2 * Glue Code for the AVX assembler implemention of the Cast6 Cipher
4 * Copyright (C) 2012 Johannes Goetzfried
5 * <Johannes.Goetzfried@informatik.stud.uni-erlangen.de>
7 * Copyright © 2013 Jussi Kivilinna <jussi.kivilinna@iki.fi>
9 * This program is free software; you can redistribute it and/or modify
10 * it under the terms of the GNU General Public License as published by
11 * the Free Software Foundation; either version 2 of the License, or
12 * (at your option) any later version.
14 * This program is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 * GNU General Public License for more details.
19 * You should have received a copy of the GNU General Public License
20 * along with this program; if not, write to the Free Software
21 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
26 #include <linux/module.h>
27 #include <linux/hardirq.h>
28 #include <linux/types.h>
29 #include <linux/crypto.h>
30 #include <linux/err.h>
31 #include <crypto/ablk_helper.h>
32 #include <crypto/algapi.h>
33 #include <crypto/cast6.h>
34 #include <crypto/cryptd.h>
35 #include <crypto/b128ops.h>
36 #include <crypto/ctr.h>
37 #include <crypto/lrw.h>
38 #include <crypto/xts.h>
39 #include <asm/fpu/api.h>
40 #include <asm/crypto/glue_helper.h>
42 #define CAST6_PARALLEL_BLOCKS 8
44 asmlinkage
void cast6_ecb_enc_8way(struct cast6_ctx
*ctx
, u8
*dst
,
46 asmlinkage
void cast6_ecb_dec_8way(struct cast6_ctx
*ctx
, u8
*dst
,
49 asmlinkage
void cast6_cbc_dec_8way(struct cast6_ctx
*ctx
, u8
*dst
,
51 asmlinkage
void cast6_ctr_8way(struct cast6_ctx
*ctx
, u8
*dst
, const u8
*src
,
54 asmlinkage
void cast6_xts_enc_8way(struct cast6_ctx
*ctx
, u8
*dst
,
55 const u8
*src
, le128
*iv
);
56 asmlinkage
void cast6_xts_dec_8way(struct cast6_ctx
*ctx
, u8
*dst
,
57 const u8
*src
, le128
*iv
);
59 static void cast6_xts_enc(void *ctx
, u128
*dst
, const u128
*src
, le128
*iv
)
61 glue_xts_crypt_128bit_one(ctx
, dst
, src
, iv
,
62 GLUE_FUNC_CAST(__cast6_encrypt
));
65 static void cast6_xts_dec(void *ctx
, u128
*dst
, const u128
*src
, le128
*iv
)
67 glue_xts_crypt_128bit_one(ctx
, dst
, src
, iv
,
68 GLUE_FUNC_CAST(__cast6_decrypt
));
71 static void cast6_crypt_ctr(void *ctx
, u128
*dst
, const u128
*src
, le128
*iv
)
75 le128_to_be128(&ctrblk
, iv
);
78 __cast6_encrypt(ctx
, (u8
*)&ctrblk
, (u8
*)&ctrblk
);
79 u128_xor(dst
, src
, (u128
*)&ctrblk
);
82 static const struct common_glue_ctx cast6_enc
= {
84 .fpu_blocks_limit
= CAST6_PARALLEL_BLOCKS
,
87 .num_blocks
= CAST6_PARALLEL_BLOCKS
,
88 .fn_u
= { .ecb
= GLUE_FUNC_CAST(cast6_ecb_enc_8way
) }
91 .fn_u
= { .ecb
= GLUE_FUNC_CAST(__cast6_encrypt
) }
95 static const struct common_glue_ctx cast6_ctr
= {
97 .fpu_blocks_limit
= CAST6_PARALLEL_BLOCKS
,
100 .num_blocks
= CAST6_PARALLEL_BLOCKS
,
101 .fn_u
= { .ctr
= GLUE_CTR_FUNC_CAST(cast6_ctr_8way
) }
104 .fn_u
= { .ctr
= GLUE_CTR_FUNC_CAST(cast6_crypt_ctr
) }
108 static const struct common_glue_ctx cast6_enc_xts
= {
110 .fpu_blocks_limit
= CAST6_PARALLEL_BLOCKS
,
113 .num_blocks
= CAST6_PARALLEL_BLOCKS
,
114 .fn_u
= { .xts
= GLUE_XTS_FUNC_CAST(cast6_xts_enc_8way
) }
117 .fn_u
= { .xts
= GLUE_XTS_FUNC_CAST(cast6_xts_enc
) }
121 static const struct common_glue_ctx cast6_dec
= {
123 .fpu_blocks_limit
= CAST6_PARALLEL_BLOCKS
,
126 .num_blocks
= CAST6_PARALLEL_BLOCKS
,
127 .fn_u
= { .ecb
= GLUE_FUNC_CAST(cast6_ecb_dec_8way
) }
130 .fn_u
= { .ecb
= GLUE_FUNC_CAST(__cast6_decrypt
) }
134 static const struct common_glue_ctx cast6_dec_cbc
= {
136 .fpu_blocks_limit
= CAST6_PARALLEL_BLOCKS
,
139 .num_blocks
= CAST6_PARALLEL_BLOCKS
,
140 .fn_u
= { .cbc
= GLUE_CBC_FUNC_CAST(cast6_cbc_dec_8way
) }
143 .fn_u
= { .cbc
= GLUE_CBC_FUNC_CAST(__cast6_decrypt
) }
147 static const struct common_glue_ctx cast6_dec_xts
= {
149 .fpu_blocks_limit
= CAST6_PARALLEL_BLOCKS
,
152 .num_blocks
= CAST6_PARALLEL_BLOCKS
,
153 .fn_u
= { .xts
= GLUE_XTS_FUNC_CAST(cast6_xts_dec_8way
) }
156 .fn_u
= { .xts
= GLUE_XTS_FUNC_CAST(cast6_xts_dec
) }
160 static int ecb_encrypt(struct blkcipher_desc
*desc
, struct scatterlist
*dst
,
161 struct scatterlist
*src
, unsigned int nbytes
)
163 return glue_ecb_crypt_128bit(&cast6_enc
, desc
, dst
, src
, nbytes
);
166 static int ecb_decrypt(struct blkcipher_desc
*desc
, struct scatterlist
*dst
,
167 struct scatterlist
*src
, unsigned int nbytes
)
169 return glue_ecb_crypt_128bit(&cast6_dec
, desc
, dst
, src
, nbytes
);
172 static int cbc_encrypt(struct blkcipher_desc
*desc
, struct scatterlist
*dst
,
173 struct scatterlist
*src
, unsigned int nbytes
)
175 return glue_cbc_encrypt_128bit(GLUE_FUNC_CAST(__cast6_encrypt
), desc
,
179 static int cbc_decrypt(struct blkcipher_desc
*desc
, struct scatterlist
*dst
,
180 struct scatterlist
*src
, unsigned int nbytes
)
182 return glue_cbc_decrypt_128bit(&cast6_dec_cbc
, desc
, dst
, src
,
186 static int ctr_crypt(struct blkcipher_desc
*desc
, struct scatterlist
*dst
,
187 struct scatterlist
*src
, unsigned int nbytes
)
189 return glue_ctr_crypt_128bit(&cast6_ctr
, desc
, dst
, src
, nbytes
);
192 static inline bool cast6_fpu_begin(bool fpu_enabled
, unsigned int nbytes
)
194 return glue_fpu_begin(CAST6_BLOCK_SIZE
, CAST6_PARALLEL_BLOCKS
,
195 NULL
, fpu_enabled
, nbytes
);
198 static inline void cast6_fpu_end(bool fpu_enabled
)
200 glue_fpu_end(fpu_enabled
);
204 struct cast6_ctx
*ctx
;
208 static void encrypt_callback(void *priv
, u8
*srcdst
, unsigned int nbytes
)
210 const unsigned int bsize
= CAST6_BLOCK_SIZE
;
211 struct crypt_priv
*ctx
= priv
;
214 ctx
->fpu_enabled
= cast6_fpu_begin(ctx
->fpu_enabled
, nbytes
);
216 if (nbytes
== bsize
* CAST6_PARALLEL_BLOCKS
) {
217 cast6_ecb_enc_8way(ctx
->ctx
, srcdst
, srcdst
);
221 for (i
= 0; i
< nbytes
/ bsize
; i
++, srcdst
+= bsize
)
222 __cast6_encrypt(ctx
->ctx
, srcdst
, srcdst
);
225 static void decrypt_callback(void *priv
, u8
*srcdst
, unsigned int nbytes
)
227 const unsigned int bsize
= CAST6_BLOCK_SIZE
;
228 struct crypt_priv
*ctx
= priv
;
231 ctx
->fpu_enabled
= cast6_fpu_begin(ctx
->fpu_enabled
, nbytes
);
233 if (nbytes
== bsize
* CAST6_PARALLEL_BLOCKS
) {
234 cast6_ecb_dec_8way(ctx
->ctx
, srcdst
, srcdst
);
238 for (i
= 0; i
< nbytes
/ bsize
; i
++, srcdst
+= bsize
)
239 __cast6_decrypt(ctx
->ctx
, srcdst
, srcdst
);
242 struct cast6_lrw_ctx
{
243 struct lrw_table_ctx lrw_table
;
244 struct cast6_ctx cast6_ctx
;
247 static int lrw_cast6_setkey(struct crypto_tfm
*tfm
, const u8
*key
,
250 struct cast6_lrw_ctx
*ctx
= crypto_tfm_ctx(tfm
);
253 err
= __cast6_setkey(&ctx
->cast6_ctx
, key
, keylen
- CAST6_BLOCK_SIZE
,
258 return lrw_init_table(&ctx
->lrw_table
, key
+ keylen
- CAST6_BLOCK_SIZE
);
261 static int lrw_encrypt(struct blkcipher_desc
*desc
, struct scatterlist
*dst
,
262 struct scatterlist
*src
, unsigned int nbytes
)
264 struct cast6_lrw_ctx
*ctx
= crypto_blkcipher_ctx(desc
->tfm
);
265 be128 buf
[CAST6_PARALLEL_BLOCKS
];
266 struct crypt_priv crypt_ctx
= {
267 .ctx
= &ctx
->cast6_ctx
,
268 .fpu_enabled
= false,
270 struct lrw_crypt_req req
= {
272 .tbuflen
= sizeof(buf
),
274 .table_ctx
= &ctx
->lrw_table
,
275 .crypt_ctx
= &crypt_ctx
,
276 .crypt_fn
= encrypt_callback
,
280 desc
->flags
&= ~CRYPTO_TFM_REQ_MAY_SLEEP
;
281 ret
= lrw_crypt(desc
, dst
, src
, nbytes
, &req
);
282 cast6_fpu_end(crypt_ctx
.fpu_enabled
);
287 static int lrw_decrypt(struct blkcipher_desc
*desc
, struct scatterlist
*dst
,
288 struct scatterlist
*src
, unsigned int nbytes
)
290 struct cast6_lrw_ctx
*ctx
= crypto_blkcipher_ctx(desc
->tfm
);
291 be128 buf
[CAST6_PARALLEL_BLOCKS
];
292 struct crypt_priv crypt_ctx
= {
293 .ctx
= &ctx
->cast6_ctx
,
294 .fpu_enabled
= false,
296 struct lrw_crypt_req req
= {
298 .tbuflen
= sizeof(buf
),
300 .table_ctx
= &ctx
->lrw_table
,
301 .crypt_ctx
= &crypt_ctx
,
302 .crypt_fn
= decrypt_callback
,
306 desc
->flags
&= ~CRYPTO_TFM_REQ_MAY_SLEEP
;
307 ret
= lrw_crypt(desc
, dst
, src
, nbytes
, &req
);
308 cast6_fpu_end(crypt_ctx
.fpu_enabled
);
313 static void lrw_exit_tfm(struct crypto_tfm
*tfm
)
315 struct cast6_lrw_ctx
*ctx
= crypto_tfm_ctx(tfm
);
317 lrw_free_table(&ctx
->lrw_table
);
320 struct cast6_xts_ctx
{
321 struct cast6_ctx tweak_ctx
;
322 struct cast6_ctx crypt_ctx
;
325 static int xts_cast6_setkey(struct crypto_tfm
*tfm
, const u8
*key
,
328 struct cast6_xts_ctx
*ctx
= crypto_tfm_ctx(tfm
);
329 u32
*flags
= &tfm
->crt_flags
;
332 err
= xts_check_key(tfm
, key
, keylen
);
336 /* first half of xts-key is for crypt */
337 err
= __cast6_setkey(&ctx
->crypt_ctx
, key
, keylen
/ 2, flags
);
341 /* second half of xts-key is for tweak */
342 return __cast6_setkey(&ctx
->tweak_ctx
, key
+ keylen
/ 2, keylen
/ 2,
346 static int xts_encrypt(struct blkcipher_desc
*desc
, struct scatterlist
*dst
,
347 struct scatterlist
*src
, unsigned int nbytes
)
349 struct cast6_xts_ctx
*ctx
= crypto_blkcipher_ctx(desc
->tfm
);
351 return glue_xts_crypt_128bit(&cast6_enc_xts
, desc
, dst
, src
, nbytes
,
352 XTS_TWEAK_CAST(__cast6_encrypt
),
353 &ctx
->tweak_ctx
, &ctx
->crypt_ctx
);
356 static int xts_decrypt(struct blkcipher_desc
*desc
, struct scatterlist
*dst
,
357 struct scatterlist
*src
, unsigned int nbytes
)
359 struct cast6_xts_ctx
*ctx
= crypto_blkcipher_ctx(desc
->tfm
);
361 return glue_xts_crypt_128bit(&cast6_dec_xts
, desc
, dst
, src
, nbytes
,
362 XTS_TWEAK_CAST(__cast6_encrypt
),
363 &ctx
->tweak_ctx
, &ctx
->crypt_ctx
);
366 static struct crypto_alg cast6_algs
[10] = { {
367 .cra_name
= "__ecb-cast6-avx",
368 .cra_driver_name
= "__driver-ecb-cast6-avx",
370 .cra_flags
= CRYPTO_ALG_TYPE_BLKCIPHER
|
372 .cra_blocksize
= CAST6_BLOCK_SIZE
,
373 .cra_ctxsize
= sizeof(struct cast6_ctx
),
375 .cra_type
= &crypto_blkcipher_type
,
376 .cra_module
= THIS_MODULE
,
379 .min_keysize
= CAST6_MIN_KEY_SIZE
,
380 .max_keysize
= CAST6_MAX_KEY_SIZE
,
381 .setkey
= cast6_setkey
,
382 .encrypt
= ecb_encrypt
,
383 .decrypt
= ecb_decrypt
,
387 .cra_name
= "__cbc-cast6-avx",
388 .cra_driver_name
= "__driver-cbc-cast6-avx",
390 .cra_flags
= CRYPTO_ALG_TYPE_BLKCIPHER
|
392 .cra_blocksize
= CAST6_BLOCK_SIZE
,
393 .cra_ctxsize
= sizeof(struct cast6_ctx
),
395 .cra_type
= &crypto_blkcipher_type
,
396 .cra_module
= THIS_MODULE
,
399 .min_keysize
= CAST6_MIN_KEY_SIZE
,
400 .max_keysize
= CAST6_MAX_KEY_SIZE
,
401 .setkey
= cast6_setkey
,
402 .encrypt
= cbc_encrypt
,
403 .decrypt
= cbc_decrypt
,
407 .cra_name
= "__ctr-cast6-avx",
408 .cra_driver_name
= "__driver-ctr-cast6-avx",
410 .cra_flags
= CRYPTO_ALG_TYPE_BLKCIPHER
|
413 .cra_ctxsize
= sizeof(struct cast6_ctx
),
415 .cra_type
= &crypto_blkcipher_type
,
416 .cra_module
= THIS_MODULE
,
419 .min_keysize
= CAST6_MIN_KEY_SIZE
,
420 .max_keysize
= CAST6_MAX_KEY_SIZE
,
421 .ivsize
= CAST6_BLOCK_SIZE
,
422 .setkey
= cast6_setkey
,
423 .encrypt
= ctr_crypt
,
424 .decrypt
= ctr_crypt
,
428 .cra_name
= "__lrw-cast6-avx",
429 .cra_driver_name
= "__driver-lrw-cast6-avx",
431 .cra_flags
= CRYPTO_ALG_TYPE_BLKCIPHER
|
433 .cra_blocksize
= CAST6_BLOCK_SIZE
,
434 .cra_ctxsize
= sizeof(struct cast6_lrw_ctx
),
436 .cra_type
= &crypto_blkcipher_type
,
437 .cra_module
= THIS_MODULE
,
438 .cra_exit
= lrw_exit_tfm
,
441 .min_keysize
= CAST6_MIN_KEY_SIZE
+
443 .max_keysize
= CAST6_MAX_KEY_SIZE
+
445 .ivsize
= CAST6_BLOCK_SIZE
,
446 .setkey
= lrw_cast6_setkey
,
447 .encrypt
= lrw_encrypt
,
448 .decrypt
= lrw_decrypt
,
452 .cra_name
= "__xts-cast6-avx",
453 .cra_driver_name
= "__driver-xts-cast6-avx",
455 .cra_flags
= CRYPTO_ALG_TYPE_BLKCIPHER
|
457 .cra_blocksize
= CAST6_BLOCK_SIZE
,
458 .cra_ctxsize
= sizeof(struct cast6_xts_ctx
),
460 .cra_type
= &crypto_blkcipher_type
,
461 .cra_module
= THIS_MODULE
,
464 .min_keysize
= CAST6_MIN_KEY_SIZE
* 2,
465 .max_keysize
= CAST6_MAX_KEY_SIZE
* 2,
466 .ivsize
= CAST6_BLOCK_SIZE
,
467 .setkey
= xts_cast6_setkey
,
468 .encrypt
= xts_encrypt
,
469 .decrypt
= xts_decrypt
,
473 .cra_name
= "ecb(cast6)",
474 .cra_driver_name
= "ecb-cast6-avx",
476 .cra_flags
= CRYPTO_ALG_TYPE_ABLKCIPHER
| CRYPTO_ALG_ASYNC
,
477 .cra_blocksize
= CAST6_BLOCK_SIZE
,
478 .cra_ctxsize
= sizeof(struct async_helper_ctx
),
480 .cra_type
= &crypto_ablkcipher_type
,
481 .cra_module
= THIS_MODULE
,
482 .cra_init
= ablk_init
,
483 .cra_exit
= ablk_exit
,
486 .min_keysize
= CAST6_MIN_KEY_SIZE
,
487 .max_keysize
= CAST6_MAX_KEY_SIZE
,
488 .setkey
= ablk_set_key
,
489 .encrypt
= ablk_encrypt
,
490 .decrypt
= ablk_decrypt
,
494 .cra_name
= "cbc(cast6)",
495 .cra_driver_name
= "cbc-cast6-avx",
497 .cra_flags
= CRYPTO_ALG_TYPE_ABLKCIPHER
| CRYPTO_ALG_ASYNC
,
498 .cra_blocksize
= CAST6_BLOCK_SIZE
,
499 .cra_ctxsize
= sizeof(struct async_helper_ctx
),
501 .cra_type
= &crypto_ablkcipher_type
,
502 .cra_module
= THIS_MODULE
,
503 .cra_init
= ablk_init
,
504 .cra_exit
= ablk_exit
,
507 .min_keysize
= CAST6_MIN_KEY_SIZE
,
508 .max_keysize
= CAST6_MAX_KEY_SIZE
,
509 .ivsize
= CAST6_BLOCK_SIZE
,
510 .setkey
= ablk_set_key
,
511 .encrypt
= __ablk_encrypt
,
512 .decrypt
= ablk_decrypt
,
516 .cra_name
= "ctr(cast6)",
517 .cra_driver_name
= "ctr-cast6-avx",
519 .cra_flags
= CRYPTO_ALG_TYPE_ABLKCIPHER
| CRYPTO_ALG_ASYNC
,
521 .cra_ctxsize
= sizeof(struct async_helper_ctx
),
523 .cra_type
= &crypto_ablkcipher_type
,
524 .cra_module
= THIS_MODULE
,
525 .cra_init
= ablk_init
,
526 .cra_exit
= ablk_exit
,
529 .min_keysize
= CAST6_MIN_KEY_SIZE
,
530 .max_keysize
= CAST6_MAX_KEY_SIZE
,
531 .ivsize
= CAST6_BLOCK_SIZE
,
532 .setkey
= ablk_set_key
,
533 .encrypt
= ablk_encrypt
,
534 .decrypt
= ablk_encrypt
,
539 .cra_name
= "lrw(cast6)",
540 .cra_driver_name
= "lrw-cast6-avx",
542 .cra_flags
= CRYPTO_ALG_TYPE_ABLKCIPHER
| CRYPTO_ALG_ASYNC
,
543 .cra_blocksize
= CAST6_BLOCK_SIZE
,
544 .cra_ctxsize
= sizeof(struct async_helper_ctx
),
546 .cra_type
= &crypto_ablkcipher_type
,
547 .cra_module
= THIS_MODULE
,
548 .cra_init
= ablk_init
,
549 .cra_exit
= ablk_exit
,
552 .min_keysize
= CAST6_MIN_KEY_SIZE
+
554 .max_keysize
= CAST6_MAX_KEY_SIZE
+
556 .ivsize
= CAST6_BLOCK_SIZE
,
557 .setkey
= ablk_set_key
,
558 .encrypt
= ablk_encrypt
,
559 .decrypt
= ablk_decrypt
,
563 .cra_name
= "xts(cast6)",
564 .cra_driver_name
= "xts-cast6-avx",
566 .cra_flags
= CRYPTO_ALG_TYPE_ABLKCIPHER
| CRYPTO_ALG_ASYNC
,
567 .cra_blocksize
= CAST6_BLOCK_SIZE
,
568 .cra_ctxsize
= sizeof(struct async_helper_ctx
),
570 .cra_type
= &crypto_ablkcipher_type
,
571 .cra_module
= THIS_MODULE
,
572 .cra_init
= ablk_init
,
573 .cra_exit
= ablk_exit
,
576 .min_keysize
= CAST6_MIN_KEY_SIZE
* 2,
577 .max_keysize
= CAST6_MAX_KEY_SIZE
* 2,
578 .ivsize
= CAST6_BLOCK_SIZE
,
579 .setkey
= ablk_set_key
,
580 .encrypt
= ablk_encrypt
,
581 .decrypt
= ablk_decrypt
,
586 static int __init
cast6_init(void)
588 const char *feature_name
;
590 if (!cpu_has_xfeatures(XFEATURE_MASK_SSE
| XFEATURE_MASK_YMM
,
592 pr_info("CPU feature '%s' is not supported.\n", feature_name
);
596 return crypto_register_algs(cast6_algs
, ARRAY_SIZE(cast6_algs
));
599 static void __exit
cast6_exit(void)
601 crypto_unregister_algs(cast6_algs
, ARRAY_SIZE(cast6_algs
));
604 module_init(cast6_init
);
605 module_exit(cast6_exit
);
607 MODULE_DESCRIPTION("Cast6 Cipher Algorithm, AVX optimized");
608 MODULE_LICENSE("GPL");
609 MODULE_ALIAS_CRYPTO("cast6");