diff options
-rw-r--r-- | arch/x86/crypto/Makefile | 2 | ||||
-rw-r--r-- | arch/x86/crypto/blowfish-x86_64-asm_64.S | 392 | ||||
-rw-r--r-- | arch/x86/crypto/blowfish_glue.c | 487 | ||||
-rw-r--r-- | crypto/Kconfig | 15 |
4 files changed, 896 insertions, 0 deletions
diff --git a/arch/x86/crypto/Makefile b/arch/x86/crypto/Makefile index 57c7f7b4436d..725addfacf0a 100644 --- a/arch/x86/crypto/Makefile +++ b/arch/x86/crypto/Makefile | |||
@@ -7,6 +7,7 @@ obj-$(CONFIG_CRYPTO_TWOFISH_586) += twofish-i586.o | |||
7 | obj-$(CONFIG_CRYPTO_SALSA20_586) += salsa20-i586.o | 7 | obj-$(CONFIG_CRYPTO_SALSA20_586) += salsa20-i586.o |
8 | 8 | ||
9 | obj-$(CONFIG_CRYPTO_AES_X86_64) += aes-x86_64.o | 9 | obj-$(CONFIG_CRYPTO_AES_X86_64) += aes-x86_64.o |
10 | obj-$(CONFIG_CRYPTO_BLOWFISH_X86_64) += blowfish-x86_64.o | ||
10 | obj-$(CONFIG_CRYPTO_TWOFISH_X86_64) += twofish-x86_64.o | 11 | obj-$(CONFIG_CRYPTO_TWOFISH_X86_64) += twofish-x86_64.o |
11 | obj-$(CONFIG_CRYPTO_SALSA20_X86_64) += salsa20-x86_64.o | 12 | obj-$(CONFIG_CRYPTO_SALSA20_X86_64) += salsa20-x86_64.o |
12 | obj-$(CONFIG_CRYPTO_AES_NI_INTEL) += aesni-intel.o | 13 | obj-$(CONFIG_CRYPTO_AES_NI_INTEL) += aesni-intel.o |
@@ -20,6 +21,7 @@ twofish-i586-y := twofish-i586-asm_32.o twofish_glue.o | |||
20 | salsa20-i586-y := salsa20-i586-asm_32.o salsa20_glue.o | 21 | salsa20-i586-y := salsa20-i586-asm_32.o salsa20_glue.o |
21 | 22 | ||
22 | aes-x86_64-y := aes-x86_64-asm_64.o aes_glue.o | 23 | aes-x86_64-y := aes-x86_64-asm_64.o aes_glue.o |
24 | blowfish-x86_64-y := blowfish-x86_64-asm_64.o blowfish_glue.o | ||
23 | twofish-x86_64-y := twofish-x86_64-asm_64.o twofish_glue.o | 25 | twofish-x86_64-y := twofish-x86_64-asm_64.o twofish_glue.o |
24 | salsa20-x86_64-y := salsa20-x86_64-asm_64.o salsa20_glue.o | 26 | salsa20-x86_64-y := salsa20-x86_64-asm_64.o salsa20_glue.o |
25 | 27 | ||
diff --git a/arch/x86/crypto/blowfish-x86_64-asm_64.S b/arch/x86/crypto/blowfish-x86_64-asm_64.S new file mode 100644 index 000000000000..44eb23ab9676 --- /dev/null +++ b/arch/x86/crypto/blowfish-x86_64-asm_64.S | |||
@@ -0,0 +1,392 @@ | |||
1 | /* | ||
2 | * Blowfish Cipher Algorithm (x86_64) | ||
3 | * | ||
4 | * Copyright (C) 2011 Jussi Kivilinna <jussi.kivilinna@mbnet.fi> | ||
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 as published by | ||
8 | * the Free Software Foundation; either version 2 of the License, or | ||
9 | * (at your option) any later version. | ||
10 | * | ||
11 | * This program is distributed in the hope that it will be useful, | ||
12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
14 | * GNU General Public License for more details. | ||
15 | * | ||
16 | * You should have received a copy of the GNU General Public License | ||
17 | * along with this program; if not, write to the Free Software | ||
18 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 | ||
19 | * USA | ||
20 | * | ||
21 | */ | ||
22 | |||
23 | .file "blowfish-x86_64-asm.S" | ||
24 | .text | ||
25 | |||
26 | /* structure of crypto context */ | ||
27 | #define p 0 | ||
28 | #define s0 ((16 + 2) * 4) | ||
29 | #define s1 ((16 + 2 + (1 * 256)) * 4) | ||
30 | #define s2 ((16 + 2 + (2 * 256)) * 4) | ||
31 | #define s3 ((16 + 2 + (3 * 256)) * 4) | ||
32 | |||
33 | /* register macros */ | ||
34 | #define CTX %rdi | ||
35 | #define RIO %rsi | ||
36 | |||
37 | #define RX0 %rax | ||
38 | #define RX1 %rbx | ||
39 | #define RX2 %rcx | ||
40 | #define RX3 %rdx | ||
41 | |||
42 | #define RX0d %eax | ||
43 | #define RX1d %ebx | ||
44 | #define RX2d %ecx | ||
45 | #define RX3d %edx | ||
46 | |||
47 | #define RX0bl %al | ||
48 | #define RX1bl %bl | ||
49 | #define RX2bl %cl | ||
50 | #define RX3bl %dl | ||
51 | |||
52 | #define RX0bh %ah | ||
53 | #define RX1bh %bh | ||
54 | #define RX2bh %ch | ||
55 | #define RX3bh %dh | ||
56 | |||
57 | #define RT0 %rbp | ||
58 | #define RT1 %rsi | ||
59 | |||
60 | #define RT0d %ebp | ||
61 | #define RT1d %esi | ||
62 | |||
63 | #define RK0 %r8 | ||
64 | #define RK1 %r9 | ||
65 | #define RK2 %r10 | ||
66 | #define RK3 %r11 | ||
67 | |||
68 | #define RK0d %r8d | ||
69 | #define RK1d %r9d | ||
70 | #define RK2d %r10d | ||
71 | #define RK3d %r11d | ||
72 | |||
73 | #define RKEY %r12 | ||
74 | |||
75 | /*********************************************************************** | ||
76 | * 1-way blowfish | ||
77 | ***********************************************************************/ | ||
78 | #define F(x, k) \ | ||
79 | rorq $16, x; \ | ||
80 | movzbl x ## bh, RT0d; \ | ||
81 | movzbl x ## bl, RT1d; \ | ||
82 | rolq $16, x; \ | ||
83 | movl s0(CTX,RT0,4), k ## d; \ | ||
84 | addl s1(CTX,RT1,4), k ## d; \ | ||
85 | movzbl x ## bh, RT0d; \ | ||
86 | movzbl x ## bl, RT1d; \ | ||
87 | rolq $32, x; \ | ||
88 | xorl s2(CTX,RT0,4), k ## d; \ | ||
89 | addl s3(CTX,RT1,4), k ## d; \ | ||
90 | xorq k, x; | ||
91 | |||
92 | #define add_roundkey_enc(n) \ | ||
93 | xorq p+4*(n)(CTX), RX0; | ||
94 | |||
95 | #define round_enc(n) \ | ||
96 | add_roundkey_enc(n); \ | ||
97 | \ | ||
98 | F(RX0, RK0); \ | ||
99 | F(RX0, RK0); | ||
100 | |||
101 | #define round_final_enc(n) \ | ||
102 | xorq p+4*(n)(CTX), RX0; | ||
103 | |||
104 | #define add_roundkey_dec(n) \ | ||
105 | movq p+4*(n-1)(CTX), RT0; \ | ||
106 | rorq $32, RT0; \ | ||
107 | xorq RT0, RX0; | ||
108 | |||
109 | #define round_dec(n) \ | ||
110 | add_roundkey_dec(n); \ | ||
111 | \ | ||
112 | F(RX0, RK0); \ | ||
113 | F(RX0, RK0); \ | ||
114 | |||
115 | #define read_block() \ | ||
116 | movq (RIO), RX0; \ | ||
117 | rorq $32, RX0; \ | ||
118 | bswapq RX0; | ||
119 | |||
120 | #define write_block() \ | ||
121 | bswapq RX0; \ | ||
122 | movq RX0, (RIO); | ||
123 | |||
124 | #define xor_block() \ | ||
125 | bswapq RX0; \ | ||
126 | xorq RX0, (RIO); | ||
127 | |||
128 | .align 8 | ||
129 | .global __blowfish_enc_blk | ||
130 | .type __blowfish_enc_blk,@function; | ||
131 | |||
132 | __blowfish_enc_blk: | ||
133 | // input: | ||
134 | // %rdi: ctx, CTX | ||
135 | // %rsi: dst | ||
136 | // %rdx: src | ||
137 | // %rcx: bool xor | ||
138 | pushq %rbp; | ||
139 | pushq %rbx; | ||
140 | |||
141 | pushq %rsi; | ||
142 | pushq %rcx; | ||
143 | movq %rdx, RIO; | ||
144 | |||
145 | read_block(); | ||
146 | |||
147 | round_enc(0); | ||
148 | round_enc(2); | ||
149 | round_enc(4); | ||
150 | round_enc(6); | ||
151 | round_enc(8); | ||
152 | round_enc(10); | ||
153 | round_enc(12); | ||
154 | round_enc(14); | ||
155 | add_roundkey_enc(16); | ||
156 | |||
157 | popq %rbp; | ||
158 | popq RIO; | ||
159 | |||
160 | test %bpl, %bpl; | ||
161 | jnz __enc_xor; | ||
162 | |||
163 | write_block(); | ||
164 | |||
165 | __enc_ret: | ||
166 | popq %rbx; | ||
167 | popq %rbp; | ||
168 | |||
169 | ret; | ||
170 | |||
171 | __enc_xor: | ||
172 | xor_block(); | ||
173 | |||
174 | jmp __enc_ret; | ||
175 | |||
176 | .align 8 | ||
177 | .global blowfish_dec_blk | ||
178 | .type blowfish_dec_blk,@function; | ||
179 | |||
180 | blowfish_dec_blk: | ||
181 | // input: | ||
182 | // %rdi: ctx, CTX | ||
183 | // %rsi: dst | ||
184 | // %rdx: src | ||
185 | pushq %rbp; | ||
186 | pushq %rbx; | ||
187 | |||
188 | pushq %rsi; | ||
189 | movq %rdx, RIO; | ||
190 | |||
191 | read_block(); | ||
192 | |||
193 | round_dec(17); | ||
194 | round_dec(15); | ||
195 | round_dec(13); | ||
196 | round_dec(11); | ||
197 | round_dec(9); | ||
198 | round_dec(7); | ||
199 | round_dec(5); | ||
200 | round_dec(3); | ||
201 | add_roundkey_dec(1); | ||
202 | |||
203 | popq RIO; | ||
204 | write_block(); | ||
205 | |||
206 | popq %rbx; | ||
207 | popq %rbp; | ||
208 | |||
209 | ret; | ||
210 | |||
211 | /********************************************************************** | ||
212 | 4-way blowfish, four blocks parallel | ||
213 | **********************************************************************/ | ||
214 | #define add_preloaded_roundkey4() \ | ||
215 | xorq RKEY, RX0; \ | ||
216 | xorq RKEY, RX1; \ | ||
217 | xorq RKEY, RX2; \ | ||
218 | xorq RKEY, RX3; | ||
219 | |||
220 | #define preload_roundkey_enc(n) \ | ||
221 | movq p+4*(n)(CTX), RKEY; | ||
222 | |||
223 | #define add_roundkey_enc4(n) \ | ||
224 | add_preloaded_roundkey4(); \ | ||
225 | preload_roundkey_enc(n + 2); | ||
226 | |||
227 | #define round_enc4(n) \ | ||
228 | add_roundkey_enc4(n); \ | ||
229 | \ | ||
230 | F(RX0, RK0); \ | ||
231 | F(RX1, RK1); \ | ||
232 | F(RX2, RK2); \ | ||
233 | F(RX3, RK3); \ | ||
234 | \ | ||
235 | F(RX0, RK0); \ | ||
236 | F(RX1, RK1); \ | ||
237 | F(RX2, RK2); \ | ||
238 | F(RX3, RK3); | ||
239 | |||
240 | #define preload_roundkey_dec(n) \ | ||
241 | movq p+4*((n)-1)(CTX), RKEY; \ | ||
242 | rorq $32, RKEY; | ||
243 | |||
244 | #define add_roundkey_dec4(n) \ | ||
245 | add_preloaded_roundkey4(); \ | ||
246 | preload_roundkey_dec(n - 2); | ||
247 | |||
248 | #define round_dec4(n) \ | ||
249 | add_roundkey_dec4(n); \ | ||
250 | \ | ||
251 | F(RX0, RK0); \ | ||
252 | F(RX1, RK1); \ | ||
253 | F(RX2, RK2); \ | ||
254 | F(RX3, RK3); \ | ||
255 | \ | ||
256 | F(RX0, RK0); \ | ||
257 | F(RX1, RK1); \ | ||
258 | F(RX2, RK2); \ | ||
259 | F(RX3, RK3); | ||
260 | |||
261 | #define read_block4() \ | ||
262 | movq (RIO), RX0; \ | ||
263 | rorq $32, RX0; \ | ||
264 | bswapq RX0; \ | ||
265 | \ | ||
266 | movq 8(RIO), RX1; \ | ||
267 | rorq $32, RX1; \ | ||
268 | bswapq RX1; \ | ||
269 | \ | ||
270 | movq 16(RIO), RX2; \ | ||
271 | rorq $32, RX2; \ | ||
272 | bswapq RX2; \ | ||
273 | \ | ||
274 | movq 24(RIO), RX3; \ | ||
275 | rorq $32, RX3; \ | ||
276 | bswapq RX3; | ||
277 | |||
278 | #define write_block4() \ | ||
279 | bswapq RX0; \ | ||
280 | movq RX0, (RIO); \ | ||
281 | \ | ||
282 | bswapq RX1; \ | ||
283 | movq RX1, 8(RIO); \ | ||
284 | \ | ||
285 | bswapq RX2; \ | ||
286 | movq RX2, 16(RIO); \ | ||
287 | \ | ||
288 | bswapq RX3; \ | ||
289 | movq RX3, 24(RIO); | ||
290 | |||
291 | #define xor_block4() \ | ||
292 | bswapq RX0; \ | ||
293 | xorq RX0, (RIO); \ | ||
294 | \ | ||
295 | bswapq RX1; \ | ||
296 | xorq RX1, 8(RIO); \ | ||
297 | \ | ||
298 | bswapq RX2; \ | ||
299 | xorq RX2, 16(RIO); \ | ||
300 | \ | ||
301 | bswapq RX3; \ | ||
302 | xorq RX3, 24(RIO); | ||
303 | |||
304 | .align 8 | ||
305 | .global __blowfish_enc_blk_4way | ||
306 | .type __blowfish_enc_blk_4way,@function; | ||
307 | |||
308 | __blowfish_enc_blk_4way: | ||
309 | // input: | ||
310 | // %rdi: ctx, CTX | ||
311 | // %rsi: dst | ||
312 | // %rdx: src | ||
313 | // %rcx: bool xor | ||
314 | pushq %rbp; | ||
315 | pushq %rbx; | ||
316 | pushq RKEY; | ||
317 | preload_roundkey_enc(0); | ||
318 | |||
319 | pushq %rsi; | ||
320 | pushq %rcx; | ||
321 | movq %rdx, RIO; | ||
322 | |||
323 | read_block4(); | ||
324 | |||
325 | round_enc4(0); | ||
326 | round_enc4(2); | ||
327 | round_enc4(4); | ||
328 | round_enc4(6); | ||
329 | round_enc4(8); | ||
330 | round_enc4(10); | ||
331 | round_enc4(12); | ||
332 | round_enc4(14); | ||
333 | add_preloaded_roundkey4(); | ||
334 | |||
335 | popq %rbp; | ||
336 | popq RIO; | ||
337 | |||
338 | test %bpl, %bpl; | ||
339 | jnz __enc_xor4; | ||
340 | |||
341 | write_block4(); | ||
342 | |||
343 | __enc_ret4: | ||
344 | popq RKEY; | ||
345 | popq %rbx; | ||
346 | popq %rbp; | ||
347 | |||
348 | ret; | ||
349 | |||
350 | __enc_xor4: | ||
351 | xor_block4(); | ||
352 | |||
353 | jmp __enc_ret4; | ||
354 | |||
355 | .align 8 | ||
356 | .global blowfish_dec_blk_4way | ||
357 | .type blowfish_dec_blk_4way,@function; | ||
358 | |||
359 | blowfish_dec_blk_4way: | ||
360 | // input: | ||
361 | // %rdi: ctx, CTX | ||
362 | // %rsi: dst | ||
363 | // %rdx: src | ||
364 | pushq %rbp; | ||
365 | pushq %rbx; | ||
366 | pushq RKEY; | ||
367 | preload_roundkey_dec(17); | ||
368 | |||
369 | pushq %rsi; | ||
370 | movq %rdx, RIO; | ||
371 | |||
372 | read_block4(); | ||
373 | |||
374 | round_dec4(17); | ||
375 | round_dec4(15); | ||
376 | round_dec4(13); | ||
377 | round_dec4(11); | ||
378 | round_dec4(9); | ||
379 | round_dec4(7); | ||
380 | round_dec4(5); | ||
381 | round_dec4(3); | ||
382 | add_preloaded_roundkey4(); | ||
383 | |||
384 | popq RIO; | ||
385 | write_block4(); | ||
386 | |||
387 | popq RKEY; | ||
388 | popq %rbx; | ||
389 | popq %rbp; | ||
390 | |||
391 | ret; | ||
392 | |||
diff --git a/arch/x86/crypto/blowfish_glue.c b/arch/x86/crypto/blowfish_glue.c new file mode 100644 index 000000000000..40911abb7b11 --- /dev/null +++ b/arch/x86/crypto/blowfish_glue.c | |||
@@ -0,0 +1,487 @@ | |||
1 | /* | ||
2 | * Glue Code for assembler optimized version of Blowfish | ||
3 | * | ||
4 | * Copyright (c) 2011 Jussi Kivilinna <jussi.kivilinna@mbnet.fi> | ||
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 as published by | ||
8 | * the Free Software Foundation; either version 2 of the License, or | ||
9 | * (at your option) any later version. | ||
10 | * | ||
11 | * This program is distributed in the hope that it will be useful, | ||
12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
14 | * GNU General Public License for more details. | ||
15 | * | ||
16 | * You should have received a copy of the GNU General Public License | ||
17 | * along with this program; if not, write to the Free Software | ||
18 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 | ||
19 | * USA | ||
20 | * | ||
21 | */ | ||
22 | |||
23 | #include <crypto/blowfish.h> | ||
24 | #include <linux/crypto.h> | ||
25 | #include <linux/init.h> | ||
26 | #include <linux/module.h> | ||
27 | #include <linux/types.h> | ||
28 | #include <crypto/algapi.h> | ||
29 | |||
30 | /* regular block cipher functions */ | ||
31 | asmlinkage void __blowfish_enc_blk(struct bf_ctx *ctx, u8 *dst, const u8 *src, | ||
32 | bool xor); | ||
33 | asmlinkage void blowfish_dec_blk(struct bf_ctx *ctx, u8 *dst, const u8 *src); | ||
34 | |||
35 | /* 4-way parallel cipher functions */ | ||
36 | asmlinkage void __blowfish_enc_blk_4way(struct bf_ctx *ctx, u8 *dst, | ||
37 | const u8 *src, bool xor); | ||
38 | asmlinkage void blowfish_dec_blk_4way(struct bf_ctx *ctx, u8 *dst, | ||
39 | const u8 *src); | ||
40 | |||
41 | static inline void blowfish_enc_blk(struct bf_ctx *ctx, u8 *dst, const u8 *src) | ||
42 | { | ||
43 | __blowfish_enc_blk(ctx, dst, src, false); | ||
44 | } | ||
45 | |||
46 | static inline void blowfish_enc_blk_xor(struct bf_ctx *ctx, u8 *dst, | ||
47 | const u8 *src) | ||
48 | { | ||
49 | __blowfish_enc_blk(ctx, dst, src, true); | ||
50 | } | ||
51 | |||
52 | static inline void blowfish_enc_blk_4way(struct bf_ctx *ctx, u8 *dst, | ||
53 | const u8 *src) | ||
54 | { | ||
55 | __blowfish_enc_blk_4way(ctx, dst, src, false); | ||
56 | } | ||
57 | |||
58 | static inline void blowfish_enc_blk_xor_4way(struct bf_ctx *ctx, u8 *dst, | ||
59 | const u8 *src) | ||
60 | { | ||
61 | __blowfish_enc_blk_4way(ctx, dst, src, true); | ||
62 | } | ||
63 | |||
64 | static void blowfish_encrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src) | ||
65 | { | ||
66 | blowfish_enc_blk(crypto_tfm_ctx(tfm), dst, src); | ||
67 | } | ||
68 | |||
69 | static void blowfish_decrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src) | ||
70 | { | ||
71 | blowfish_dec_blk(crypto_tfm_ctx(tfm), dst, src); | ||
72 | } | ||
73 | |||
74 | static struct crypto_alg bf_alg = { | ||
75 | .cra_name = "blowfish", | ||
76 | .cra_driver_name = "blowfish-asm", | ||
77 | .cra_priority = 200, | ||
78 | .cra_flags = CRYPTO_ALG_TYPE_CIPHER, | ||
79 | .cra_blocksize = BF_BLOCK_SIZE, | ||
80 | .cra_ctxsize = sizeof(struct bf_ctx), | ||
81 | .cra_alignmask = 3, | ||
82 | .cra_module = THIS_MODULE, | ||
83 | .cra_list = LIST_HEAD_INIT(bf_alg.cra_list), | ||
84 | .cra_u = { | ||
85 | .cipher = { | ||
86 | .cia_min_keysize = BF_MIN_KEY_SIZE, | ||
87 | .cia_max_keysize = BF_MAX_KEY_SIZE, | ||
88 | .cia_setkey = blowfish_setkey, | ||
89 | .cia_encrypt = blowfish_encrypt, | ||
90 | .cia_decrypt = blowfish_decrypt, | ||
91 | } | ||
92 | } | ||
93 | }; | ||
94 | |||
95 | static int ecb_crypt(struct blkcipher_desc *desc, struct blkcipher_walk *walk, | ||
96 | void (*fn)(struct bf_ctx *, u8 *, const u8 *), | ||
97 | void (*fn_4way)(struct bf_ctx *, u8 *, const u8 *)) | ||
98 | { | ||
99 | struct bf_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); | ||
100 | unsigned int bsize = BF_BLOCK_SIZE; | ||
101 | unsigned int nbytes; | ||
102 | int err; | ||
103 | |||
104 | err = blkcipher_walk_virt(desc, walk); | ||
105 | |||
106 | while ((nbytes = walk->nbytes)) { | ||
107 | u8 *wsrc = walk->src.virt.addr; | ||
108 | u8 *wdst = walk->dst.virt.addr; | ||
109 | |||
110 | /* Process four block batch */ | ||
111 | if (nbytes >= bsize * 4) { | ||
112 | do { | ||
113 | fn_4way(ctx, wdst, wsrc); | ||
114 | |||
115 | wsrc += bsize * 4; | ||
116 | wdst += bsize * 4; | ||
117 | nbytes -= bsize * 4; | ||
118 | } while (nbytes >= bsize * 4); | ||
119 | |||
120 | if (nbytes < bsize) | ||
121 | goto done; | ||
122 | } | ||
123 | |||
124 | /* Handle leftovers */ | ||
125 | do { | ||
126 | fn(ctx, wdst, wsrc); | ||
127 | |||
128 | wsrc += bsize; | ||
129 | wdst += bsize; | ||
130 | nbytes -= bsize; | ||
131 | } while (nbytes >= bsize); | ||
132 | |||
133 | done: | ||
134 | err = blkcipher_walk_done(desc, walk, nbytes); | ||
135 | } | ||
136 | |||
137 | return err; | ||
138 | } | ||
139 | |||
140 | static int ecb_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, | ||
141 | struct scatterlist *src, unsigned int nbytes) | ||
142 | { | ||
143 | struct blkcipher_walk walk; | ||
144 | |||
145 | blkcipher_walk_init(&walk, dst, src, nbytes); | ||
146 | return ecb_crypt(desc, &walk, blowfish_enc_blk, blowfish_enc_blk_4way); | ||
147 | } | ||
148 | |||
149 | static int ecb_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, | ||
150 | struct scatterlist *src, unsigned int nbytes) | ||
151 | { | ||
152 | struct blkcipher_walk walk; | ||
153 | |||
154 | blkcipher_walk_init(&walk, dst, src, nbytes); | ||
155 | return ecb_crypt(desc, &walk, blowfish_dec_blk, blowfish_dec_blk_4way); | ||
156 | } | ||
157 | |||
158 | static struct crypto_alg blk_ecb_alg = { | ||
159 | .cra_name = "ecb(blowfish)", | ||
160 | .cra_driver_name = "ecb-blowfish-asm", | ||
161 | .cra_priority = 300, | ||
162 | .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, | ||
163 | .cra_blocksize = BF_BLOCK_SIZE, | ||
164 | .cra_ctxsize = sizeof(struct bf_ctx), | ||
165 | .cra_alignmask = 0, | ||
166 | .cra_type = &crypto_blkcipher_type, | ||
167 | .cra_module = THIS_MODULE, | ||
168 | .cra_list = LIST_HEAD_INIT(blk_ecb_alg.cra_list), | ||
169 | .cra_u = { | ||
170 | .blkcipher = { | ||
171 | .min_keysize = BF_MIN_KEY_SIZE, | ||
172 | .max_keysize = BF_MAX_KEY_SIZE, | ||
173 | .setkey = blowfish_setkey, | ||
174 | .encrypt = ecb_encrypt, | ||
175 | .decrypt = ecb_decrypt, | ||
176 | }, | ||
177 | }, | ||
178 | }; | ||
179 | |||
180 | static unsigned int __cbc_encrypt(struct blkcipher_desc *desc, | ||
181 | struct blkcipher_walk *walk) | ||
182 | { | ||
183 | struct bf_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); | ||
184 | unsigned int bsize = BF_BLOCK_SIZE; | ||
185 | unsigned int nbytes = walk->nbytes; | ||
186 | u64 *src = (u64 *)walk->src.virt.addr; | ||
187 | u64 *dst = (u64 *)walk->dst.virt.addr; | ||
188 | u64 *iv = (u64 *)walk->iv; | ||
189 | |||
190 | do { | ||
191 | *dst = *src ^ *iv; | ||
192 | blowfish_enc_blk(ctx, (u8 *)dst, (u8 *)dst); | ||
193 | iv = dst; | ||
194 | |||
195 | src += 1; | ||
196 | dst += 1; | ||
197 | nbytes -= bsize; | ||
198 | } while (nbytes >= bsize); | ||
199 | |||
200 | *(u64 *)walk->iv = *iv; | ||
201 | return nbytes; | ||
202 | } | ||
203 | |||
204 | static int cbc_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, | ||
205 | struct scatterlist *src, unsigned int nbytes) | ||
206 | { | ||
207 | struct blkcipher_walk walk; | ||
208 | int err; | ||
209 | |||
210 | blkcipher_walk_init(&walk, dst, src, nbytes); | ||
211 | err = blkcipher_walk_virt(desc, &walk); | ||
212 | |||
213 | while ((nbytes = walk.nbytes)) { | ||
214 | nbytes = __cbc_encrypt(desc, &walk); | ||
215 | err = blkcipher_walk_done(desc, &walk, nbytes); | ||
216 | } | ||
217 | |||
218 | return err; | ||
219 | } | ||
220 | |||
221 | static unsigned int __cbc_decrypt(struct blkcipher_desc *desc, | ||
222 | struct blkcipher_walk *walk) | ||
223 | { | ||
224 | struct bf_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); | ||
225 | unsigned int bsize = BF_BLOCK_SIZE; | ||
226 | unsigned int nbytes = walk->nbytes; | ||
227 | u64 *src = (u64 *)walk->src.virt.addr; | ||
228 | u64 *dst = (u64 *)walk->dst.virt.addr; | ||
229 | u64 ivs[4 - 1]; | ||
230 | u64 last_iv; | ||
231 | |||
232 | /* Start of the last block. */ | ||
233 | src += nbytes / bsize - 1; | ||
234 | dst += nbytes / bsize - 1; | ||
235 | |||
236 | last_iv = *src; | ||
237 | |||
238 | /* Process four block batch */ | ||
239 | if (nbytes >= bsize * 4) { | ||
240 | do { | ||
241 | nbytes -= bsize * 4 - bsize; | ||
242 | src -= 4 - 1; | ||
243 | dst -= 4 - 1; | ||
244 | |||
245 | ivs[0] = src[0]; | ||
246 | ivs[1] = src[1]; | ||
247 | ivs[2] = src[2]; | ||
248 | |||
249 | blowfish_dec_blk_4way(ctx, (u8 *)dst, (u8 *)src); | ||
250 | |||
251 | dst[1] ^= ivs[0]; | ||
252 | dst[2] ^= ivs[1]; | ||
253 | dst[3] ^= ivs[2]; | ||
254 | |||
255 | nbytes -= bsize; | ||
256 | if (nbytes < bsize) | ||
257 | goto done; | ||
258 | |||
259 | *dst ^= *(src - 1); | ||
260 | src -= 1; | ||
261 | dst -= 1; | ||
262 | } while (nbytes >= bsize * 4); | ||
263 | |||
264 | if (nbytes < bsize) | ||
265 | goto done; | ||
266 | } | ||
267 | |||
268 | /* Handle leftovers */ | ||
269 | for (;;) { | ||
270 | blowfish_dec_blk(ctx, (u8 *)dst, (u8 *)src); | ||
271 | |||
272 | nbytes -= bsize; | ||
273 | if (nbytes < bsize) | ||
274 | break; | ||
275 | |||
276 | *dst ^= *(src - 1); | ||
277 | src -= 1; | ||
278 | dst -= 1; | ||
279 | } | ||
280 | |||
281 | done: | ||
282 | *dst ^= *(u64 *)walk->iv; | ||
283 | *(u64 *)walk->iv = last_iv; | ||
284 | |||
285 | return nbytes; | ||
286 | } | ||
287 | |||
288 | static int cbc_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, | ||
289 | struct scatterlist *src, unsigned int nbytes) | ||
290 | { | ||
291 | struct blkcipher_walk walk; | ||
292 | int err; | ||
293 | |||
294 | blkcipher_walk_init(&walk, dst, src, nbytes); | ||
295 | err = blkcipher_walk_virt(desc, &walk); | ||
296 | |||
297 | while ((nbytes = walk.nbytes)) { | ||
298 | nbytes = __cbc_decrypt(desc, &walk); | ||
299 | err = blkcipher_walk_done(desc, &walk, nbytes); | ||
300 | } | ||
301 | |||
302 | return err; | ||
303 | } | ||
304 | |||
305 | static struct crypto_alg blk_cbc_alg = { | ||
306 | .cra_name = "cbc(blowfish)", | ||
307 | .cra_driver_name = "cbc-blowfish-asm", | ||
308 | .cra_priority = 300, | ||
309 | .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, | ||
310 | .cra_blocksize = BF_BLOCK_SIZE, | ||
311 | .cra_ctxsize = sizeof(struct bf_ctx), | ||
312 | .cra_alignmask = 0, | ||
313 | .cra_type = &crypto_blkcipher_type, | ||
314 | .cra_module = THIS_MODULE, | ||
315 | .cra_list = LIST_HEAD_INIT(blk_cbc_alg.cra_list), | ||
316 | .cra_u = { | ||
317 | .blkcipher = { | ||
318 | .min_keysize = BF_MIN_KEY_SIZE, | ||
319 | .max_keysize = BF_MAX_KEY_SIZE, | ||
320 | .ivsize = BF_BLOCK_SIZE, | ||
321 | .setkey = blowfish_setkey, | ||
322 | .encrypt = cbc_encrypt, | ||
323 | .decrypt = cbc_decrypt, | ||
324 | }, | ||
325 | }, | ||
326 | }; | ||
327 | |||
328 | static void ctr_crypt_final(struct bf_ctx *ctx, struct blkcipher_walk *walk) | ||
329 | { | ||
330 | u8 *ctrblk = walk->iv; | ||
331 | u8 keystream[BF_BLOCK_SIZE]; | ||
332 | u8 *src = walk->src.virt.addr; | ||
333 | u8 *dst = walk->dst.virt.addr; | ||
334 | unsigned int nbytes = walk->nbytes; | ||
335 | |||
336 | blowfish_enc_blk(ctx, keystream, ctrblk); | ||
337 | crypto_xor(keystream, src, nbytes); | ||
338 | memcpy(dst, keystream, nbytes); | ||
339 | |||
340 | crypto_inc(ctrblk, BF_BLOCK_SIZE); | ||
341 | } | ||
342 | |||
343 | static unsigned int __ctr_crypt(struct blkcipher_desc *desc, | ||
344 | struct blkcipher_walk *walk) | ||
345 | { | ||
346 | struct bf_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); | ||
347 | unsigned int bsize = BF_BLOCK_SIZE; | ||
348 | unsigned int nbytes = walk->nbytes; | ||
349 | u64 *src = (u64 *)walk->src.virt.addr; | ||
350 | u64 *dst = (u64 *)walk->dst.virt.addr; | ||
351 | u64 ctrblk = be64_to_cpu(*(__be64 *)walk->iv); | ||
352 | __be64 ctrblocks[4]; | ||
353 | |||
354 | /* Process four block batch */ | ||
355 | if (nbytes >= bsize * 4) { | ||
356 | do { | ||
357 | if (dst != src) { | ||
358 | dst[0] = src[0]; | ||
359 | dst[1] = src[1]; | ||
360 | dst[2] = src[2]; | ||
361 | dst[3] = src[3]; | ||
362 | } | ||
363 | |||
364 | /* create ctrblks for parallel encrypt */ | ||
365 | ctrblocks[0] = cpu_to_be64(ctrblk++); | ||
366 | ctrblocks[1] = cpu_to_be64(ctrblk++); | ||
367 | ctrblocks[2] = cpu_to_be64(ctrblk++); | ||
368 | ctrblocks[3] = cpu_to_be64(ctrblk++); | ||
369 | |||
370 | blowfish_enc_blk_xor_4way(ctx, (u8 *)dst, | ||
371 | (u8 *)ctrblocks); | ||
372 | |||
373 | src += 4; | ||
374 | dst += 4; | ||
375 | } while ((nbytes -= bsize * 4) >= bsize * 4); | ||
376 | |||
377 | if (nbytes < bsize) | ||
378 | goto done; | ||
379 | } | ||
380 | |||
381 | /* Handle leftovers */ | ||
382 | do { | ||
383 | if (dst != src) | ||
384 | *dst = *src; | ||
385 | |||
386 | ctrblocks[0] = cpu_to_be64(ctrblk++); | ||
387 | |||
388 | blowfish_enc_blk_xor(ctx, (u8 *)dst, (u8 *)ctrblocks); | ||
389 | |||
390 | src += 1; | ||
391 | dst += 1; | ||
392 | } while ((nbytes -= bsize) >= bsize); | ||
393 | |||
394 | done: | ||
395 | *(__be64 *)walk->iv = cpu_to_be64(ctrblk); | ||
396 | return nbytes; | ||
397 | } | ||
398 | |||
399 | static int ctr_crypt(struct blkcipher_desc *desc, struct scatterlist *dst, | ||
400 | struct scatterlist *src, unsigned int nbytes) | ||
401 | { | ||
402 | struct blkcipher_walk walk; | ||
403 | int err; | ||
404 | |||
405 | blkcipher_walk_init(&walk, dst, src, nbytes); | ||
406 | err = blkcipher_walk_virt_block(desc, &walk, BF_BLOCK_SIZE); | ||
407 | |||
408 | while ((nbytes = walk.nbytes) >= BF_BLOCK_SIZE) { | ||
409 | nbytes = __ctr_crypt(desc, &walk); | ||
410 | err = blkcipher_walk_done(desc, &walk, nbytes); | ||
411 | } | ||
412 | |||
413 | if (walk.nbytes) { | ||
414 | ctr_crypt_final(crypto_blkcipher_ctx(desc->tfm), &walk); | ||
415 | err = blkcipher_walk_done(desc, &walk, 0); | ||
416 | } | ||
417 | |||
418 | return err; | ||
419 | } | ||
420 | |||
421 | static struct crypto_alg blk_ctr_alg = { | ||
422 | .cra_name = "ctr(blowfish)", | ||
423 | .cra_driver_name = "ctr-blowfish-asm", | ||
424 | .cra_priority = 300, | ||
425 | .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, | ||
426 | .cra_blocksize = BF_BLOCK_SIZE, | ||
427 | .cra_ctxsize = sizeof(struct bf_ctx), | ||
428 | .cra_alignmask = 0, | ||
429 | .cra_type = &crypto_blkcipher_type, | ||
430 | .cra_module = THIS_MODULE, | ||
431 | .cra_list = LIST_HEAD_INIT(blk_ctr_alg.cra_list), | ||
432 | .cra_u = { | ||
433 | .blkcipher = { | ||
434 | .min_keysize = BF_MIN_KEY_SIZE, | ||
435 | .max_keysize = BF_MAX_KEY_SIZE, | ||
436 | .ivsize = BF_BLOCK_SIZE, | ||
437 | .setkey = blowfish_setkey, | ||
438 | .encrypt = ctr_crypt, | ||
439 | .decrypt = ctr_crypt, | ||
440 | }, | ||
441 | }, | ||
442 | }; | ||
443 | |||
444 | static int __init init(void) | ||
445 | { | ||
446 | int err; | ||
447 | |||
448 | err = crypto_register_alg(&bf_alg); | ||
449 | if (err) | ||
450 | goto bf_err; | ||
451 | err = crypto_register_alg(&blk_ecb_alg); | ||
452 | if (err) | ||
453 | goto ecb_err; | ||
454 | err = crypto_register_alg(&blk_cbc_alg); | ||
455 | if (err) | ||
456 | goto cbc_err; | ||
457 | err = crypto_register_alg(&blk_ctr_alg); | ||
458 | if (err) | ||
459 | goto ctr_err; | ||
460 | |||
461 | return 0; | ||
462 | |||
463 | ctr_err: | ||
464 | crypto_unregister_alg(&blk_cbc_alg); | ||
465 | cbc_err: | ||
466 | crypto_unregister_alg(&blk_ecb_alg); | ||
467 | ecb_err: | ||
468 | crypto_unregister_alg(&bf_alg); | ||
469 | bf_err: | ||
470 | return err; | ||
471 | } | ||
472 | |||
473 | static void __exit fini(void) | ||
474 | { | ||
475 | crypto_unregister_alg(&blk_ctr_alg); | ||
476 | crypto_unregister_alg(&blk_cbc_alg); | ||
477 | crypto_unregister_alg(&blk_ecb_alg); | ||
478 | crypto_unregister_alg(&bf_alg); | ||
479 | } | ||
480 | |||
481 | module_init(init); | ||
482 | module_exit(fini); | ||
483 | |||
484 | MODULE_LICENSE("GPL"); | ||
485 | MODULE_DESCRIPTION("Blowfish Cipher Algorithm, asm optimized"); | ||
486 | MODULE_ALIAS("blowfish"); | ||
487 | MODULE_ALIAS("blowfish-asm"); | ||
diff --git a/crypto/Kconfig b/crypto/Kconfig index 108cb98e2170..07637745e234 100644 --- a/crypto/Kconfig +++ b/crypto/Kconfig | |||
@@ -620,6 +620,21 @@ config CRYPTO_BLOWFISH_COMMON | |||
620 | See also: | 620 | See also: |
621 | <http://www.schneier.com/blowfish.html> | 621 | <http://www.schneier.com/blowfish.html> |
622 | 622 | ||
623 | config CRYPTO_BLOWFISH_X86_64 | ||
624 | tristate "Blowfish cipher algorithm (x86_64)" | ||
625 | depends on (X86 || UML_X86) && 64BIT | ||
626 | select CRYPTO_ALGAPI | ||
627 | select CRYPTO_BLOWFISH_COMMON | ||
628 | help | ||
629 | Blowfish cipher algorithm (x86_64), by Bruce Schneier. | ||
630 | |||
631 | This is a variable key length cipher which can use keys from 32 | ||
632 | bits to 448 bits in length. It's fast, simple and specifically | ||
633 | designed for use on "large microprocessors". | ||
634 | |||
635 | See also: | ||
636 | <http://www.schneier.com/blowfish.html> | ||
637 | |||
623 | config CRYPTO_CAMELLIA | 638 | config CRYPTO_CAMELLIA |
624 | tristate "Camellia cipher algorithms" | 639 | tristate "Camellia cipher algorithms" |
625 | depends on CRYPTO | 640 | depends on CRYPTO |