16) ^ \
195 ror32(t, 24)
196
197/* initialise the key schedule from the user supplied key */
198
199#define loop4(i) \
200{ \
201 t = ror32(t, 8); t = ls_box(t) ^ rco_tab[i]; \
202 t ^= E_KEY[4 * i]; E_KEY[4 * i + 4] = t; \
203 t ^= E_KEY[4 * i + 1]; E_KEY[4 * i + 5] = t; \
204 t ^= E_KEY[4 * i + 2]; E_KEY[4 * i + 6] = t; \
205 t ^= E_KEY[4 * i + 3]; E_KEY[4 * i + 7] = t; \
206}
207
208#define loop6(i) \
209{ \
210 t = ror32(t, 8); t = ls_box(t) ^ rco_tab[i]; \
211 t ^= E_KEY[6 * i]; E_KEY[6 * i + 6] = t; \
212 t ^= E_KEY[6 * i + 1]; E_KEY[6 * i + 7] = t; \
213 t ^= E_KEY[6 * i + 2]; E_KEY[6 * i + 8] = t; \
214 t ^= E_KEY[6 * i + 3]; E_KEY[6 * i + 9] = t; \
215 t ^= E_KEY[6 * i + 4]; E_KEY[6 * i + 10] = t; \
216 t ^= E_KEY[6 * i + 5]; E_KEY[6 * i + 11] = t; \
217}
218
219#define loop8(i) \
220{ \
221 t = ror32(t, 8); ; t = ls_box(t) ^ rco_tab[i]; \
222 t ^= E_KEY[8 * i]; E_KEY[8 * i + 8] = t; \
223 t ^= E_KEY[8 * i + 1]; E_KEY[8 * i + 9] = t; \
224 t ^= E_KEY[8 * i + 2]; E_KEY[8 * i + 10] = t; \
225 t ^= E_KEY[8 * i + 3]; E_KEY[8 * i + 11] = t; \
226 t = E_KEY[8 * i + 4] ^ ls_box(t); \
227 E_KEY[8 * i + 12] = t; \
228 t ^= E_KEY[8 * i + 5]; E_KEY[8 * i + 13] = t; \
229 t ^= E_KEY[8 * i + 6]; E_KEY[8 * i + 14] = t; \
230 t ^= E_KEY[8 * i + 7]; E_KEY[8 * i + 15] = t; \
231}
232
233static int aes_set_key(void *ctx_arg, const u8 *in_key, unsigned int key_len,
234 u32 *flags)
235{
236 struct aes_ctx *ctx = ctx_arg;
237 u32 i, j, t, u, v, w;
238
239 if (key_len != 16 && key_len != 24 && key_len != 32) {
240 *flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
241 return -EINVAL;
242 }
243
244 ctx->key_length = key_len;
245
246 D_KEY[key_len + 24] = E_KEY[0] = u32_in(in_key);
247 D_KEY[key_len + 25] = E_KEY[1] = u32_in(in_key + 4);
248 D_KEY[key_len + 26] = E_KEY[2] = u32_in(in_key + 8);
249 D_KEY[key_len + 27] = E_KEY[3] = u32_in(in_key + 12);
250
251 switch (key_len) {
252 case 16:
253 t = E_KEY[3];
254 for (i = 0; i < 10; ++i)
255 loop4(i);
256 break;
257
258 case 24:
259 E_KEY[4] = u32_in(in_key + 16);
260 t = E_KEY[5] = u32_in(in_key + 20);
261 for (i = 0; i < 8; ++i)
262 loop6 (i);
263 break;
264
265 case 32:
266 E_KEY[4] = u32_in(in_key + 16);
267 E_KEY[5] = u32_in(in_key + 20);
268 E_KEY[6] = u32_in(in_key + 24);
269 t = E_KEY[7] = u32_in(in_key + 28);
270 for (i = 0; i < 7; ++i)
271 loop8(i);
272 break;
273 }
274
275 D_KEY[0] = E_KEY[key_len + 24];
276 D_KEY[1] = E_KEY[key_len + 25];
277 D_KEY[2] = E_KEY[key_len + 26];
278 D_KEY[3] = E_KEY[key_len + 27];
279
280 for (i = 4; i < key_len + 24; ++i) {
281 j = key_len + 24 - (i & ~3) + (i & 3);
282 imix_col(D_KEY[j], E_KEY[i]);
283 }
284
285 return 0;
286}
287
288extern void aes_encrypt(void *ctx_arg, u8 *out, const u8 *in);
289extern void aes_decrypt(void *ctx_arg, u8 *out, const u8 *in);
290
291static struct crypto_alg aes_alg = {
292 .cra_name = "aes",
293 .cra_flags = CRYPTO_ALG_TYPE_CIPHER,
294 .cra_blocksize = AES_BLOCK_SIZE,
295 .cra_ctxsize = sizeof(struct aes_ctx),
296 .cra_module = THIS_MODULE,
297 .cra_list = LIST_HEAD_INIT(aes_alg.cra_list),
298 .cra_u = {
299 .cipher = {
300 .cia_min_keysize = AES_MIN_KEY_SIZE,
301 .cia_max_keysize = AES_MAX_KEY_SIZE,
302 .cia_setkey = aes_set_key,
303 .cia_encrypt = aes_encrypt,
304 .cia_decrypt = aes_decrypt
305 }
306 }
307};
308
309static int __init aes_init(void)
310{
311 gen_tabs();
312 return crypto_register_alg(&aes_alg);
313}
314
315static void __exit aes_fini(void)
316{
317 crypto_unregister_alg(&aes_alg);
318}
319
320module_init(aes_init);
321module_exit(aes_fini);
322
323MODULE_DESCRIPTION("Rijndael (AES) Cipher Algorithm");
324MODULE_LICENSE("GPL");
generated by cgit v1.2.2 (git 2.25.0) at 2024-12-14 20:17:43 -0500