kernel_optimize_test/arch/x86/crypto/aes_64.c

337 lines
8.5 KiB
C
Raw Normal View History

[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
/*
* Cryptographic API.
*
* AES Cipher Algorithm.
*
* Based on Brian Gladman's code.
*
* Linux developers:
* Alexander Kjeldaas <astor@fast.no>
* Herbert Valerio Riedel <hvr@hvrlab.org>
* Kyle McMartin <kyle@debian.org>
* Adam J. Richter <adam@yggdrasil.com> (conversion to 2.5 API).
* Andreas Steinmetz <ast@domdv.de> (adapted to x86_64 assembler)
*
* 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.
*
* ---------------------------------------------------------------------------
* Copyright (c) 2002, Dr Brian Gladman <brg@gladman.me.uk>, Worcester, UK.
* All rights reserved.
*
* LICENSE TERMS
*
* The free distribution and use of this software in both source and binary
* form is allowed (with or without changes) provided that:
*
* 1. distributions of this source code include the above copyright
* notice, this list of conditions and the following disclaimer;
*
* 2. distributions in binary form include the above copyright
* notice, this list of conditions and the following disclaimer
* in the documentation and/or other associated materials;
*
* 3. the copyright holder's name is not used to endorse products
* built using this software without specific written permission.
*
* ALTERNATIVELY, provided that this notice is retained in full, this product
* may be distributed under the terms of the GNU General Public License (GPL),
* in which case the provisions of the GPL apply INSTEAD OF those given above.
*
* DISCLAIMER
*
* This software is provided 'as is' with no explicit or implied warranties
* in respect of its properties, including, but not limited to, correctness
* and/or fitness for purpose.
* ---------------------------------------------------------------------------
*/
/* Some changes from the Gladman version:
s/RIJNDAEL(e_key)/E_KEY/g
s/RIJNDAEL(d_key)/D_KEY/g
*/
#include <asm/byteorder.h>
#include <linux/bitops.h>
#include <linux/crypto.h>
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/types.h>
#define AES_MIN_KEY_SIZE 16
#define AES_MAX_KEY_SIZE 32
#define AES_BLOCK_SIZE 16
/*
* #define byte(x, nr) ((unsigned char)((x) >> (nr*8)))
*/
static inline u8 byte(const u32 x, const unsigned n)
{
return x >> (n << 3);
}
struct aes_ctx
{
u32 key_length;
u32 buf[120];
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
};
#define E_KEY (&ctx->buf[0])
#define D_KEY (&ctx->buf[60])
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
static u8 pow_tab[256] __initdata;
static u8 log_tab[256] __initdata;
static u8 sbx_tab[256] __initdata;
static u8 isb_tab[256] __initdata;
static u32 rco_tab[10];
u32 aes_ft_tab[4][256];
u32 aes_it_tab[4][256];
u32 aes_fl_tab[4][256];
u32 aes_il_tab[4][256];
static inline u8 f_mult(u8 a, u8 b)
{
u8 aa = log_tab[a], cc = aa + log_tab[b];
return pow_tab[cc + (cc < aa ? 1 : 0)];
}
#define ff_mult(a, b) (a && b ? f_mult(a, b) : 0)
#define ls_box(x) \
(aes_fl_tab[0][byte(x, 0)] ^ \
aes_fl_tab[1][byte(x, 1)] ^ \
aes_fl_tab[2][byte(x, 2)] ^ \
aes_fl_tab[3][byte(x, 3)])
static void __init gen_tabs(void)
{
u32 i, t;
u8 p, q;
/* log and power tables for GF(2**8) finite field with
0x011b as modular polynomial - the simplest primitive
root is 0x03, used here to generate the tables */
for (i = 0, p = 1; i < 256; ++i) {
pow_tab[i] = (u8)p;
log_tab[p] = (u8)i;
p ^= (p << 1) ^ (p & 0x80 ? 0x01b : 0);
}
log_tab[1] = 0;
for (i = 0, p = 1; i < 10; ++i) {
rco_tab[i] = p;
p = (p << 1) ^ (p & 0x80 ? 0x01b : 0);
}
for (i = 0; i < 256; ++i) {
p = (i ? pow_tab[255 - log_tab[i]] : 0);
q = ((p >> 7) | (p << 1)) ^ ((p >> 6) | (p << 2));
p ^= 0x63 ^ q ^ ((q >> 6) | (q << 2));
sbx_tab[i] = p;
isb_tab[p] = (u8)i;
}
for (i = 0; i < 256; ++i) {
p = sbx_tab[i];
t = p;
aes_fl_tab[0][i] = t;
aes_fl_tab[1][i] = rol32(t, 8);
aes_fl_tab[2][i] = rol32(t, 16);
aes_fl_tab[3][i] = rol32(t, 24);
t = ((u32)ff_mult(2, p)) |
((u32)p << 8) |
((u32)p << 16) | ((u32)ff_mult(3, p) << 24);
aes_ft_tab[0][i] = t;
aes_ft_tab[1][i] = rol32(t, 8);
aes_ft_tab[2][i] = rol32(t, 16);
aes_ft_tab[3][i] = rol32(t, 24);
p = isb_tab[i];
t = p;
aes_il_tab[0][i] = t;
aes_il_tab[1][i] = rol32(t, 8);
aes_il_tab[2][i] = rol32(t, 16);
aes_il_tab[3][i] = rol32(t, 24);
t = ((u32)ff_mult(14, p)) |
((u32)ff_mult(9, p) << 8) |
((u32)ff_mult(13, p) << 16) |
((u32)ff_mult(11, p) << 24);
aes_it_tab[0][i] = t;
aes_it_tab[1][i] = rol32(t, 8);
aes_it_tab[2][i] = rol32(t, 16);
aes_it_tab[3][i] = rol32(t, 24);
}
}
#define star_x(x) (((x) & 0x7f7f7f7f) << 1) ^ ((((x) & 0x80808080) >> 7) * 0x1b)
#define imix_col(y, x) \
u = star_x(x); \
v = star_x(u); \
w = star_x(v); \
t = w ^ (x); \
(y) = u ^ v ^ w; \
(y) ^= ror32(u ^ t, 8) ^ \
ror32(v ^ t, 16) ^ \
ror32(t, 24)
/* initialise the key schedule from the user supplied key */
#define loop4(i) \
{ \
t = ror32(t, 8); t = ls_box(t) ^ rco_tab[i]; \
t ^= E_KEY[4 * i]; E_KEY[4 * i + 4] = t; \
t ^= E_KEY[4 * i + 1]; E_KEY[4 * i + 5] = t; \
t ^= E_KEY[4 * i + 2]; E_KEY[4 * i + 6] = t; \
t ^= E_KEY[4 * i + 3]; E_KEY[4 * i + 7] = t; \
}
#define loop6(i) \
{ \
t = ror32(t, 8); t = ls_box(t) ^ rco_tab[i]; \
t ^= E_KEY[6 * i]; E_KEY[6 * i + 6] = t; \
t ^= E_KEY[6 * i + 1]; E_KEY[6 * i + 7] = t; \
t ^= E_KEY[6 * i + 2]; E_KEY[6 * i + 8] = t; \
t ^= E_KEY[6 * i + 3]; E_KEY[6 * i + 9] = t; \
t ^= E_KEY[6 * i + 4]; E_KEY[6 * i + 10] = t; \
t ^= E_KEY[6 * i + 5]; E_KEY[6 * i + 11] = t; \
}
#define loop8(i) \
{ \
t = ror32(t, 8); ; t = ls_box(t) ^ rco_tab[i]; \
t ^= E_KEY[8 * i]; E_KEY[8 * i + 8] = t; \
t ^= E_KEY[8 * i + 1]; E_KEY[8 * i + 9] = t; \
t ^= E_KEY[8 * i + 2]; E_KEY[8 * i + 10] = t; \
t ^= E_KEY[8 * i + 3]; E_KEY[8 * i + 11] = t; \
t = E_KEY[8 * i + 4] ^ ls_box(t); \
E_KEY[8 * i + 12] = t; \
t ^= E_KEY[8 * i + 5]; E_KEY[8 * i + 13] = t; \
t ^= E_KEY[8 * i + 6]; E_KEY[8 * i + 14] = t; \
t ^= E_KEY[8 * i + 7]; E_KEY[8 * i + 15] = t; \
}
static int aes_set_key(struct crypto_tfm *tfm, const u8 *in_key,
unsigned int key_len)
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
{
struct aes_ctx *ctx = crypto_tfm_ctx(tfm);
const __le32 *key = (const __le32 *)in_key;
u32 *flags = &tfm->crt_flags;
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
u32 i, j, t, u, v, w;
if (key_len % 8) {
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
*flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
return -EINVAL;
}
ctx->key_length = key_len;
D_KEY[key_len + 24] = E_KEY[0] = le32_to_cpu(key[0]);
D_KEY[key_len + 25] = E_KEY[1] = le32_to_cpu(key[1]);
D_KEY[key_len + 26] = E_KEY[2] = le32_to_cpu(key[2]);
D_KEY[key_len + 27] = E_KEY[3] = le32_to_cpu(key[3]);
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
switch (key_len) {
case 16:
t = E_KEY[3];
for (i = 0; i < 10; ++i)
loop4(i);
break;
case 24:
E_KEY[4] = le32_to_cpu(key[4]);
t = E_KEY[5] = le32_to_cpu(key[5]);
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
for (i = 0; i < 8; ++i)
loop6 (i);
break;
case 32:
E_KEY[4] = le32_to_cpu(key[4]);
E_KEY[5] = le32_to_cpu(key[5]);
E_KEY[6] = le32_to_cpu(key[6]);
t = E_KEY[7] = le32_to_cpu(key[7]);
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
for (i = 0; i < 7; ++i)
loop8(i);
break;
}
D_KEY[0] = E_KEY[key_len + 24];
D_KEY[1] = E_KEY[key_len + 25];
D_KEY[2] = E_KEY[key_len + 26];
D_KEY[3] = E_KEY[key_len + 27];
for (i = 4; i < key_len + 24; ++i) {
j = key_len + 24 - (i & ~3) + (i & 3);
imix_col(D_KEY[j], E_KEY[i]);
}
return 0;
}
asmlinkage void aes_enc_blk(struct crypto_tfm *tfm, u8 *out, const u8 *in);
asmlinkage void aes_dec_blk(struct crypto_tfm *tfm, u8 *out, const u8 *in);
static void aes_encrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
{
aes_enc_blk(tfm, dst, src);
}
static void aes_decrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
{
aes_dec_blk(tfm, dst, src);
}
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
static struct crypto_alg aes_alg = {
.cra_name = "aes",
.cra_driver_name = "aes-x86_64",
.cra_priority = 200,
[CRYPTO] Add x86_64 asm AES Implementation: =============== The encrypt/decrypt code is based on an x86 implementation I did a while ago which I never published. This unpublished implementation does include an assembler based key schedule and precomputed tables. For simplicity and best acceptance, however, I took Gladman's in-kernel code for table generation and key schedule for the kernel port of my assembler code and modified this code to produce the key schedule as required by my assembler implementation. File locations and Kconfig are kept similar to the i586 AES assembler implementation. It may seem a little bit strange to use 32 bit I/O and registers in the assembler implementation but this gives the best code size. My implementation takes one instruction more per round compared to Gladman's x86 assembler but it doesn't require any stack for local variables or saved registers and it is less serialized than Gladman's code. Note that all comparisons to Gladman's code were done after my code was implemented. I did only use FIPS PUB 197 for the implementation so my implementation is independent work. If anybody has a better assembler solution for x86_64 I'll be pleased to have my code replaced with the better solution. Testing: ======== The implementation passes the in-kernel crypto testing module and I'm running it without any problems on my laptop where it is mainly used for dm-crypt. Microbenchmark: =============== The microbenchmark was done in userspace with similar compile flags as used during kernel compile. Encrypt/decrypt is about 35% faster than the generic C implementation. As the generic C as well as my assembler implementation are both table I don't really expect that there is much room for further improvements though I'll be glad to be corrected here. The key schedule is about 5% slower than the generic C implementation. This is due to the fact that some more work has to be done in the key schedule routine to fit the schedule to the assembler implementation. Code Size: ========== Encrypt and decrypt are together about 2.1 Kbytes smaller than the generic C implementation which is important with regard to L1 cache usage. The key schedule routine is about 100 bytes larger than the generic C implementation. Data Size: ========== There's no difference in data size requirements between the assembler implementation and the generic C implementation. License: ======== Gladmans's code is dual BSD/GPL whereas my assembler code is GPLv2 only (I'm not going to change the license for my code). So I had to change the module license for the x86_64 aes module from 'Dual BSD/GPL' to 'GPL' to reflect the most restrictive license within the module. Signed-off-by: Andreas Steinmetz <ast@domdv.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-07 04:55:00 +08:00
.cra_flags = CRYPTO_ALG_TYPE_CIPHER,
.cra_blocksize = AES_BLOCK_SIZE,
.cra_ctxsize = sizeof(struct aes_ctx),
.cra_module = THIS_MODULE,
.cra_list = LIST_HEAD_INIT(aes_alg.cra_list),
.cra_u = {
.cipher = {
.cia_min_keysize = AES_MIN_KEY_SIZE,
.cia_max_keysize = AES_MAX_KEY_SIZE,
.cia_setkey = aes_set_key,
.cia_encrypt = aes_encrypt,
.cia_decrypt = aes_decrypt
}
}
};
static int __init aes_init(void)
{
gen_tabs();
return crypto_register_alg(&aes_alg);
}
static void __exit aes_fini(void)
{
crypto_unregister_alg(&aes_alg);
}
module_init(aes_init);
module_exit(aes_fini);
MODULE_DESCRIPTION("Rijndael (AES) Cipher Algorithm");
MODULE_LICENSE("GPL");
MODULE_ALIAS("aes");