crypto: twofish - add 3-way parallel x86_64 assembler implemention

Patch adds 3-way parallel x86_64 assembly implementation of twofish as new
module. New assembler functions crypt data in three blocks chunks, improving
cipher performance on out-of-order CPUs.

Patch has been tested with tcrypt and automated filesystem tests.

Summary of the tcrypt benchmarks:

Twofish 3-way-asm vs twofish asm (128bit 8kb block ECB)
 encrypt: 1.3x speed
 decrypt: 1.3x speed

Twofish 3-way-asm vs twofish asm (128bit 8kb block CBC)
 encrypt: 1.07x speed
 decrypt: 1.4x speed

Twofish 3-way-asm vs twofish asm (128bit 8kb block CTR)
 encrypt: 1.4x speed

Twofish 3-way-asm vs AES asm (128bit 8kb block ECB)
 encrypt: 1.0x speed
 decrypt: 1.0x speed

Twofish 3-way-asm vs AES asm (128bit 8kb block CBC)
 encrypt: 0.84x speed
 decrypt: 1.09x speed

Twofish 3-way-asm vs AES asm (128bit 8kb block CTR)
 encrypt: 1.15x speed

Full output:
 http://koti.mbnet.fi/axh/kernel/crypto/tcrypt-speed-twofish-3way-asm-x86_64.txt
 http://koti.mbnet.fi/axh/kernel/crypto/tcrypt-speed-twofish-asm-x86_64.txt
 http://koti.mbnet.fi/axh/kernel/crypto/tcrypt-speed-aes-asm-x86_64.txt

Tests were run on:
 vendor_id  : AuthenticAMD
 cpu family : 16
 model      : 10
 model name : AMD Phenom(tm) II X6 1055T Processor

Also userspace test were run on:
 vendor_id  : GenuineIntel
 cpu family : 6
 model      : 15
 model name : Intel(R) Xeon(R) CPU           E7330  @ 2.40GHz
 stepping   : 11

Userspace test results:

Encryption/decryption of twofish 3-way vs x86_64-asm on AMD Phenom II:
 encrypt: 1.27x
 decrypt: 1.25x

Encryption/decryption of twofish 3-way vs x86_64-asm on Intel Xeon E7330:
 encrypt: 1.36x
 decrypt: 1.36x

Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This commit is contained in:
Jussi Kivilinna 2011-09-26 16:47:25 +03:00 committed by Herbert Xu
parent 91d41f159d
commit 8280daad43
4 changed files with 810 additions and 0 deletions

View file

@ -9,6 +9,7 @@ obj-$(CONFIG_CRYPTO_SALSA20_586) += salsa20-i586.o
obj-$(CONFIG_CRYPTO_AES_X86_64) += aes-x86_64.o
obj-$(CONFIG_CRYPTO_BLOWFISH_X86_64) += blowfish-x86_64.o
obj-$(CONFIG_CRYPTO_TWOFISH_X86_64) += twofish-x86_64.o
obj-$(CONFIG_CRYPTO_TWOFISH_X86_64_3WAY) += twofish-x86_64-3way.o
obj-$(CONFIG_CRYPTO_SALSA20_X86_64) += salsa20-x86_64.o
obj-$(CONFIG_CRYPTO_AES_NI_INTEL) += aesni-intel.o
obj-$(CONFIG_CRYPTO_GHASH_CLMUL_NI_INTEL) += ghash-clmulni-intel.o
@ -23,6 +24,7 @@ salsa20-i586-y := salsa20-i586-asm_32.o salsa20_glue.o
aes-x86_64-y := aes-x86_64-asm_64.o aes_glue.o
blowfish-x86_64-y := blowfish-x86_64-asm_64.o blowfish_glue.o
twofish-x86_64-y := twofish-x86_64-asm_64.o twofish_glue.o
twofish-x86_64-3way-y := twofish-x86_64-asm_64-3way.o twofish_glue_3way.o
salsa20-x86_64-y := salsa20-x86_64-asm_64.o salsa20_glue.o
aesni-intel-y := aesni-intel_asm.o aesni-intel_glue.o fpu.o

View file

@ -0,0 +1,316 @@
/*
* Twofish Cipher 3-way parallel algorithm (x86_64)
*
* Copyright (C) 2011 Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
* USA
*
*/
.file "twofish-x86_64-asm-3way.S"
.text
/* structure of crypto context */
#define s0 0
#define s1 1024
#define s2 2048
#define s3 3072
#define w 4096
#define k 4128
/**********************************************************************
3-way twofish
**********************************************************************/
#define CTX %rdi
#define RIO %rdx
#define RAB0 %rax
#define RAB1 %rbx
#define RAB2 %rcx
#define RAB0d %eax
#define RAB1d %ebx
#define RAB2d %ecx
#define RAB0bh %ah
#define RAB1bh %bh
#define RAB2bh %ch
#define RAB0bl %al
#define RAB1bl %bl
#define RAB2bl %cl
#define RCD0 %r8
#define RCD1 %r9
#define RCD2 %r10
#define RCD0d %r8d
#define RCD1d %r9d
#define RCD2d %r10d
#define RX0 %rbp
#define RX1 %r11
#define RX2 %r12
#define RX0d %ebp
#define RX1d %r11d
#define RX2d %r12d
#define RY0 %r13
#define RY1 %r14
#define RY2 %r15
#define RY0d %r13d
#define RY1d %r14d
#define RY2d %r15d
#define RT0 %rdx
#define RT1 %rsi
#define RT0d %edx
#define RT1d %esi
#define do16bit_ror(rot, op1, op2, T0, T1, tmp1, tmp2, ab, dst) \
movzbl ab ## bl, tmp2 ## d; \
movzbl ab ## bh, tmp1 ## d; \
rorq $(rot), ab; \
op1##l T0(CTX, tmp2, 4), dst ## d; \
op2##l T1(CTX, tmp1, 4), dst ## d;
/*
* Combined G1 & G2 function. Reordered with help of rotates to have moves
* at begining.
*/
#define g1g2_3(ab, cd, Tx0, Tx1, Tx2, Tx3, Ty0, Ty1, Ty2, Ty3, x, y) \
/* G1,1 && G2,1 */ \
do16bit_ror(32, mov, xor, Tx0, Tx1, RT0, x ## 0, ab ## 0, x ## 0); \
do16bit_ror(48, mov, xor, Ty1, Ty2, RT0, y ## 0, ab ## 0, y ## 0); \
\
do16bit_ror(32, mov, xor, Tx0, Tx1, RT0, x ## 1, ab ## 1, x ## 1); \
do16bit_ror(48, mov, xor, Ty1, Ty2, RT0, y ## 1, ab ## 1, y ## 1); \
\
do16bit_ror(32, mov, xor, Tx0, Tx1, RT0, x ## 2, ab ## 2, x ## 2); \
do16bit_ror(48, mov, xor, Ty1, Ty2, RT0, y ## 2, ab ## 2, y ## 2); \
\
/* G1,2 && G2,2 */ \
do16bit_ror(32, xor, xor, Tx2, Tx3, RT0, RT1, ab ## 0, x ## 0); \
do16bit_ror(16, xor, xor, Ty3, Ty0, RT0, RT1, ab ## 0, y ## 0); \
xchgq cd ## 0, ab ## 0; \
\
do16bit_ror(32, xor, xor, Tx2, Tx3, RT0, RT1, ab ## 1, x ## 1); \
do16bit_ror(16, xor, xor, Ty3, Ty0, RT0, RT1, ab ## 1, y ## 1); \
xchgq cd ## 1, ab ## 1; \
\
do16bit_ror(32, xor, xor, Tx2, Tx3, RT0, RT1, ab ## 2, x ## 2); \
do16bit_ror(16, xor, xor, Ty3, Ty0, RT0, RT1, ab ## 2, y ## 2); \
xchgq cd ## 2, ab ## 2;
#define enc_round_end(ab, x, y, n) \
addl y ## d, x ## d; \
addl x ## d, y ## d; \
addl k+4*(2*(n))(CTX), x ## d; \
xorl ab ## d, x ## d; \
addl k+4*(2*(n)+1)(CTX), y ## d; \
shrq $32, ab; \
roll $1, ab ## d; \
xorl y ## d, ab ## d; \
shlq $32, ab; \
rorl $1, x ## d; \
orq x, ab;
#define dec_round_end(ba, x, y, n) \
addl y ## d, x ## d; \
addl x ## d, y ## d; \
addl k+4*(2*(n))(CTX), x ## d; \
addl k+4*(2*(n)+1)(CTX), y ## d; \
xorl ba ## d, y ## d; \
shrq $32, ba; \
roll $1, ba ## d; \
xorl x ## d, ba ## d; \
shlq $32, ba; \
rorl $1, y ## d; \
orq y, ba;
#define encrypt_round3(ab, cd, n) \
g1g2_3(ab, cd, s0, s1, s2, s3, s0, s1, s2, s3, RX, RY); \
\
enc_round_end(ab ## 0, RX0, RY0, n); \
enc_round_end(ab ## 1, RX1, RY1, n); \
enc_round_end(ab ## 2, RX2, RY2, n);
#define decrypt_round3(ba, dc, n) \
g1g2_3(ba, dc, s1, s2, s3, s0, s3, s0, s1, s2, RY, RX); \
\
dec_round_end(ba ## 0, RX0, RY0, n); \
dec_round_end(ba ## 1, RX1, RY1, n); \
dec_round_end(ba ## 2, RX2, RY2, n);
#define encrypt_cycle3(ab, cd, n) \
encrypt_round3(ab, cd, n*2); \
encrypt_round3(ab, cd, (n*2)+1);
#define decrypt_cycle3(ba, dc, n) \
decrypt_round3(ba, dc, (n*2)+1); \
decrypt_round3(ba, dc, (n*2));
#define inpack3(in, n, xy, m) \
movq 4*(n)(in), xy ## 0; \
xorq w+4*m(CTX), xy ## 0; \
\
movq 4*(4+(n))(in), xy ## 1; \
xorq w+4*m(CTX), xy ## 1; \
\
movq 4*(8+(n))(in), xy ## 2; \
xorq w+4*m(CTX), xy ## 2;
#define outunpack3(op, out, n, xy, m) \
xorq w+4*m(CTX), xy ## 0; \
op ## q xy ## 0, 4*(n)(out); \
\
xorq w+4*m(CTX), xy ## 1; \
op ## q xy ## 1, 4*(4+(n))(out); \
\
xorq w+4*m(CTX), xy ## 2; \
op ## q xy ## 2, 4*(8+(n))(out);
#define inpack_enc3() \
inpack3(RIO, 0, RAB, 0); \
inpack3(RIO, 2, RCD, 2);
#define outunpack_enc3(op) \
outunpack3(op, RIO, 2, RAB, 6); \
outunpack3(op, RIO, 0, RCD, 4);
#define inpack_dec3() \
inpack3(RIO, 0, RAB, 4); \
rorq $32, RAB0; \
rorq $32, RAB1; \
rorq $32, RAB2; \
inpack3(RIO, 2, RCD, 6); \
rorq $32, RCD0; \
rorq $32, RCD1; \
rorq $32, RCD2;
#define outunpack_dec3() \
rorq $32, RCD0; \
rorq $32, RCD1; \
rorq $32, RCD2; \
outunpack3(mov, RIO, 0, RCD, 0); \
rorq $32, RAB0; \
rorq $32, RAB1; \
rorq $32, RAB2; \
outunpack3(mov, RIO, 2, RAB, 2);
.align 8
.global __twofish_enc_blk_3way
.type __twofish_enc_blk_3way,@function;
__twofish_enc_blk_3way:
/* input:
* %rdi: ctx, CTX
* %rsi: dst
* %rdx: src, RIO
* %rcx: bool, if true: xor output
*/
pushq %r15;
pushq %r14;
pushq %r13;
pushq %r12;
pushq %rbp;
pushq %rbx;
pushq %rcx; /* bool xor */
pushq %rsi; /* dst */
inpack_enc3();
encrypt_cycle3(RAB, RCD, 0);
encrypt_cycle3(RAB, RCD, 1);
encrypt_cycle3(RAB, RCD, 2);
encrypt_cycle3(RAB, RCD, 3);
encrypt_cycle3(RAB, RCD, 4);
encrypt_cycle3(RAB, RCD, 5);
encrypt_cycle3(RAB, RCD, 6);
encrypt_cycle3(RAB, RCD, 7);
popq RIO; /* dst */
popq %rbp; /* bool xor */
testb %bpl, %bpl;
jnz __enc_xor3;
outunpack_enc3(mov);
popq %rbx;
popq %rbp;
popq %r12;
popq %r13;
popq %r14;
popq %r15;
ret;
__enc_xor3:
outunpack_enc3(xor);
popq %rbx;
popq %rbp;
popq %r12;
popq %r13;
popq %r14;
popq %r15;
ret;
.global twofish_dec_blk_3way
.type twofish_dec_blk_3way,@function;
twofish_dec_blk_3way:
/* input:
* %rdi: ctx, CTX
* %rsi: dst
* %rdx: src, RIO
*/
pushq %r15;
pushq %r14;
pushq %r13;
pushq %r12;
pushq %rbp;
pushq %rbx;
pushq %rsi; /* dst */
inpack_dec3();
decrypt_cycle3(RAB, RCD, 7);
decrypt_cycle3(RAB, RCD, 6);
decrypt_cycle3(RAB, RCD, 5);
decrypt_cycle3(RAB, RCD, 4);
decrypt_cycle3(RAB, RCD, 3);
decrypt_cycle3(RAB, RCD, 2);
decrypt_cycle3(RAB, RCD, 1);
decrypt_cycle3(RAB, RCD, 0);
popq RIO; /* dst */
outunpack_dec3();
popq %rbx;
popq %rbp;
popq %r12;
popq %r13;
popq %r14;
popq %r15;
ret;

View file

@ -0,0 +1,472 @@
/*
* Glue Code for 3-way parallel assembler optimized version of Twofish
*
* Copyright (c) 2011 Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
*
* CBC & ECB parts based on code (crypto/cbc.c,ecb.c) by:
* Copyright (c) 2006 Herbert Xu <herbert@gondor.apana.org.au>
* CTR part based on code (crypto/ctr.c) by:
* (C) Copyright IBM Corp. 2007 - Joy Latten <latten@us.ibm.com>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
* USA
*
*/
#include <linux/crypto.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/types.h>
#include <crypto/algapi.h>
#include <crypto/twofish.h>
#include <crypto/b128ops.h>
/* regular block cipher functions from twofish_x86_64 module */
asmlinkage void twofish_enc_blk(struct twofish_ctx *ctx, u8 *dst,
const u8 *src);
asmlinkage void twofish_dec_blk(struct twofish_ctx *ctx, u8 *dst,
const u8 *src);
/* 3-way parallel cipher functions */
asmlinkage void __twofish_enc_blk_3way(struct twofish_ctx *ctx, u8 *dst,
const u8 *src, bool xor);
asmlinkage void twofish_dec_blk_3way(struct twofish_ctx *ctx, u8 *dst,
const u8 *src);
static inline void twofish_enc_blk_3way(struct twofish_ctx *ctx, u8 *dst,
const u8 *src)
{
__twofish_enc_blk_3way(ctx, dst, src, false);
}
static inline void twofish_enc_blk_xor_3way(struct twofish_ctx *ctx, u8 *dst,
const u8 *src)
{
__twofish_enc_blk_3way(ctx, dst, src, true);
}
static int ecb_crypt(struct blkcipher_desc *desc, struct blkcipher_walk *walk,
void (*fn)(struct twofish_ctx *, u8 *, const u8 *),
void (*fn_3way)(struct twofish_ctx *, u8 *, const u8 *))
{
struct twofish_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
unsigned int bsize = TF_BLOCK_SIZE;
unsigned int nbytes;
int err;
err = blkcipher_walk_virt(desc, walk);
while ((nbytes = walk->nbytes)) {
u8 *wsrc = walk->src.virt.addr;
u8 *wdst = walk->dst.virt.addr;
/* Process three block batch */
if (nbytes >= bsize * 3) {
do {
fn_3way(ctx, wdst, wsrc);
wsrc += bsize * 3;
wdst += bsize * 3;
nbytes -= bsize * 3;
} while (nbytes >= bsize * 3);
if (nbytes < bsize)
goto done;
}
/* Handle leftovers */
do {
fn(ctx, wdst, wsrc);
wsrc += bsize;
wdst += bsize;
nbytes -= bsize;
} while (nbytes >= bsize);
done:
err = blkcipher_walk_done(desc, walk, nbytes);
}
return err;
}
static int ecb_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
struct scatterlist *src, unsigned int nbytes)
{
struct blkcipher_walk walk;
blkcipher_walk_init(&walk, dst, src, nbytes);
return ecb_crypt(desc, &walk, twofish_enc_blk, twofish_enc_blk_3way);
}
static int ecb_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
struct scatterlist *src, unsigned int nbytes)
{
struct blkcipher_walk walk;
blkcipher_walk_init(&walk, dst, src, nbytes);
return ecb_crypt(desc, &walk, twofish_dec_blk, twofish_dec_blk_3way);
}
static struct crypto_alg blk_ecb_alg = {
.cra_name = "ecb(twofish)",
.cra_driver_name = "ecb-twofish-3way",
.cra_priority = 300,
.cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER,
.cra_blocksize = TF_BLOCK_SIZE,
.cra_ctxsize = sizeof(struct twofish_ctx),
.cra_alignmask = 0,
.cra_type = &crypto_blkcipher_type,
.cra_module = THIS_MODULE,
.cra_list = LIST_HEAD_INIT(blk_ecb_alg.cra_list),
.cra_u = {
.blkcipher = {
.min_keysize = TF_MIN_KEY_SIZE,
.max_keysize = TF_MAX_KEY_SIZE,
.setkey = twofish_setkey,
.encrypt = ecb_encrypt,
.decrypt = ecb_decrypt,
},
},
};
static unsigned int __cbc_encrypt(struct blkcipher_desc *desc,
struct blkcipher_walk *walk)
{
struct twofish_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
unsigned int bsize = TF_BLOCK_SIZE;
unsigned int nbytes = walk->nbytes;
u128 *src = (u128 *)walk->src.virt.addr;
u128 *dst = (u128 *)walk->dst.virt.addr;
u128 *iv = (u128 *)walk->iv;
do {
u128_xor(dst, src, iv);
twofish_enc_blk(ctx, (u8 *)dst, (u8 *)dst);
iv = dst;
src += 1;
dst += 1;
nbytes -= bsize;
} while (nbytes >= bsize);
u128_xor((u128 *)walk->iv, (u128 *)walk->iv, iv);
return nbytes;
}
static int cbc_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
struct scatterlist *src, unsigned int nbytes)
{
struct blkcipher_walk walk;
int err;
blkcipher_walk_init(&walk, dst, src, nbytes);
err = blkcipher_walk_virt(desc, &walk);
while ((nbytes = walk.nbytes)) {
nbytes = __cbc_encrypt(desc, &walk);
err = blkcipher_walk_done(desc, &walk, nbytes);
}
return err;
}
static unsigned int __cbc_decrypt(struct blkcipher_desc *desc,
struct blkcipher_walk *walk)
{
struct twofish_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
unsigned int bsize = TF_BLOCK_SIZE;
unsigned int nbytes = walk->nbytes;
u128 *src = (u128 *)walk->src.virt.addr;
u128 *dst = (u128 *)walk->dst.virt.addr;
u128 ivs[3 - 1];
u128 last_iv;
/* Start of the last block. */
src += nbytes / bsize - 1;
dst += nbytes / bsize - 1;
last_iv = *src;
/* Process three block batch */
if (nbytes >= bsize * 3) {
do {
nbytes -= bsize * (3 - 1);
src -= 3 - 1;
dst -= 3 - 1;
ivs[0] = src[0];
ivs[1] = src[1];
twofish_dec_blk_3way(ctx, (u8 *)dst, (u8 *)src);
u128_xor(dst + 1, dst + 1, ivs + 0);
u128_xor(dst + 2, dst + 2, ivs + 1);
nbytes -= bsize;
if (nbytes < bsize)
goto done;
u128_xor(dst, dst, src - 1);
src -= 1;
dst -= 1;
} while (nbytes >= bsize * 3);
if (nbytes < bsize)
goto done;
}
/* Handle leftovers */
for (;;) {
twofish_dec_blk(ctx, (u8 *)dst, (u8 *)src);
nbytes -= bsize;
if (nbytes < bsize)
break;
u128_xor(dst, dst, src - 1);
src -= 1;
dst -= 1;
}
done:
u128_xor(dst, dst, (u128 *)walk->iv);
*(u128 *)walk->iv = last_iv;
return nbytes;
}
static int cbc_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
struct scatterlist *src, unsigned int nbytes)
{
struct blkcipher_walk walk;
int err;
blkcipher_walk_init(&walk, dst, src, nbytes);
err = blkcipher_walk_virt(desc, &walk);
while ((nbytes = walk.nbytes)) {
nbytes = __cbc_decrypt(desc, &walk);
err = blkcipher_walk_done(desc, &walk, nbytes);
}
return err;
}
static struct crypto_alg blk_cbc_alg = {
.cra_name = "cbc(twofish)",
.cra_driver_name = "cbc-twofish-3way",
.cra_priority = 300,
.cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER,
.cra_blocksize = TF_BLOCK_SIZE,
.cra_ctxsize = sizeof(struct twofish_ctx),
.cra_alignmask = 0,
.cra_type = &crypto_blkcipher_type,
.cra_module = THIS_MODULE,
.cra_list = LIST_HEAD_INIT(blk_cbc_alg.cra_list),
.cra_u = {
.blkcipher = {
.min_keysize = TF_MIN_KEY_SIZE,
.max_keysize = TF_MAX_KEY_SIZE,
.ivsize = TF_BLOCK_SIZE,
.setkey = twofish_setkey,
.encrypt = cbc_encrypt,
.decrypt = cbc_decrypt,
},
},
};
static inline void u128_to_be128(be128 *dst, const u128 *src)
{
dst->a = cpu_to_be64(src->a);
dst->b = cpu_to_be64(src->b);
}
static inline void be128_to_u128(u128 *dst, const be128 *src)
{
dst->a = be64_to_cpu(src->a);
dst->b = be64_to_cpu(src->b);
}
static inline void u128_inc(u128 *i)
{
i->b++;
if (!i->b)
i->a++;
}
static void ctr_crypt_final(struct blkcipher_desc *desc,
struct blkcipher_walk *walk)
{
struct twofish_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
u8 *ctrblk = walk->iv;
u8 keystream[TF_BLOCK_SIZE];
u8 *src = walk->src.virt.addr;
u8 *dst = walk->dst.virt.addr;
unsigned int nbytes = walk->nbytes;
twofish_enc_blk(ctx, keystream, ctrblk);
crypto_xor(keystream, src, nbytes);
memcpy(dst, keystream, nbytes);
crypto_inc(ctrblk, TF_BLOCK_SIZE);
}
static unsigned int __ctr_crypt(struct blkcipher_desc *desc,
struct blkcipher_walk *walk)
{
struct twofish_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
unsigned int bsize = TF_BLOCK_SIZE;
unsigned int nbytes = walk->nbytes;
u128 *src = (u128 *)walk->src.virt.addr;
u128 *dst = (u128 *)walk->dst.virt.addr;
u128 ctrblk;
be128 ctrblocks[3];
be128_to_u128(&ctrblk, (be128 *)walk->iv);
/* Process three block batch */
if (nbytes >= bsize * 3) {
do {
if (dst != src) {
dst[0] = src[0];
dst[1] = src[1];
dst[2] = src[2];
}
/* create ctrblks for parallel encrypt */
u128_to_be128(&ctrblocks[0], &ctrblk);
u128_inc(&ctrblk);
u128_to_be128(&ctrblocks[1], &ctrblk);
u128_inc(&ctrblk);
u128_to_be128(&ctrblocks[2], &ctrblk);
u128_inc(&ctrblk);
twofish_enc_blk_xor_3way(ctx, (u8 *)dst,
(u8 *)ctrblocks);
src += 3;
dst += 3;
nbytes -= bsize * 3;
} while (nbytes >= bsize * 3);
if (nbytes < bsize)
goto done;
}
/* Handle leftovers */
do {
if (dst != src)
*dst = *src;
u128_to_be128(&ctrblocks[0], &ctrblk);
u128_inc(&ctrblk);
twofish_enc_blk(ctx, (u8 *)ctrblocks, (u8 *)ctrblocks);
u128_xor(dst, dst, (u128 *)ctrblocks);
src += 1;
dst += 1;
nbytes -= bsize;
} while (nbytes >= bsize);
done:
u128_to_be128((be128 *)walk->iv, &ctrblk);
return nbytes;
}
static int ctr_crypt(struct blkcipher_desc *desc, struct scatterlist *dst,
struct scatterlist *src, unsigned int nbytes)
{
struct blkcipher_walk walk;
int err;
blkcipher_walk_init(&walk, dst, src, nbytes);
err = blkcipher_walk_virt_block(desc, &walk, TF_BLOCK_SIZE);
while ((nbytes = walk.nbytes) >= TF_BLOCK_SIZE) {
nbytes = __ctr_crypt(desc, &walk);
err = blkcipher_walk_done(desc, &walk, nbytes);
}
if (walk.nbytes) {
ctr_crypt_final(desc, &walk);
err = blkcipher_walk_done(desc, &walk, 0);
}
return err;
}
static struct crypto_alg blk_ctr_alg = {
.cra_name = "ctr(twofish)",
.cra_driver_name = "ctr-twofish-3way",
.cra_priority = 300,
.cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER,
.cra_blocksize = TF_BLOCK_SIZE,
.cra_ctxsize = sizeof(struct twofish_ctx),
.cra_alignmask = 0,
.cra_type = &crypto_blkcipher_type,
.cra_module = THIS_MODULE,
.cra_list = LIST_HEAD_INIT(blk_ctr_alg.cra_list),
.cra_u = {
.blkcipher = {
.min_keysize = TF_MIN_KEY_SIZE,
.max_keysize = TF_MAX_KEY_SIZE,
.ivsize = TF_BLOCK_SIZE,
.setkey = twofish_setkey,
.encrypt = ctr_crypt,
.decrypt = ctr_crypt,
},
},
};
int __init init(void)
{
int err;
err = crypto_register_alg(&blk_ecb_alg);
if (err)
goto ecb_err;
err = crypto_register_alg(&blk_cbc_alg);
if (err)
goto cbc_err;
err = crypto_register_alg(&blk_ctr_alg);
if (err)
goto ctr_err;
return 0;
ctr_err:
crypto_unregister_alg(&blk_cbc_alg);
cbc_err:
crypto_unregister_alg(&blk_ecb_alg);
ecb_err:
return err;
}
void __exit fini(void)
{
crypto_unregister_alg(&blk_ctr_alg);
crypto_unregister_alg(&blk_cbc_alg);
crypto_unregister_alg(&blk_ecb_alg);
}
module_init(init);
module_exit(fini);
MODULE_LICENSE("GPL");
MODULE_DESCRIPTION("Twofish Cipher Algorithm, 3-way parallel asm optimized");
MODULE_ALIAS("twofish");
MODULE_ALIAS("twofish-asm");

View file

@ -828,6 +828,26 @@ config CRYPTO_TWOFISH_X86_64
See also:
<http://www.schneier.com/twofish.html>
config CRYPTO_TWOFISH_X86_64_3WAY
tristate "Twofish cipher algorithm (x86_64, 3-way parallel)"
depends on (X86 || UML_X86) && 64BIT
select CRYPTO_ALGAPI
select CRYPTO_TWOFISH_COMMON
select CRYPTO_TWOFISH_X86_64
help
Twofish cipher algorithm (x86_64, 3-way parallel).
Twofish was submitted as an AES (Advanced Encryption Standard)
candidate cipher by researchers at CounterPane Systems. It is a
16 round block cipher supporting key sizes of 128, 192, and 256
bits.
This module provides Twofish cipher algorithm that processes three
blocks parallel, utilizing resources of out-of-order CPUs better.
See also:
<http://www.schneier.com/twofish.html>
comment "Compression"
config CRYPTO_DEFLATE