Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 1 | /* |
| 2 | * Glue Code for the AVX assembler implemention of the Cast6 Cipher |
| 3 | * |
| 4 | * Copyright (C) 2012 Johannes Goetzfried |
| 5 | * <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> |
| 6 | * |
Jussi Kivilinna | 7017728 | 2013-04-08 21:51:05 +0300 | [diff] [blame] | 7 | * Copyright © 2013 Jussi Kivilinna <jussi.kivilinna@iki.fi> |
| 8 | * |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 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. |
| 13 | * |
| 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. |
| 18 | * |
| 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 |
| 22 | * USA |
| 23 | * |
| 24 | */ |
| 25 | |
| 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/algapi.h> |
| 32 | #include <crypto/cast6.h> |
| 33 | #include <crypto/cryptd.h> |
| 34 | #include <crypto/b128ops.h> |
| 35 | #include <crypto/ctr.h> |
| 36 | #include <crypto/lrw.h> |
| 37 | #include <crypto/xts.h> |
| 38 | #include <asm/xcr.h> |
| 39 | #include <asm/xsave.h> |
| 40 | #include <asm/crypto/ablk_helper.h> |
| 41 | #include <asm/crypto/glue_helper.h> |
| 42 | |
| 43 | #define CAST6_PARALLEL_BLOCKS 8 |
| 44 | |
Jussi Kivilinna | cba1cce | 2012-10-20 15:06:41 +0300 | [diff] [blame] | 45 | asmlinkage void cast6_ecb_enc_8way(struct cast6_ctx *ctx, u8 *dst, |
| 46 | const u8 *src); |
| 47 | asmlinkage void cast6_ecb_dec_8way(struct cast6_ctx *ctx, u8 *dst, |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 48 | const u8 *src); |
| 49 | |
Jussi Kivilinna | cba1cce | 2012-10-20 15:06:41 +0300 | [diff] [blame] | 50 | asmlinkage void cast6_cbc_dec_8way(struct cast6_ctx *ctx, u8 *dst, |
| 51 | const u8 *src); |
| 52 | asmlinkage void cast6_ctr_8way(struct cast6_ctx *ctx, u8 *dst, const u8 *src, |
| 53 | le128 *iv); |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 54 | |
Jussi Kivilinna | 7017728 | 2013-04-08 21:51:05 +0300 | [diff] [blame] | 55 | asmlinkage void cast6_xts_enc_8way(struct cast6_ctx *ctx, u8 *dst, |
| 56 | const u8 *src, le128 *iv); |
| 57 | asmlinkage void cast6_xts_dec_8way(struct cast6_ctx *ctx, u8 *dst, |
| 58 | const u8 *src, le128 *iv); |
| 59 | |
| 60 | static void cast6_xts_enc(void *ctx, u128 *dst, const u128 *src, le128 *iv) |
| 61 | { |
| 62 | glue_xts_crypt_128bit_one(ctx, dst, src, iv, |
| 63 | GLUE_FUNC_CAST(__cast6_encrypt)); |
| 64 | } |
| 65 | |
| 66 | static void cast6_xts_dec(void *ctx, u128 *dst, const u128 *src, le128 *iv) |
| 67 | { |
| 68 | glue_xts_crypt_128bit_one(ctx, dst, src, iv, |
| 69 | GLUE_FUNC_CAST(__cast6_decrypt)); |
| 70 | } |
| 71 | |
Jussi Kivilinna | 5899098 | 2012-10-20 15:06:36 +0300 | [diff] [blame] | 72 | static void cast6_crypt_ctr(void *ctx, u128 *dst, const u128 *src, le128 *iv) |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 73 | { |
| 74 | be128 ctrblk; |
| 75 | |
Jussi Kivilinna | 5899098 | 2012-10-20 15:06:36 +0300 | [diff] [blame] | 76 | le128_to_be128(&ctrblk, iv); |
| 77 | le128_inc(iv); |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 78 | |
| 79 | __cast6_encrypt(ctx, (u8 *)&ctrblk, (u8 *)&ctrblk); |
| 80 | u128_xor(dst, src, (u128 *)&ctrblk); |
| 81 | } |
| 82 | |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 83 | static const struct common_glue_ctx cast6_enc = { |
| 84 | .num_funcs = 2, |
| 85 | .fpu_blocks_limit = CAST6_PARALLEL_BLOCKS, |
| 86 | |
| 87 | .funcs = { { |
| 88 | .num_blocks = CAST6_PARALLEL_BLOCKS, |
Jussi Kivilinna | cba1cce | 2012-10-20 15:06:41 +0300 | [diff] [blame] | 89 | .fn_u = { .ecb = GLUE_FUNC_CAST(cast6_ecb_enc_8way) } |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 90 | }, { |
| 91 | .num_blocks = 1, |
| 92 | .fn_u = { .ecb = GLUE_FUNC_CAST(__cast6_encrypt) } |
| 93 | } } |
| 94 | }; |
| 95 | |
| 96 | static const struct common_glue_ctx cast6_ctr = { |
| 97 | .num_funcs = 2, |
| 98 | .fpu_blocks_limit = CAST6_PARALLEL_BLOCKS, |
| 99 | |
| 100 | .funcs = { { |
| 101 | .num_blocks = CAST6_PARALLEL_BLOCKS, |
Jussi Kivilinna | cba1cce | 2012-10-20 15:06:41 +0300 | [diff] [blame] | 102 | .fn_u = { .ctr = GLUE_CTR_FUNC_CAST(cast6_ctr_8way) } |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 103 | }, { |
| 104 | .num_blocks = 1, |
| 105 | .fn_u = { .ctr = GLUE_CTR_FUNC_CAST(cast6_crypt_ctr) } |
| 106 | } } |
| 107 | }; |
| 108 | |
Jussi Kivilinna | 7017728 | 2013-04-08 21:51:05 +0300 | [diff] [blame] | 109 | static const struct common_glue_ctx cast6_enc_xts = { |
| 110 | .num_funcs = 2, |
| 111 | .fpu_blocks_limit = CAST6_PARALLEL_BLOCKS, |
| 112 | |
| 113 | .funcs = { { |
| 114 | .num_blocks = CAST6_PARALLEL_BLOCKS, |
| 115 | .fn_u = { .xts = GLUE_XTS_FUNC_CAST(cast6_xts_enc_8way) } |
| 116 | }, { |
| 117 | .num_blocks = 1, |
| 118 | .fn_u = { .xts = GLUE_XTS_FUNC_CAST(cast6_xts_enc) } |
| 119 | } } |
| 120 | }; |
| 121 | |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 122 | static const struct common_glue_ctx cast6_dec = { |
| 123 | .num_funcs = 2, |
| 124 | .fpu_blocks_limit = CAST6_PARALLEL_BLOCKS, |
| 125 | |
| 126 | .funcs = { { |
| 127 | .num_blocks = CAST6_PARALLEL_BLOCKS, |
Jussi Kivilinna | cba1cce | 2012-10-20 15:06:41 +0300 | [diff] [blame] | 128 | .fn_u = { .ecb = GLUE_FUNC_CAST(cast6_ecb_dec_8way) } |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 129 | }, { |
| 130 | .num_blocks = 1, |
| 131 | .fn_u = { .ecb = GLUE_FUNC_CAST(__cast6_decrypt) } |
| 132 | } } |
| 133 | }; |
| 134 | |
| 135 | static const struct common_glue_ctx cast6_dec_cbc = { |
| 136 | .num_funcs = 2, |
| 137 | .fpu_blocks_limit = CAST6_PARALLEL_BLOCKS, |
| 138 | |
| 139 | .funcs = { { |
| 140 | .num_blocks = CAST6_PARALLEL_BLOCKS, |
Jussi Kivilinna | cba1cce | 2012-10-20 15:06:41 +0300 | [diff] [blame] | 141 | .fn_u = { .cbc = GLUE_CBC_FUNC_CAST(cast6_cbc_dec_8way) } |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 142 | }, { |
| 143 | .num_blocks = 1, |
| 144 | .fn_u = { .cbc = GLUE_CBC_FUNC_CAST(__cast6_decrypt) } |
| 145 | } } |
| 146 | }; |
| 147 | |
Jussi Kivilinna | 7017728 | 2013-04-08 21:51:05 +0300 | [diff] [blame] | 148 | static const struct common_glue_ctx cast6_dec_xts = { |
| 149 | .num_funcs = 2, |
| 150 | .fpu_blocks_limit = CAST6_PARALLEL_BLOCKS, |
| 151 | |
| 152 | .funcs = { { |
| 153 | .num_blocks = CAST6_PARALLEL_BLOCKS, |
| 154 | .fn_u = { .xts = GLUE_XTS_FUNC_CAST(cast6_xts_dec_8way) } |
| 155 | }, { |
| 156 | .num_blocks = 1, |
| 157 | .fn_u = { .xts = GLUE_XTS_FUNC_CAST(cast6_xts_dec) } |
| 158 | } } |
| 159 | }; |
| 160 | |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 161 | static int ecb_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, |
| 162 | struct scatterlist *src, unsigned int nbytes) |
| 163 | { |
| 164 | return glue_ecb_crypt_128bit(&cast6_enc, desc, dst, src, nbytes); |
| 165 | } |
| 166 | |
| 167 | static int ecb_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, |
| 168 | struct scatterlist *src, unsigned int nbytes) |
| 169 | { |
| 170 | return glue_ecb_crypt_128bit(&cast6_dec, desc, dst, src, nbytes); |
| 171 | } |
| 172 | |
| 173 | static int cbc_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, |
| 174 | struct scatterlist *src, unsigned int nbytes) |
| 175 | { |
| 176 | return glue_cbc_encrypt_128bit(GLUE_FUNC_CAST(__cast6_encrypt), desc, |
| 177 | dst, src, nbytes); |
| 178 | } |
| 179 | |
| 180 | static int cbc_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, |
| 181 | struct scatterlist *src, unsigned int nbytes) |
| 182 | { |
| 183 | return glue_cbc_decrypt_128bit(&cast6_dec_cbc, desc, dst, src, |
| 184 | nbytes); |
| 185 | } |
| 186 | |
| 187 | static int ctr_crypt(struct blkcipher_desc *desc, struct scatterlist *dst, |
| 188 | struct scatterlist *src, unsigned int nbytes) |
| 189 | { |
| 190 | return glue_ctr_crypt_128bit(&cast6_ctr, desc, dst, src, nbytes); |
| 191 | } |
| 192 | |
| 193 | static inline bool cast6_fpu_begin(bool fpu_enabled, unsigned int nbytes) |
| 194 | { |
| 195 | return glue_fpu_begin(CAST6_BLOCK_SIZE, CAST6_PARALLEL_BLOCKS, |
| 196 | NULL, fpu_enabled, nbytes); |
| 197 | } |
| 198 | |
| 199 | static inline void cast6_fpu_end(bool fpu_enabled) |
| 200 | { |
| 201 | glue_fpu_end(fpu_enabled); |
| 202 | } |
| 203 | |
| 204 | struct crypt_priv { |
| 205 | struct cast6_ctx *ctx; |
| 206 | bool fpu_enabled; |
| 207 | }; |
| 208 | |
| 209 | static void encrypt_callback(void *priv, u8 *srcdst, unsigned int nbytes) |
| 210 | { |
| 211 | const unsigned int bsize = CAST6_BLOCK_SIZE; |
| 212 | struct crypt_priv *ctx = priv; |
| 213 | int i; |
| 214 | |
| 215 | ctx->fpu_enabled = cast6_fpu_begin(ctx->fpu_enabled, nbytes); |
| 216 | |
| 217 | if (nbytes == bsize * CAST6_PARALLEL_BLOCKS) { |
Jussi Kivilinna | cba1cce | 2012-10-20 15:06:41 +0300 | [diff] [blame] | 218 | cast6_ecb_enc_8way(ctx->ctx, srcdst, srcdst); |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 219 | return; |
| 220 | } |
| 221 | |
| 222 | for (i = 0; i < nbytes / bsize; i++, srcdst += bsize) |
| 223 | __cast6_encrypt(ctx->ctx, srcdst, srcdst); |
| 224 | } |
| 225 | |
| 226 | static void decrypt_callback(void *priv, u8 *srcdst, unsigned int nbytes) |
| 227 | { |
| 228 | const unsigned int bsize = CAST6_BLOCK_SIZE; |
| 229 | struct crypt_priv *ctx = priv; |
| 230 | int i; |
| 231 | |
| 232 | ctx->fpu_enabled = cast6_fpu_begin(ctx->fpu_enabled, nbytes); |
| 233 | |
| 234 | if (nbytes == bsize * CAST6_PARALLEL_BLOCKS) { |
Jussi Kivilinna | cba1cce | 2012-10-20 15:06:41 +0300 | [diff] [blame] | 235 | cast6_ecb_dec_8way(ctx->ctx, srcdst, srcdst); |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 236 | return; |
| 237 | } |
| 238 | |
| 239 | for (i = 0; i < nbytes / bsize; i++, srcdst += bsize) |
| 240 | __cast6_decrypt(ctx->ctx, srcdst, srcdst); |
| 241 | } |
| 242 | |
| 243 | struct cast6_lrw_ctx { |
| 244 | struct lrw_table_ctx lrw_table; |
| 245 | struct cast6_ctx cast6_ctx; |
| 246 | }; |
| 247 | |
| 248 | static int lrw_cast6_setkey(struct crypto_tfm *tfm, const u8 *key, |
| 249 | unsigned int keylen) |
| 250 | { |
| 251 | struct cast6_lrw_ctx *ctx = crypto_tfm_ctx(tfm); |
| 252 | int err; |
| 253 | |
| 254 | err = __cast6_setkey(&ctx->cast6_ctx, key, keylen - CAST6_BLOCK_SIZE, |
| 255 | &tfm->crt_flags); |
| 256 | if (err) |
| 257 | return err; |
| 258 | |
| 259 | return lrw_init_table(&ctx->lrw_table, key + keylen - CAST6_BLOCK_SIZE); |
| 260 | } |
| 261 | |
| 262 | static int lrw_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, |
| 263 | struct scatterlist *src, unsigned int nbytes) |
| 264 | { |
| 265 | struct cast6_lrw_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); |
| 266 | be128 buf[CAST6_PARALLEL_BLOCKS]; |
| 267 | struct crypt_priv crypt_ctx = { |
| 268 | .ctx = &ctx->cast6_ctx, |
| 269 | .fpu_enabled = false, |
| 270 | }; |
| 271 | struct lrw_crypt_req req = { |
| 272 | .tbuf = buf, |
| 273 | .tbuflen = sizeof(buf), |
| 274 | |
| 275 | .table_ctx = &ctx->lrw_table, |
| 276 | .crypt_ctx = &crypt_ctx, |
| 277 | .crypt_fn = encrypt_callback, |
| 278 | }; |
| 279 | int ret; |
| 280 | |
| 281 | desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; |
| 282 | ret = lrw_crypt(desc, dst, src, nbytes, &req); |
| 283 | cast6_fpu_end(crypt_ctx.fpu_enabled); |
| 284 | |
| 285 | return ret; |
| 286 | } |
| 287 | |
| 288 | static int lrw_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, |
| 289 | struct scatterlist *src, unsigned int nbytes) |
| 290 | { |
| 291 | struct cast6_lrw_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); |
| 292 | be128 buf[CAST6_PARALLEL_BLOCKS]; |
| 293 | struct crypt_priv crypt_ctx = { |
| 294 | .ctx = &ctx->cast6_ctx, |
| 295 | .fpu_enabled = false, |
| 296 | }; |
| 297 | struct lrw_crypt_req req = { |
| 298 | .tbuf = buf, |
| 299 | .tbuflen = sizeof(buf), |
| 300 | |
| 301 | .table_ctx = &ctx->lrw_table, |
| 302 | .crypt_ctx = &crypt_ctx, |
| 303 | .crypt_fn = decrypt_callback, |
| 304 | }; |
| 305 | int ret; |
| 306 | |
| 307 | desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; |
| 308 | ret = lrw_crypt(desc, dst, src, nbytes, &req); |
| 309 | cast6_fpu_end(crypt_ctx.fpu_enabled); |
| 310 | |
| 311 | return ret; |
| 312 | } |
| 313 | |
| 314 | static void lrw_exit_tfm(struct crypto_tfm *tfm) |
| 315 | { |
| 316 | struct cast6_lrw_ctx *ctx = crypto_tfm_ctx(tfm); |
| 317 | |
| 318 | lrw_free_table(&ctx->lrw_table); |
| 319 | } |
| 320 | |
| 321 | struct cast6_xts_ctx { |
| 322 | struct cast6_ctx tweak_ctx; |
| 323 | struct cast6_ctx crypt_ctx; |
| 324 | }; |
| 325 | |
| 326 | static int xts_cast6_setkey(struct crypto_tfm *tfm, const u8 *key, |
| 327 | unsigned int keylen) |
| 328 | { |
| 329 | struct cast6_xts_ctx *ctx = crypto_tfm_ctx(tfm); |
| 330 | u32 *flags = &tfm->crt_flags; |
| 331 | int err; |
| 332 | |
| 333 | /* key consists of keys of equal size concatenated, therefore |
| 334 | * the length must be even |
| 335 | */ |
| 336 | if (keylen % 2) { |
| 337 | *flags |= CRYPTO_TFM_RES_BAD_KEY_LEN; |
| 338 | return -EINVAL; |
| 339 | } |
| 340 | |
| 341 | /* first half of xts-key is for crypt */ |
| 342 | err = __cast6_setkey(&ctx->crypt_ctx, key, keylen / 2, flags); |
| 343 | if (err) |
| 344 | return err; |
| 345 | |
| 346 | /* second half of xts-key is for tweak */ |
| 347 | return __cast6_setkey(&ctx->tweak_ctx, key + keylen / 2, keylen / 2, |
| 348 | flags); |
| 349 | } |
| 350 | |
| 351 | static int xts_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, |
| 352 | struct scatterlist *src, unsigned int nbytes) |
| 353 | { |
| 354 | struct cast6_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 355 | |
Jussi Kivilinna | 7017728 | 2013-04-08 21:51:05 +0300 | [diff] [blame] | 356 | return glue_xts_crypt_128bit(&cast6_enc_xts, desc, dst, src, nbytes, |
| 357 | XTS_TWEAK_CAST(__cast6_encrypt), |
| 358 | &ctx->tweak_ctx, &ctx->crypt_ctx); |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 359 | } |
| 360 | |
| 361 | static int xts_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, |
| 362 | struct scatterlist *src, unsigned int nbytes) |
| 363 | { |
| 364 | struct cast6_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 365 | |
Jussi Kivilinna | 7017728 | 2013-04-08 21:51:05 +0300 | [diff] [blame] | 366 | return glue_xts_crypt_128bit(&cast6_dec_xts, desc, dst, src, nbytes, |
| 367 | XTS_TWEAK_CAST(__cast6_encrypt), |
| 368 | &ctx->tweak_ctx, &ctx->crypt_ctx); |
Johannes Goetzfried | 4ea1277 | 2012-07-11 19:38:57 +0200 | [diff] [blame] | 369 | } |
| 370 | |
| 371 | static struct crypto_alg cast6_algs[10] = { { |
| 372 | .cra_name = "__ecb-cast6-avx", |
| 373 | .cra_driver_name = "__driver-ecb-cast6-avx", |
| 374 | .cra_priority = 0, |
| 375 | .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, |
| 376 | .cra_blocksize = CAST6_BLOCK_SIZE, |
| 377 | .cra_ctxsize = sizeof(struct cast6_ctx), |
| 378 | .cra_alignmask = 0, |
| 379 | .cra_type = &crypto_blkcipher_type, |
| 380 | .cra_module = THIS_MODULE, |
| 381 | .cra_u = { |
| 382 | .blkcipher = { |
| 383 | .min_keysize = CAST6_MIN_KEY_SIZE, |
| 384 | .max_keysize = CAST6_MAX_KEY_SIZE, |
| 385 | .setkey = cast6_setkey, |
| 386 | .encrypt = ecb_encrypt, |
| 387 | .decrypt = ecb_decrypt, |
| 388 | }, |
| 389 | }, |
| 390 | }, { |
| 391 | .cra_name = "__cbc-cast6-avx", |
| 392 | .cra_driver_name = "__driver-cbc-cast6-avx", |
| 393 | .cra_priority = 0, |
| 394 | .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, |
| 395 | .cra_blocksize = CAST6_BLOCK_SIZE, |
| 396 | .cra_ctxsize = sizeof(struct cast6_ctx), |
| 397 | .cra_alignmask = 0, |
| 398 | .cra_type = &crypto_blkcipher_type, |
| 399 | .cra_module = THIS_MODULE, |
| 400 | .cra_u = { |
| 401 | .blkcipher = { |
| 402 | .min_keysize = CAST6_MIN_KEY_SIZE, |
| 403 | .max_keysize = CAST6_MAX_KEY_SIZE, |
| 404 | .setkey = cast6_setkey, |
| 405 | .encrypt = cbc_encrypt, |
| 406 | .decrypt = cbc_decrypt, |
| 407 | }, |
| 408 | }, |
| 409 | }, { |
| 410 | .cra_name = "__ctr-cast6-avx", |
| 411 | .cra_driver_name = "__driver-ctr-cast6-avx", |
| 412 | .cra_priority = 0, |
| 413 | .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, |
| 414 | .cra_blocksize = 1, |
| 415 | .cra_ctxsize = sizeof(struct cast6_ctx), |
| 416 | .cra_alignmask = 0, |
| 417 | .cra_type = &crypto_blkcipher_type, |
| 418 | .cra_module = THIS_MODULE, |
| 419 | .cra_u = { |
| 420 | .blkcipher = { |
| 421 | .min_keysize = CAST6_MIN_KEY_SIZE, |
| 422 | .max_keysize = CAST6_MAX_KEY_SIZE, |
| 423 | .ivsize = CAST6_BLOCK_SIZE, |
| 424 | .setkey = cast6_setkey, |
| 425 | .encrypt = ctr_crypt, |
| 426 | .decrypt = ctr_crypt, |
| 427 | }, |
| 428 | }, |
| 429 | }, { |
| 430 | .cra_name = "__lrw-cast6-avx", |
| 431 | .cra_driver_name = "__driver-lrw-cast6-avx", |
| 432 | .cra_priority = 0, |
| 433 | .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, |
| 434 | .cra_blocksize = CAST6_BLOCK_SIZE, |
| 435 | .cra_ctxsize = sizeof(struct cast6_lrw_ctx), |
| 436 | .cra_alignmask = 0, |
| 437 | .cra_type = &crypto_blkcipher_type, |
| 438 | .cra_module = THIS_MODULE, |
| 439 | .cra_exit = lrw_exit_tfm, |
| 440 | .cra_u = { |
| 441 | .blkcipher = { |
| 442 | .min_keysize = CAST6_MIN_KEY_SIZE + |
| 443 | CAST6_BLOCK_SIZE, |
| 444 | .max_keysize = CAST6_MAX_KEY_SIZE + |
| 445 | CAST6_BLOCK_SIZE, |
| 446 | .ivsize = CAST6_BLOCK_SIZE, |
| 447 | .setkey = lrw_cast6_setkey, |
| 448 | .encrypt = lrw_encrypt, |
| 449 | .decrypt = lrw_decrypt, |
| 450 | }, |
| 451 | }, |
| 452 | }, { |
| 453 | .cra_name = "__xts-cast6-avx", |
| 454 | .cra_driver_name = "__driver-xts-cast6-avx", |
| 455 | .cra_priority = 0, |
| 456 | .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, |
| 457 | .cra_blocksize = CAST6_BLOCK_SIZE, |
| 458 | .cra_ctxsize = sizeof(struct cast6_xts_ctx), |
| 459 | .cra_alignmask = 0, |
| 460 | .cra_type = &crypto_blkcipher_type, |
| 461 | .cra_module = THIS_MODULE, |
| 462 | .cra_u = { |
| 463 | .blkcipher = { |
| 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, |
| 470 | }, |
| 471 | }, |
| 472 | }, { |
| 473 | .cra_name = "ecb(cast6)", |
| 474 | .cra_driver_name = "ecb-cast6-avx", |
| 475 | .cra_priority = 200, |
| 476 | .cra_flags = CRYPTO_ALG_TYPE_ABLKCIPHER | CRYPTO_ALG_ASYNC, |
| 477 | .cra_blocksize = CAST6_BLOCK_SIZE, |
| 478 | .cra_ctxsize = sizeof(struct async_helper_ctx), |
| 479 | .cra_alignmask = 0, |
| 480 | .cra_type = &crypto_ablkcipher_type, |
| 481 | .cra_module = THIS_MODULE, |
| 482 | .cra_init = ablk_init, |
| 483 | .cra_exit = ablk_exit, |
| 484 | .cra_u = { |
| 485 | .ablkcipher = { |
| 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, |
| 491 | }, |
| 492 | }, |
| 493 | }, { |
| 494 | .cra_name = "cbc(cast6)", |
| 495 | .cra_driver_name = "cbc-cast6-avx", |
| 496 | .cra_priority = 200, |
| 497 | .cra_flags = CRYPTO_ALG_TYPE_ABLKCIPHER | CRYPTO_ALG_ASYNC, |
| 498 | .cra_blocksize = CAST6_BLOCK_SIZE, |
| 499 | .cra_ctxsize = sizeof(struct async_helper_ctx), |
| 500 | .cra_alignmask = 0, |
| 501 | .cra_type = &crypto_ablkcipher_type, |
| 502 | .cra_module = THIS_MODULE, |
| 503 | .cra_init = ablk_init, |
| 504 | .cra_exit = ablk_exit, |
| 505 | .cra_u = { |
| 506 | .ablkcipher = { |
| 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, |
| 513 | }, |
| 514 | }, |
| 515 | }, { |
| 516 | .cra_name = "ctr(cast6)", |
| 517 | .cra_driver_name = "ctr-cast6-avx", |
| 518 | .cra_priority = 200, |
| 519 | .cra_flags = CRYPTO_ALG_TYPE_ABLKCIPHER | CRYPTO_ALG_ASYNC, |
| 520 | .cra_blocksize = 1, |
| 521 | .cra_ctxsize = sizeof(struct async_helper_ctx), |
| 522 | .cra_alignmask = 0, |
| 523 | .cra_type = &crypto_ablkcipher_type, |
| 524 | .cra_module = THIS_MODULE, |
| 525 | .cra_init = ablk_init, |
| 526 | .cra_exit = ablk_exit, |
| 527 | .cra_u = { |
| 528 | .ablkcipher = { |
| 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, |
| 535 | .geniv = "chainiv", |
| 536 | }, |
| 537 | }, |
| 538 | }, { |
| 539 | .cra_name = "lrw(cast6)", |
| 540 | .cra_driver_name = "lrw-cast6-avx", |
| 541 | .cra_priority = 200, |
| 542 | .cra_flags = CRYPTO_ALG_TYPE_ABLKCIPHER | CRYPTO_ALG_ASYNC, |
| 543 | .cra_blocksize = CAST6_BLOCK_SIZE, |
| 544 | .cra_ctxsize = sizeof(struct async_helper_ctx), |
| 545 | .cra_alignmask = 0, |
| 546 | .cra_type = &crypto_ablkcipher_type, |
| 547 | .cra_module = THIS_MODULE, |
| 548 | .cra_init = ablk_init, |
| 549 | .cra_exit = ablk_exit, |
| 550 | .cra_u = { |
| 551 | .ablkcipher = { |
| 552 | .min_keysize = CAST6_MIN_KEY_SIZE + |
| 553 | CAST6_BLOCK_SIZE, |
| 554 | .max_keysize = CAST6_MAX_KEY_SIZE + |
| 555 | CAST6_BLOCK_SIZE, |
| 556 | .ivsize = CAST6_BLOCK_SIZE, |
| 557 | .setkey = ablk_set_key, |
| 558 | .encrypt = ablk_encrypt, |
| 559 | .decrypt = ablk_decrypt, |
| 560 | }, |
| 561 | }, |
| 562 | }, { |
| 563 | .cra_name = "xts(cast6)", |
| 564 | .cra_driver_name = "xts-cast6-avx", |
| 565 | .cra_priority = 200, |
| 566 | .cra_flags = CRYPTO_ALG_TYPE_ABLKCIPHER | CRYPTO_ALG_ASYNC, |
| 567 | .cra_blocksize = CAST6_BLOCK_SIZE, |
| 568 | .cra_ctxsize = sizeof(struct async_helper_ctx), |
| 569 | .cra_alignmask = 0, |
| 570 | .cra_type = &crypto_ablkcipher_type, |
| 571 | .cra_module = THIS_MODULE, |
| 572 | .cra_init = ablk_init, |
| 573 | .cra_exit = ablk_exit, |
| 574 | .cra_u = { |
| 575 | .ablkcipher = { |
| 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, |
| 582 | }, |
| 583 | }, |
| 584 | } }; |
| 585 | |
| 586 | static int __init cast6_init(void) |
| 587 | { |
| 588 | u64 xcr0; |
| 589 | |
| 590 | if (!cpu_has_avx || !cpu_has_osxsave) { |
| 591 | pr_info("AVX instructions are not detected.\n"); |
| 592 | return -ENODEV; |
| 593 | } |
| 594 | |
| 595 | xcr0 = xgetbv(XCR_XFEATURE_ENABLED_MASK); |
| 596 | if ((xcr0 & (XSTATE_SSE | XSTATE_YMM)) != (XSTATE_SSE | XSTATE_YMM)) { |
| 597 | pr_info("AVX detected but unusable.\n"); |
| 598 | return -ENODEV; |
| 599 | } |
| 600 | |
| 601 | return crypto_register_algs(cast6_algs, ARRAY_SIZE(cast6_algs)); |
| 602 | } |
| 603 | |
| 604 | static void __exit cast6_exit(void) |
| 605 | { |
| 606 | crypto_unregister_algs(cast6_algs, ARRAY_SIZE(cast6_algs)); |
| 607 | } |
| 608 | |
| 609 | module_init(cast6_init); |
| 610 | module_exit(cast6_exit); |
| 611 | |
| 612 | MODULE_DESCRIPTION("Cast6 Cipher Algorithm, AVX optimized"); |
| 613 | MODULE_LICENSE("GPL"); |
| 614 | MODULE_ALIAS("cast6"); |