2021-07-20 03:48:54

by Tianjia Zhang

[permalink] [raw]
Subject: [PATCH v3 0/4] Introduce x86 assembler accelerated implementation for SM4 algorithm

This patchset extracts the public SM4 algorithm as a separate library,
At the same time, the acceleration implementation of SM4 in arm64 was
adjusted to adapt to this SM4 library. Then introduces an accelerated
implementation of the instruction set on x86_64.

This optimization supports the four modes of SM4, ECB, CBC, CFB, and
CTR. Since CBC and CFB do not support multiple block parallel
encryption, the optimization effect is not obvious. And all selftests
have passed already.

The main algorithm implementation comes from SM4 AES-NI work by
libgcrypt and Markku-Juhani O. Saarinen at:
https://github.com/mjosaarinen/sm4ni

Benchmark on Intel Xeon Cascadelake, the data comes from the mode 218
and mode 518 of tcrypt. The abscissas are blocks of different lengths.
The data is tabulated and the unit is Mb/s:

sm4-generic | 16 64 128 256 1024 1420 4096
ECB enc | 40.99 46.50 48.05 48.41 49.20 49.25 49.28
ECB dec | 41.07 46.99 48.15 48.67 49.20 49.25 49.29
CBC enc | 37.71 45.28 46.77 47.60 48.32 48.37 48.40
CBC dec | 36.48 44.82 46.43 47.45 48.23 48.30 48.36
CFB enc | 37.94 44.84 46.12 46.94 47.57 47.46 47.68
CFB dec | 37.50 42.84 43.74 44.37 44.85 44.80 44.96
CTR enc | 39.20 45.63 46.75 47.49 48.09 47.85 48.08
CTR dec | 39.64 45.70 46.72 47.47 47.98 47.88 48.06
sm4-aesni-avx
ECB enc | 33.75 134.47 221.64 243.43 264.05 251.58 258.13
ECB dec | 34.02 134.92 223.11 245.14 264.12 251.04 258.33
CBC enc | 38.85 46.18 47.67 48.34 49.00 48.96 49.14
CBC dec | 33.54 131.29 223.88 245.27 265.50 252.41 263.78
CFB enc | 38.70 46.10 47.58 48.29 49.01 48.94 49.19
CFB dec | 32.79 128.40 223.23 244.87 265.77 253.31 262.79
CTR enc | 32.58 122.23 220.29 241.16 259.57 248.32 256.69
CTR dec | 32.81 122.47 218.99 241.54 258.42 248.58 256.61

---
v3 changes:
* Remove single block algorithm that does not greatly improve performance
* Remove accelerated for sm4 key expand, which is not performance-critical
* Fix the warning on arm64/sm4-ce

v2 changes:
* SM4 library functions use "sm4_" prefix instead of "crypto_" prefix
* sm4-aesni-avx supports accelerated implementation of four specific modes
* tcrypt benchmark supports sm4-aesni-avx
* fixes of other reviews


Tianjia Zhang (4):
crypto: sm4 - create SM4 library based on sm4 generic code
crypto: arm64/sm4-ce - Make dependent on sm4 library instead of
sm4-generic
crypto: x86/sm4 - add AES-NI/AVX/x86_64 implementation
crypto: tcrypt - add the asynchronous speed test for SM4

arch/arm64/crypto/Kconfig | 2 +-
arch/arm64/crypto/sm4-ce-glue.c | 20 +-
arch/x86/crypto/Makefile | 3 +
arch/x86/crypto/sm4-aesni-avx-asm_64.S | 589 +++++++++++++++++++++++++
arch/x86/crypto/sm4_aesni_avx_glue.c | 459 +++++++++++++++++++
crypto/Kconfig | 22 +
crypto/sm4_generic.c | 180 +-------
crypto/tcrypt.c | 26 +-
include/crypto/sm4.h | 25 +-
lib/crypto/Kconfig | 3 +
lib/crypto/Makefile | 3 +
lib/crypto/sm4.c | 176 ++++++++
12 files changed, 1330 insertions(+), 178 deletions(-)
create mode 100644 arch/x86/crypto/sm4-aesni-avx-asm_64.S
create mode 100644 arch/x86/crypto/sm4_aesni_avx_glue.c
create mode 100644 lib/crypto/sm4.c

--
2.19.1.3.ge56e4f7


2021-07-20 03:49:13

by Tianjia Zhang

[permalink] [raw]
Subject: [PATCH v3 2/4] crypto: arm64/sm4-ce - Make dependent on sm4 library instead of sm4-generic

SM4 library is abstracted from sm4-generic algorithm, sm4-ce can depend on
the SM4 library instead of sm4-generic, and some functions in sm4-generic
do not need to be exported.

Signed-off-by: Tianjia Zhang <[email protected]>
---
arch/arm64/crypto/Kconfig | 2 +-
arch/arm64/crypto/sm4-ce-glue.c | 20 ++++++++++++++------
crypto/sm4_generic.c | 27 ++++++++++++---------------
include/crypto/sm4.h | 9 ++-------
lib/crypto/sm4.c | 2 +-
5 files changed, 30 insertions(+), 30 deletions(-)

diff --git a/arch/arm64/crypto/Kconfig b/arch/arm64/crypto/Kconfig
index b8eb0453123d..55f19450091b 100644
--- a/arch/arm64/crypto/Kconfig
+++ b/arch/arm64/crypto/Kconfig
@@ -51,7 +51,7 @@ config CRYPTO_SM4_ARM64_CE
tristate "SM4 symmetric cipher (ARMv8.2 Crypto Extensions)"
depends on KERNEL_MODE_NEON
select CRYPTO_ALGAPI
- select CRYPTO_SM4
+ select CRYPTO_LIB_SM4

config CRYPTO_GHASH_ARM64_CE
tristate "GHASH/AES-GCM using ARMv8 Crypto Extensions"
diff --git a/arch/arm64/crypto/sm4-ce-glue.c b/arch/arm64/crypto/sm4-ce-glue.c
index 2754c875d39c..9c93cfc4841b 100644
--- a/arch/arm64/crypto/sm4-ce-glue.c
+++ b/arch/arm64/crypto/sm4-ce-glue.c
@@ -17,12 +17,20 @@ MODULE_LICENSE("GPL v2");

asmlinkage void sm4_ce_do_crypt(const u32 *rk, void *out, const void *in);

+static int sm4_ce_setkey(struct crypto_tfm *tfm, const u8 *key,
+ unsigned int key_len)
+{
+ struct sm4_ctx *ctx = crypto_tfm_ctx(tfm);
+
+ return sm4_expandkey(ctx, key, key_len);
+}
+
static void sm4_ce_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
{
- const struct crypto_sm4_ctx *ctx = crypto_tfm_ctx(tfm);
+ const struct sm4_ctx *ctx = crypto_tfm_ctx(tfm);

if (!crypto_simd_usable()) {
- crypto_sm4_encrypt(tfm, out, in);
+ sm4_crypt_block(ctx->rkey_enc, out, in);
} else {
kernel_neon_begin();
sm4_ce_do_crypt(ctx->rkey_enc, out, in);
@@ -32,10 +40,10 @@ static void sm4_ce_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)

static void sm4_ce_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
{
- const struct crypto_sm4_ctx *ctx = crypto_tfm_ctx(tfm);
+ const struct sm4_ctx *ctx = crypto_tfm_ctx(tfm);

if (!crypto_simd_usable()) {
- crypto_sm4_decrypt(tfm, out, in);
+ sm4_crypt_block(ctx->rkey_dec, out, in);
} else {
kernel_neon_begin();
sm4_ce_do_crypt(ctx->rkey_dec, out, in);
@@ -49,12 +57,12 @@ static struct crypto_alg sm4_ce_alg = {
.cra_priority = 200,
.cra_flags = CRYPTO_ALG_TYPE_CIPHER,
.cra_blocksize = SM4_BLOCK_SIZE,
- .cra_ctxsize = sizeof(struct crypto_sm4_ctx),
+ .cra_ctxsize = sizeof(struct sm4_ctx),
.cra_module = THIS_MODULE,
.cra_u.cipher = {
.cia_min_keysize = SM4_KEY_SIZE,
.cia_max_keysize = SM4_KEY_SIZE,
- .cia_setkey = crypto_sm4_set_key,
+ .cia_setkey = sm4_ce_setkey,
.cia_encrypt = sm4_ce_encrypt,
.cia_decrypt = sm4_ce_decrypt
}
diff --git a/crypto/sm4_generic.c b/crypto/sm4_generic.c
index d19d01f852a9..4a6480a27fee 100644
--- a/crypto/sm4_generic.c
+++ b/crypto/sm4_generic.c
@@ -17,45 +17,42 @@
#include <asm/unaligned.h>

/**
- * crypto_sm4_set_key - Set the SM4 key.
+ * sm4_setkey - Set the SM4 key.
* @tfm: The %crypto_tfm that is used in the context.
* @in_key: The input key.
* @key_len: The size of the key.
*
* This function uses sm4_expandkey() to expand the key.
- * &crypto_sm4_ctx _must_ be the private data embedded in @tfm which is
+ * &sm4_ctx _must_ be the private data embedded in @tfm which is
* retrieved with crypto_tfm_ctx().
*
* Return: 0 on success; -EINVAL on failure (only happens for bad key lengths)
*/
-int crypto_sm4_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+static int sm4_setkey(struct crypto_tfm *tfm, const u8 *in_key,
unsigned int key_len)
{
- struct crypto_sm4_ctx *ctx = crypto_tfm_ctx(tfm);
+ struct sm4_ctx *ctx = crypto_tfm_ctx(tfm);

return sm4_expandkey(ctx, in_key, key_len);
}
-EXPORT_SYMBOL_GPL(crypto_sm4_set_key);

/* encrypt a block of text */

-void crypto_sm4_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
+static void sm4_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
{
- const struct crypto_sm4_ctx *ctx = crypto_tfm_ctx(tfm);
+ const struct sm4_ctx *ctx = crypto_tfm_ctx(tfm);

sm4_crypt_block(ctx->rkey_enc, out, in);
}
-EXPORT_SYMBOL_GPL(crypto_sm4_encrypt);

/* decrypt a block of text */

-void crypto_sm4_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
+static void sm4_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
{
- const struct crypto_sm4_ctx *ctx = crypto_tfm_ctx(tfm);
+ const struct sm4_ctx *ctx = crypto_tfm_ctx(tfm);

sm4_crypt_block(ctx->rkey_dec, out, in);
}
-EXPORT_SYMBOL_GPL(crypto_sm4_decrypt);

static struct crypto_alg sm4_alg = {
.cra_name = "sm4",
@@ -63,15 +60,15 @@ static struct crypto_alg sm4_alg = {
.cra_priority = 100,
.cra_flags = CRYPTO_ALG_TYPE_CIPHER,
.cra_blocksize = SM4_BLOCK_SIZE,
- .cra_ctxsize = sizeof(struct crypto_sm4_ctx),
+ .cra_ctxsize = sizeof(struct sm4_ctx),
.cra_module = THIS_MODULE,
.cra_u = {
.cipher = {
.cia_min_keysize = SM4_KEY_SIZE,
.cia_max_keysize = SM4_KEY_SIZE,
- .cia_setkey = crypto_sm4_set_key,
- .cia_encrypt = crypto_sm4_encrypt,
- .cia_decrypt = crypto_sm4_decrypt
+ .cia_setkey = sm4_setkey,
+ .cia_encrypt = sm4_encrypt,
+ .cia_decrypt = sm4_decrypt
}
}
};
diff --git a/include/crypto/sm4.h b/include/crypto/sm4.h
index 06322325f862..709f286e7b25 100644
--- a/include/crypto/sm4.h
+++ b/include/crypto/sm4.h
@@ -16,7 +16,7 @@
#define SM4_BLOCK_SIZE 16
#define SM4_RKEY_WORDS 32

-struct crypto_sm4_ctx {
+struct sm4_ctx {
u32 rkey_enc[SM4_RKEY_WORDS];
u32 rkey_dec[SM4_RKEY_WORDS];
};
@@ -30,7 +30,7 @@ struct crypto_sm4_ctx {
* Returns 0 on success. The function fails only if an invalid key size (or
* pointer) is supplied.
*/
-int sm4_expandkey(struct crypto_sm4_ctx *ctx, const u8 *in_key,
+int sm4_expandkey(struct sm4_ctx *ctx, const u8 *in_key,
unsigned int key_len);

/**
@@ -41,9 +41,4 @@ int sm4_expandkey(struct crypto_sm4_ctx *ctx, const u8 *in_key,
*/
void sm4_crypt_block(const u32 *rk, u8 *out, const u8 *in);

-int crypto_sm4_set_key(struct crypto_tfm *tfm, const u8 *in_key,
- unsigned int key_len);
-void crypto_sm4_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in);
-void crypto_sm4_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in);
-
#endif
diff --git a/lib/crypto/sm4.c b/lib/crypto/sm4.c
index 5fbf8c741a2f..633b59fed9db 100644
--- a/lib/crypto/sm4.c
+++ b/lib/crypto/sm4.c
@@ -108,7 +108,7 @@ static inline u32 sm4_round(u32 x0, u32 x1, u32 x2, u32 x3, u32 rk)
* Returns 0 on success. The function fails only if an invalid key size (or
* pointer) is supplied.
*/
-int sm4_expandkey(struct crypto_sm4_ctx *ctx, const u8 *in_key,
+int sm4_expandkey(struct sm4_ctx *ctx, const u8 *in_key,
unsigned int key_len)
{
u32 rk[4];
--
2.19.1.3.ge56e4f7

2021-07-20 03:50:12

by Tianjia Zhang

[permalink] [raw]
Subject: [PATCH v3 1/4] crypto: sm4 - create SM4 library based on sm4 generic code

Take the existing small footprint and mostly time invariant C code
and turn it into a SM4 library that can be used for non-performance
critical, casual use of SM4, and as a fallback for, e.g., SIMD code
that needs a secondary path that can be taken in contexts where the
SIMD unit is off limits.

Secondly, some codes have been optimized, such as unrolling small
times loop, removing unnecessary memory shifts, exporting sbox, fk,
ck arrays, and basic encryption and decryption functions.

Signed-off-by: Tianjia Zhang <[email protected]>
---
crypto/Kconfig | 1 +
crypto/sm4_generic.c | 153 +------------------------------------
include/crypto/sm4.h | 24 +++++-
lib/crypto/Kconfig | 3 +
lib/crypto/Makefile | 3 +
lib/crypto/sm4.c | 176 +++++++++++++++++++++++++++++++++++++++++++
6 files changed, 208 insertions(+), 152 deletions(-)
create mode 100644 lib/crypto/sm4.c

diff --git a/crypto/Kconfig b/crypto/Kconfig
index ca3b02dcbbfa..4fbc9c080ca9 100644
--- a/crypto/Kconfig
+++ b/crypto/Kconfig
@@ -1547,6 +1547,7 @@ config CRYPTO_SERPENT_AVX2_X86_64
config CRYPTO_SM4
tristate "SM4 cipher algorithm"
select CRYPTO_ALGAPI
+ select CRYPTO_LIB_SM4
help
SM4 cipher algorithms (OSCCA GB/T 32907-2016).

diff --git a/crypto/sm4_generic.c b/crypto/sm4_generic.c
index 016dbc595705..d19d01f852a9 100644
--- a/crypto/sm4_generic.c
+++ b/crypto/sm4_generic.c
@@ -16,139 +16,13 @@
#include <asm/byteorder.h>
#include <asm/unaligned.h>

-static const u32 fk[4] = {
- 0xa3b1bac6, 0x56aa3350, 0x677d9197, 0xb27022dc
-};
-
-static const u8 sbox[256] = {
- 0xd6, 0x90, 0xe9, 0xfe, 0xcc, 0xe1, 0x3d, 0xb7,
- 0x16, 0xb6, 0x14, 0xc2, 0x28, 0xfb, 0x2c, 0x05,
- 0x2b, 0x67, 0x9a, 0x76, 0x2a, 0xbe, 0x04, 0xc3,
- 0xaa, 0x44, 0x13, 0x26, 0x49, 0x86, 0x06, 0x99,
- 0x9c, 0x42, 0x50, 0xf4, 0x91, 0xef, 0x98, 0x7a,
- 0x33, 0x54, 0x0b, 0x43, 0xed, 0xcf, 0xac, 0x62,
- 0xe4, 0xb3, 0x1c, 0xa9, 0xc9, 0x08, 0xe8, 0x95,
- 0x80, 0xdf, 0x94, 0xfa, 0x75, 0x8f, 0x3f, 0xa6,
- 0x47, 0x07, 0xa7, 0xfc, 0xf3, 0x73, 0x17, 0xba,
- 0x83, 0x59, 0x3c, 0x19, 0xe6, 0x85, 0x4f, 0xa8,
- 0x68, 0x6b, 0x81, 0xb2, 0x71, 0x64, 0xda, 0x8b,
- 0xf8, 0xeb, 0x0f, 0x4b, 0x70, 0x56, 0x9d, 0x35,
- 0x1e, 0x24, 0x0e, 0x5e, 0x63, 0x58, 0xd1, 0xa2,
- 0x25, 0x22, 0x7c, 0x3b, 0x01, 0x21, 0x78, 0x87,
- 0xd4, 0x00, 0x46, 0x57, 0x9f, 0xd3, 0x27, 0x52,
- 0x4c, 0x36, 0x02, 0xe7, 0xa0, 0xc4, 0xc8, 0x9e,
- 0xea, 0xbf, 0x8a, 0xd2, 0x40, 0xc7, 0x38, 0xb5,
- 0xa3, 0xf7, 0xf2, 0xce, 0xf9, 0x61, 0x15, 0xa1,
- 0xe0, 0xae, 0x5d, 0xa4, 0x9b, 0x34, 0x1a, 0x55,
- 0xad, 0x93, 0x32, 0x30, 0xf5, 0x8c, 0xb1, 0xe3,
- 0x1d, 0xf6, 0xe2, 0x2e, 0x82, 0x66, 0xca, 0x60,
- 0xc0, 0x29, 0x23, 0xab, 0x0d, 0x53, 0x4e, 0x6f,
- 0xd5, 0xdb, 0x37, 0x45, 0xde, 0xfd, 0x8e, 0x2f,
- 0x03, 0xff, 0x6a, 0x72, 0x6d, 0x6c, 0x5b, 0x51,
- 0x8d, 0x1b, 0xaf, 0x92, 0xbb, 0xdd, 0xbc, 0x7f,
- 0x11, 0xd9, 0x5c, 0x41, 0x1f, 0x10, 0x5a, 0xd8,
- 0x0a, 0xc1, 0x31, 0x88, 0xa5, 0xcd, 0x7b, 0xbd,
- 0x2d, 0x74, 0xd0, 0x12, 0xb8, 0xe5, 0xb4, 0xb0,
- 0x89, 0x69, 0x97, 0x4a, 0x0c, 0x96, 0x77, 0x7e,
- 0x65, 0xb9, 0xf1, 0x09, 0xc5, 0x6e, 0xc6, 0x84,
- 0x18, 0xf0, 0x7d, 0xec, 0x3a, 0xdc, 0x4d, 0x20,
- 0x79, 0xee, 0x5f, 0x3e, 0xd7, 0xcb, 0x39, 0x48
-};
-
-static const u32 ck[] = {
- 0x00070e15, 0x1c232a31, 0x383f464d, 0x545b6269,
- 0x70777e85, 0x8c939aa1, 0xa8afb6bd, 0xc4cbd2d9,
- 0xe0e7eef5, 0xfc030a11, 0x181f262d, 0x343b4249,
- 0x50575e65, 0x6c737a81, 0x888f969d, 0xa4abb2b9,
- 0xc0c7ced5, 0xdce3eaf1, 0xf8ff060d, 0x141b2229,
- 0x30373e45, 0x4c535a61, 0x686f767d, 0x848b9299,
- 0xa0a7aeb5, 0xbcc3cad1, 0xd8dfe6ed, 0xf4fb0209,
- 0x10171e25, 0x2c333a41, 0x484f565d, 0x646b7279
-};
-
-static u32 sm4_t_non_lin_sub(u32 x)
-{
- int i;
- u8 *b = (u8 *)&x;
-
- for (i = 0; i < 4; ++i)
- b[i] = sbox[b[i]];
-
- return x;
-}
-
-static u32 sm4_key_lin_sub(u32 x)
-{
- return x ^ rol32(x, 13) ^ rol32(x, 23);
-
-}
-
-static u32 sm4_enc_lin_sub(u32 x)
-{
- return x ^ rol32(x, 2) ^ rol32(x, 10) ^ rol32(x, 18) ^ rol32(x, 24);
-}
-
-static u32 sm4_key_sub(u32 x)
-{
- return sm4_key_lin_sub(sm4_t_non_lin_sub(x));
-}
-
-static u32 sm4_enc_sub(u32 x)
-{
- return sm4_enc_lin_sub(sm4_t_non_lin_sub(x));
-}
-
-static u32 sm4_round(const u32 *x, const u32 rk)
-{
- return x[0] ^ sm4_enc_sub(x[1] ^ x[2] ^ x[3] ^ rk);
-}
-
-
-/**
- * crypto_sm4_expand_key - Expands the SM4 key as described in GB/T 32907-2016
- * @ctx: The location where the computed key will be stored.
- * @in_key: The supplied key.
- * @key_len: The length of the supplied key.
- *
- * Returns 0 on success. The function fails only if an invalid key size (or
- * pointer) is supplied.
- */
-int crypto_sm4_expand_key(struct crypto_sm4_ctx *ctx, const u8 *in_key,
- unsigned int key_len)
-{
- u32 rk[4], t;
- const u32 *key = (u32 *)in_key;
- int i;
-
- if (key_len != SM4_KEY_SIZE)
- return -EINVAL;
-
- for (i = 0; i < 4; ++i)
- rk[i] = get_unaligned_be32(&key[i]) ^ fk[i];
-
- for (i = 0; i < 32; ++i) {
- t = rk[0] ^ sm4_key_sub(rk[1] ^ rk[2] ^ rk[3] ^ ck[i]);
- ctx->rkey_enc[i] = t;
- rk[0] = rk[1];
- rk[1] = rk[2];
- rk[2] = rk[3];
- rk[3] = t;
- }
-
- for (i = 0; i < 32; ++i)
- ctx->rkey_dec[i] = ctx->rkey_enc[31 - i];
-
- return 0;
-}
-EXPORT_SYMBOL_GPL(crypto_sm4_expand_key);
-
/**
* crypto_sm4_set_key - Set the SM4 key.
* @tfm: The %crypto_tfm that is used in the context.
* @in_key: The input key.
* @key_len: The size of the key.
*
- * This function uses crypto_sm4_expand_key() to expand the key.
+ * This function uses sm4_expandkey() to expand the key.
* &crypto_sm4_ctx _must_ be the private data embedded in @tfm which is
* retrieved with crypto_tfm_ctx().
*
@@ -159,36 +33,17 @@ int crypto_sm4_set_key(struct crypto_tfm *tfm, const u8 *in_key,
{
struct crypto_sm4_ctx *ctx = crypto_tfm_ctx(tfm);

- return crypto_sm4_expand_key(ctx, in_key, key_len);
+ return sm4_expandkey(ctx, in_key, key_len);
}
EXPORT_SYMBOL_GPL(crypto_sm4_set_key);

-static void sm4_do_crypt(const u32 *rk, u32 *out, const u32 *in)
-{
- u32 x[4], i, t;
-
- for (i = 0; i < 4; ++i)
- x[i] = get_unaligned_be32(&in[i]);
-
- for (i = 0; i < 32; ++i) {
- t = sm4_round(x, rk[i]);
- x[0] = x[1];
- x[1] = x[2];
- x[2] = x[3];
- x[3] = t;
- }
-
- for (i = 0; i < 4; ++i)
- put_unaligned_be32(x[3 - i], &out[i]);
-}
-
/* encrypt a block of text */

void crypto_sm4_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
{
const struct crypto_sm4_ctx *ctx = crypto_tfm_ctx(tfm);

- sm4_do_crypt(ctx->rkey_enc, (u32 *)out, (u32 *)in);
+ sm4_crypt_block(ctx->rkey_enc, out, in);
}
EXPORT_SYMBOL_GPL(crypto_sm4_encrypt);

@@ -198,7 +53,7 @@ void crypto_sm4_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
{
const struct crypto_sm4_ctx *ctx = crypto_tfm_ctx(tfm);

- sm4_do_crypt(ctx->rkey_dec, (u32 *)out, (u32 *)in);
+ sm4_crypt_block(ctx->rkey_dec, out, in);
}
EXPORT_SYMBOL_GPL(crypto_sm4_decrypt);

diff --git a/include/crypto/sm4.h b/include/crypto/sm4.h
index 7afd730d16ff..06322325f862 100644
--- a/include/crypto/sm4.h
+++ b/include/crypto/sm4.h
@@ -3,6 +3,7 @@
/*
* Common values for the SM4 algorithm
* Copyright (C) 2018 ARM Limited or its affiliates.
+ * Copyright (c) 2021 Tianjia Zhang <[email protected]>
*/

#ifndef _CRYPTO_SM4_H
@@ -20,11 +21,28 @@ struct crypto_sm4_ctx {
u32 rkey_dec[SM4_RKEY_WORDS];
};

-int crypto_sm4_set_key(struct crypto_tfm *tfm, const u8 *in_key,
- unsigned int key_len);
-int crypto_sm4_expand_key(struct crypto_sm4_ctx *ctx, const u8 *in_key,
+/**
+ * sm4_expandkey - Expands the SM4 key as described in GB/T 32907-2016
+ * @ctx: The location where the computed key will be stored.
+ * @in_key: The supplied key.
+ * @key_len: The length of the supplied key.
+ *
+ * Returns 0 on success. The function fails only if an invalid key size (or
+ * pointer) is supplied.
+ */
+int sm4_expandkey(struct crypto_sm4_ctx *ctx, const u8 *in_key,
unsigned int key_len);

+/**
+ * sm4_crypt_block - Encrypt or decrypt a single SM4 block
+ * @rk: The rkey_enc for encrypt or rkey_dec for decrypt
+ * @out: Buffer to store output data
+ * @in: Buffer containing the input data
+ */
+void sm4_crypt_block(const u32 *rk, u8 *out, const u8 *in);
+
+int crypto_sm4_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+ unsigned int key_len);
void crypto_sm4_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in);
void crypto_sm4_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in);

diff --git a/lib/crypto/Kconfig b/lib/crypto/Kconfig
index 14c032de276e..545ccbddf6a1 100644
--- a/lib/crypto/Kconfig
+++ b/lib/crypto/Kconfig
@@ -128,3 +128,6 @@ config CRYPTO_LIB_CHACHA20POLY1305

config CRYPTO_LIB_SHA256
tristate
+
+config CRYPTO_LIB_SM4
+ tristate
diff --git a/lib/crypto/Makefile b/lib/crypto/Makefile
index 3a435629d9ce..73205ed269ba 100644
--- a/lib/crypto/Makefile
+++ b/lib/crypto/Makefile
@@ -38,6 +38,9 @@ libpoly1305-y += poly1305.o
obj-$(CONFIG_CRYPTO_LIB_SHA256) += libsha256.o
libsha256-y := sha256.o

+obj-$(CONFIG_CRYPTO_LIB_SM4) += libsm4.o
+libsm4-y := sm4.o
+
ifneq ($(CONFIG_CRYPTO_MANAGER_DISABLE_TESTS),y)
libblake2s-y += blake2s-selftest.o
libchacha20poly1305-y += chacha20poly1305-selftest.o
diff --git a/lib/crypto/sm4.c b/lib/crypto/sm4.c
new file mode 100644
index 000000000000..5fbf8c741a2f
--- /dev/null
+++ b/lib/crypto/sm4.c
@@ -0,0 +1,176 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * SM4, as specified in
+ * https://tools.ietf.org/id/draft-ribose-cfrg-sm4-10.html
+ *
+ * Copyright (C) 2018 ARM Limited or its affiliates.
+ * Copyright (c) 2021 Tianjia Zhang <[email protected]>
+ */
+
+#include <linux/module.h>
+#include <asm/unaligned.h>
+#include <crypto/sm4.h>
+
+static const u32 fk[4] = {
+ 0xa3b1bac6, 0x56aa3350, 0x677d9197, 0xb27022dc
+};
+
+static const u32 __cacheline_aligned ck[32] = {
+ 0x00070e15, 0x1c232a31, 0x383f464d, 0x545b6269,
+ 0x70777e85, 0x8c939aa1, 0xa8afb6bd, 0xc4cbd2d9,
+ 0xe0e7eef5, 0xfc030a11, 0x181f262d, 0x343b4249,
+ 0x50575e65, 0x6c737a81, 0x888f969d, 0xa4abb2b9,
+ 0xc0c7ced5, 0xdce3eaf1, 0xf8ff060d, 0x141b2229,
+ 0x30373e45, 0x4c535a61, 0x686f767d, 0x848b9299,
+ 0xa0a7aeb5, 0xbcc3cad1, 0xd8dfe6ed, 0xf4fb0209,
+ 0x10171e25, 0x2c333a41, 0x484f565d, 0x646b7279
+};
+
+static const u8 __cacheline_aligned sbox[256] = {
+ 0xd6, 0x90, 0xe9, 0xfe, 0xcc, 0xe1, 0x3d, 0xb7,
+ 0x16, 0xb6, 0x14, 0xc2, 0x28, 0xfb, 0x2c, 0x05,
+ 0x2b, 0x67, 0x9a, 0x76, 0x2a, 0xbe, 0x04, 0xc3,
+ 0xaa, 0x44, 0x13, 0x26, 0x49, 0x86, 0x06, 0x99,
+ 0x9c, 0x42, 0x50, 0xf4, 0x91, 0xef, 0x98, 0x7a,
+ 0x33, 0x54, 0x0b, 0x43, 0xed, 0xcf, 0xac, 0x62,
+ 0xe4, 0xb3, 0x1c, 0xa9, 0xc9, 0x08, 0xe8, 0x95,
+ 0x80, 0xdf, 0x94, 0xfa, 0x75, 0x8f, 0x3f, 0xa6,
+ 0x47, 0x07, 0xa7, 0xfc, 0xf3, 0x73, 0x17, 0xba,
+ 0x83, 0x59, 0x3c, 0x19, 0xe6, 0x85, 0x4f, 0xa8,
+ 0x68, 0x6b, 0x81, 0xb2, 0x71, 0x64, 0xda, 0x8b,
+ 0xf8, 0xeb, 0x0f, 0x4b, 0x70, 0x56, 0x9d, 0x35,
+ 0x1e, 0x24, 0x0e, 0x5e, 0x63, 0x58, 0xd1, 0xa2,
+ 0x25, 0x22, 0x7c, 0x3b, 0x01, 0x21, 0x78, 0x87,
+ 0xd4, 0x00, 0x46, 0x57, 0x9f, 0xd3, 0x27, 0x52,
+ 0x4c, 0x36, 0x02, 0xe7, 0xa0, 0xc4, 0xc8, 0x9e,
+ 0xea, 0xbf, 0x8a, 0xd2, 0x40, 0xc7, 0x38, 0xb5,
+ 0xa3, 0xf7, 0xf2, 0xce, 0xf9, 0x61, 0x15, 0xa1,
+ 0xe0, 0xae, 0x5d, 0xa4, 0x9b, 0x34, 0x1a, 0x55,
+ 0xad, 0x93, 0x32, 0x30, 0xf5, 0x8c, 0xb1, 0xe3,
+ 0x1d, 0xf6, 0xe2, 0x2e, 0x82, 0x66, 0xca, 0x60,
+ 0xc0, 0x29, 0x23, 0xab, 0x0d, 0x53, 0x4e, 0x6f,
+ 0xd5, 0xdb, 0x37, 0x45, 0xde, 0xfd, 0x8e, 0x2f,
+ 0x03, 0xff, 0x6a, 0x72, 0x6d, 0x6c, 0x5b, 0x51,
+ 0x8d, 0x1b, 0xaf, 0x92, 0xbb, 0xdd, 0xbc, 0x7f,
+ 0x11, 0xd9, 0x5c, 0x41, 0x1f, 0x10, 0x5a, 0xd8,
+ 0x0a, 0xc1, 0x31, 0x88, 0xa5, 0xcd, 0x7b, 0xbd,
+ 0x2d, 0x74, 0xd0, 0x12, 0xb8, 0xe5, 0xb4, 0xb0,
+ 0x89, 0x69, 0x97, 0x4a, 0x0c, 0x96, 0x77, 0x7e,
+ 0x65, 0xb9, 0xf1, 0x09, 0xc5, 0x6e, 0xc6, 0x84,
+ 0x18, 0xf0, 0x7d, 0xec, 0x3a, 0xdc, 0x4d, 0x20,
+ 0x79, 0xee, 0x5f, 0x3e, 0xd7, 0xcb, 0x39, 0x48
+};
+
+static inline u32 sm4_t_non_lin_sub(u32 x)
+{
+ u32 out;
+
+ out = (u32)sbox[x & 0xff];
+ out |= (u32)sbox[(x >> 8) & 0xff] << 8;
+ out |= (u32)sbox[(x >> 16) & 0xff] << 16;
+ out |= (u32)sbox[(x >> 24) & 0xff] << 24;
+
+ return out;
+}
+
+static inline u32 sm4_key_lin_sub(u32 x)
+{
+ return x ^ rol32(x, 13) ^ rol32(x, 23);
+}
+
+static inline u32 sm4_enc_lin_sub(u32 x)
+{
+ return x ^ rol32(x, 2) ^ rol32(x, 10) ^ rol32(x, 18) ^ rol32(x, 24);
+}
+
+static inline u32 sm4_key_sub(u32 x)
+{
+ return sm4_key_lin_sub(sm4_t_non_lin_sub(x));
+}
+
+static inline u32 sm4_enc_sub(u32 x)
+{
+ return sm4_enc_lin_sub(sm4_t_non_lin_sub(x));
+}
+
+static inline u32 sm4_round(u32 x0, u32 x1, u32 x2, u32 x3, u32 rk)
+{
+ return x0 ^ sm4_enc_sub(x1 ^ x2 ^ x3 ^ rk);
+}
+
+
+/**
+ * sm4_expandkey - Expands the SM4 key as described in GB/T 32907-2016
+ * @ctx: The location where the computed key will be stored.
+ * @in_key: The supplied key.
+ * @key_len: The length of the supplied key.
+ *
+ * Returns 0 on success. The function fails only if an invalid key size (or
+ * pointer) is supplied.
+ */
+int sm4_expandkey(struct crypto_sm4_ctx *ctx, const u8 *in_key,
+ unsigned int key_len)
+{
+ u32 rk[4];
+ const u32 *key = (u32 *)in_key;
+ int i;
+
+ if (key_len != SM4_KEY_SIZE)
+ return -EINVAL;
+
+ rk[0] = get_unaligned_be32(&key[0]) ^ fk[0];
+ rk[1] = get_unaligned_be32(&key[1]) ^ fk[1];
+ rk[2] = get_unaligned_be32(&key[2]) ^ fk[2];
+ rk[3] = get_unaligned_be32(&key[3]) ^ fk[3];
+
+ for (i = 0; i < 32; i += 4) {
+ rk[0] ^= sm4_key_sub(rk[1] ^ rk[2] ^ rk[3] ^ ck[i + 0]);
+ rk[1] ^= sm4_key_sub(rk[2] ^ rk[3] ^ rk[0] ^ ck[i + 1]);
+ rk[2] ^= sm4_key_sub(rk[3] ^ rk[0] ^ rk[1] ^ ck[i + 2]);
+ rk[3] ^= sm4_key_sub(rk[0] ^ rk[1] ^ rk[2] ^ ck[i + 3]);
+
+ ctx->rkey_enc[i + 0] = rk[0];
+ ctx->rkey_enc[i + 1] = rk[1];
+ ctx->rkey_enc[i + 2] = rk[2];
+ ctx->rkey_enc[i + 3] = rk[3];
+ ctx->rkey_dec[31 - 0 - i] = rk[0];
+ ctx->rkey_dec[31 - 1 - i] = rk[1];
+ ctx->rkey_dec[31 - 2 - i] = rk[2];
+ ctx->rkey_dec[31 - 3 - i] = rk[3];
+ }
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(sm4_expandkey);
+
+/**
+ * sm4_crypt_block - Encrypt or decrypt a single SM4 block
+ * @rk: The rkey_enc for encrypt or rkey_dec for decrypt
+ * @out: Buffer to store output data
+ * @in: Buffer containing the input data
+ */
+void sm4_crypt_block(const u32 *rk, u8 *out, const u8 *in)
+{
+ u32 x[4], i;
+
+ x[0] = get_unaligned_be32(in + 0 * 4);
+ x[1] = get_unaligned_be32(in + 1 * 4);
+ x[2] = get_unaligned_be32(in + 2 * 4);
+ x[3] = get_unaligned_be32(in + 3 * 4);
+
+ for (i = 0; i < 32; i += 4) {
+ x[0] = sm4_round(x[0], x[1], x[2], x[3], rk[i + 0]);
+ x[1] = sm4_round(x[1], x[2], x[3], x[0], rk[i + 1]);
+ x[2] = sm4_round(x[2], x[3], x[0], x[1], rk[i + 2]);
+ x[3] = sm4_round(x[3], x[0], x[1], x[2], rk[i + 3]);
+ }
+
+ put_unaligned_be32(x[3 - 0], out + 0 * 4);
+ put_unaligned_be32(x[3 - 1], out + 1 * 4);
+ put_unaligned_be32(x[3 - 2], out + 2 * 4);
+ put_unaligned_be32(x[3 - 3], out + 3 * 4);
+}
+EXPORT_SYMBOL_GPL(sm4_crypt_block);
+
+MODULE_DESCRIPTION("Generic SM4 library");
+MODULE_LICENSE("GPL v2");
--
2.19.1.3.ge56e4f7

2021-07-30 03:11:44

by Herbert Xu

[permalink] [raw]
Subject: Re: [PATCH v3 0/4] Introduce x86 assembler accelerated implementation for SM4 algorithm

On Tue, Jul 20, 2021 at 11:46:38AM +0800, Tianjia Zhang wrote:
> This patchset extracts the public SM4 algorithm as a separate library,
> At the same time, the acceleration implementation of SM4 in arm64 was
> adjusted to adapt to this SM4 library. Then introduces an accelerated
> implementation of the instruction set on x86_64.
>
> This optimization supports the four modes of SM4, ECB, CBC, CFB, and
> CTR. Since CBC and CFB do not support multiple block parallel
> encryption, the optimization effect is not obvious. And all selftests
> have passed already.
>
> The main algorithm implementation comes from SM4 AES-NI work by
> libgcrypt and Markku-Juhani O. Saarinen at:
> https://github.com/mjosaarinen/sm4ni
>
> Benchmark on Intel Xeon Cascadelake, the data comes from the mode 218
> and mode 518 of tcrypt. The abscissas are blocks of different lengths.
> The data is tabulated and the unit is Mb/s:
>
> sm4-generic | 16 64 128 256 1024 1420 4096
> ECB enc | 40.99 46.50 48.05 48.41 49.20 49.25 49.28
> ECB dec | 41.07 46.99 48.15 48.67 49.20 49.25 49.29
> CBC enc | 37.71 45.28 46.77 47.60 48.32 48.37 48.40
> CBC dec | 36.48 44.82 46.43 47.45 48.23 48.30 48.36
> CFB enc | 37.94 44.84 46.12 46.94 47.57 47.46 47.68
> CFB dec | 37.50 42.84 43.74 44.37 44.85 44.80 44.96
> CTR enc | 39.20 45.63 46.75 47.49 48.09 47.85 48.08
> CTR dec | 39.64 45.70 46.72 47.47 47.98 47.88 48.06
> sm4-aesni-avx
> ECB enc | 33.75 134.47 221.64 243.43 264.05 251.58 258.13
> ECB dec | 34.02 134.92 223.11 245.14 264.12 251.04 258.33
> CBC enc | 38.85 46.18 47.67 48.34 49.00 48.96 49.14
> CBC dec | 33.54 131.29 223.88 245.27 265.50 252.41 263.78
> CFB enc | 38.70 46.10 47.58 48.29 49.01 48.94 49.19
> CFB dec | 32.79 128.40 223.23 244.87 265.77 253.31 262.79
> CTR enc | 32.58 122.23 220.29 241.16 259.57 248.32 256.69
> CTR dec | 32.81 122.47 218.99 241.54 258.42 248.58 256.61
>
> ---
> v3 changes:
> * Remove single block algorithm that does not greatly improve performance
> * Remove accelerated for sm4 key expand, which is not performance-critical
> * Fix the warning on arm64/sm4-ce
>
> v2 changes:
> * SM4 library functions use "sm4_" prefix instead of "crypto_" prefix
> * sm4-aesni-avx supports accelerated implementation of four specific modes
> * tcrypt benchmark supports sm4-aesni-avx
> * fixes of other reviews
>
>
> Tianjia Zhang (4):
> crypto: sm4 - create SM4 library based on sm4 generic code
> crypto: arm64/sm4-ce - Make dependent on sm4 library instead of
> sm4-generic
> crypto: x86/sm4 - add AES-NI/AVX/x86_64 implementation
> crypto: tcrypt - add the asynchronous speed test for SM4
>
> arch/arm64/crypto/Kconfig | 2 +-
> arch/arm64/crypto/sm4-ce-glue.c | 20 +-
> arch/x86/crypto/Makefile | 3 +
> arch/x86/crypto/sm4-aesni-avx-asm_64.S | 589 +++++++++++++++++++++++++
> arch/x86/crypto/sm4_aesni_avx_glue.c | 459 +++++++++++++++++++
> crypto/Kconfig | 22 +
> crypto/sm4_generic.c | 180 +-------
> crypto/tcrypt.c | 26 +-
> include/crypto/sm4.h | 25 +-
> lib/crypto/Kconfig | 3 +
> lib/crypto/Makefile | 3 +
> lib/crypto/sm4.c | 176 ++++++++
> 12 files changed, 1330 insertions(+), 178 deletions(-)
> create mode 100644 arch/x86/crypto/sm4-aesni-avx-asm_64.S
> create mode 100644 arch/x86/crypto/sm4_aesni_avx_glue.c
> create mode 100644 lib/crypto/sm4.c

All applied. Thanks.
--
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt