crypto: arc4 - improve performance by adding ecb(arc4)

Currently arc4.c provides simple one-byte blocksize cipher which is wrapped by ecb() module, giving function call overhead on every encrypted byte. This patch adds ecb(arc4) directly into arc4.c for higher performance. tcrypt results (speed ratios: new/old): AMD Phenom II, x86-64 : x2.7 Intel Core 2, x86-64 : x1.9 Intel Atom N260, i386 : x1.4 Cc: Jon Oberheide <jon@oberheide.org> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

crypto: arc4 - improve performance by adding ecb(arc4)
Currently arc4.c provides simple one-byte blocksize cipher which is wrapped by ecb() module, giving function call overhead on every encrypted byte. This patch adds ecb(arc4) directly into arc4.c for higher performance. tcrypt results (speed ratios: new/old): AMD Phenom II, x86-64 : x2.7 Intel Core 2, x86-64 : x1.9 Intel Atom N260, i386 : x1.4 Cc: Jon Oberheide <jon@oberheide.org> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Jussi Kivilinna · Herbert Xu
1 parent 31b4cd2907
Showing 1 changed file with 87 additions and 22 deletions Side-by-side Diff
crypto/arc4.c
@@ -11,9 +11,11 @@
  * (at your option) any later version.
  *
  */
+
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/crypto.h>
+#include <crypto/algapi.h>
  
 #define ARC4_MIN_KEY_SIZE	1
 #define ARC4_MAX_KEY_SIZE	256
  
  
  
  
  
  
  
  
  
  
@@ -48,51 +50,114 @@
 	return 0;
 }
  
-static void arc4_crypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
+static void arc4_crypt(struct arc4_ctx *ctx, u8 *out, const u8 *in,
+		       unsigned int len)
 {
-	struct arc4_ctx *ctx = crypto_tfm_ctx(tfm);
-
 	u8 *const S = ctx->S;
-	u8 x = ctx->x;
-	u8 y = ctx->y;
-	u8 a, b;
+	u8 x, y, a, b;
+	u8 ty, ta, tb;
  
+	if (len == 0)
+		return;
+
+	x = ctx->x;
+	y = ctx->y;
+
 	a = S[x];
 	y = (y + a) & 0xff;
 	b = S[y];
-	S[x] = b;
-	S[y] = a;
-	x = (x + 1) & 0xff;
-	*out++ = *in ^ S[(a + b) & 0xff];
  
+	do {
+		S[y] = a;
+		a = (a + b) & 0xff;
+		S[x] = b;
+		x = (x + 1) & 0xff;
+		ta = S[x];
+		ty = (y + ta) & 0xff;
+		tb = S[ty];
+		*out++ = *in++ ^ S[a];
+		if (--len == 0)
+			break;
+		y = ty;
+		a = ta;
+		b = tb;
+	} while (true);
+
 	ctx->x = x;
 	ctx->y = y;
 }
  
-static struct crypto_alg arc4_alg = {
+static void arc4_crypt_one(struct crypto_tfm *tfm, u8 *out, const u8 *in)
+{
+	arc4_crypt(crypto_tfm_ctx(tfm), out, in, 1);
+}
+
+static int ecb_arc4_crypt(struct blkcipher_desc *desc, struct scatterlist *dst,
+			  struct scatterlist *src, unsigned int nbytes)
+{
+	struct arc4_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+
+	err = blkcipher_walk_virt(desc, &walk);
+
+	while (walk.nbytes > 0) {
+		u8 *wsrc = walk.src.virt.addr;
+		u8 *wdst = walk.dst.virt.addr;
+
+		arc4_crypt(ctx, wdst, wsrc, walk.nbytes);
+
+		err = blkcipher_walk_done(desc, &walk, 0);
+	}
+
+	return err;
+}
+
+static struct crypto_alg arc4_algs[2] = { {
 	.cra_name		=	"arc4",
 	.cra_flags		=	CRYPTO_ALG_TYPE_CIPHER,
 	.cra_blocksize		=	ARC4_BLOCK_SIZE,
 	.cra_ctxsize		=	sizeof(struct arc4_ctx),
 	.cra_module		=	THIS_MODULE,
-	.cra_list		=	LIST_HEAD_INIT(arc4_alg.cra_list),
-	.cra_u			=	{ .cipher = {
-	.cia_min_keysize	=	ARC4_MIN_KEY_SIZE,
-	.cia_max_keysize	=	ARC4_MAX_KEY_SIZE,
-	.cia_setkey		=	arc4_set_key,
-	.cia_encrypt		=	arc4_crypt,
-	.cia_decrypt		=	arc4_crypt } }
-};
+	.cra_u			=	{
+		.cipher = {
+			.cia_min_keysize	=	ARC4_MIN_KEY_SIZE,
+			.cia_max_keysize	=	ARC4_MAX_KEY_SIZE,
+			.cia_setkey		=	arc4_set_key,
+			.cia_encrypt		=	arc4_crypt_one,
+			.cia_decrypt		=	arc4_crypt_one,
+		},
+	},
+}, {
+	.cra_name		=	"ecb(arc4)",
+	.cra_priority		=	100,
+	.cra_flags		=	CRYPTO_ALG_TYPE_BLKCIPHER,
+	.cra_blocksize		=	ARC4_BLOCK_SIZE,
+	.cra_ctxsize		=	sizeof(struct arc4_ctx),
+	.cra_alignmask		=	0,
+	.cra_type		=	&crypto_blkcipher_type,
+	.cra_module		=	THIS_MODULE,
+	.cra_u			=	{
+		.blkcipher = {
+			.min_keysize	=	ARC4_MIN_KEY_SIZE,
+			.max_keysize	=	ARC4_MAX_KEY_SIZE,
+			.setkey		=	arc4_set_key,
+			.encrypt	=	ecb_arc4_crypt,
+			.decrypt	=	ecb_arc4_crypt,
+		},
+	},
+} };
  
 static int __init arc4_init(void)
 {
-	return crypto_register_alg(&arc4_alg);
+	return crypto_register_algs(arc4_algs, ARRAY_SIZE(arc4_algs));
 }
  
-
 static void __exit arc4_exit(void)
 {
-	crypto_unregister_alg(&arc4_alg);
+	crypto_unregister_algs(arc4_algs, ARRAY_SIZE(arc4_algs));
 }
  
 module_init(arc4_init);
...	...	@@ -11,9 +11,11 @@
11	11	* (at your option) any later version.
12	12	*
13	13	*/
	14	+
14	15	#include <linux/module.h>
15	16	#include <linux/init.h>
16	17	#include <linux/crypto.h>
	18	+#include <crypto/algapi.h>
17	19
18	20	#define ARC4_MIN_KEY_SIZE 1
19	21	#define ARC4_MAX_KEY_SIZE 256
20	22
21	23
22	24
23	25
24	26
25	27
26	28
27	29
28	30
29	31
...	...	@@ -48,51 +50,114 @@
48	50	return 0;
49	51	}
50	52
51		-static void arc4_crypt(struct crypto_tfm tfm, u8 out, const u8 *in)
	53	+static void arc4_crypt(struct arc4_ctx ctx, u8 out, const u8 *in,
	54	+ unsigned int len)
52	55	{
53		- struct arc4_ctx *ctx = crypto_tfm_ctx(tfm);
54		-
55	56	u8 *const S = ctx->S;
56		- u8 x = ctx->x;
57		- u8 y = ctx->y;
58		- u8 a, b;
	57	+ u8 x, y, a, b;
	58	+ u8 ty, ta, tb;
59	59
	60	+ if (len == 0)
	61	+ return;
	62	+
	63	+ x = ctx->x;
	64	+ y = ctx->y;
	65	+
60	66	a = S[x];
61	67	y = (y + a) & 0xff;
62	68	b = S[y];
63		- S[x] = b;
64		- S[y] = a;
65		- x = (x + 1) & 0xff;
66		- out++ = in ^ S[(a + b) & 0xff];
67	69
	70	+ do {
	71	+ S[y] = a;
	72	+ a = (a + b) & 0xff;
	73	+ S[x] = b;
	74	+ x = (x + 1) & 0xff;
	75	+ ta = S[x];
	76	+ ty = (y + ta) & 0xff;
	77	+ tb = S[ty];
	78	+ out++ = in++ ^ S[a];
	79	+ if (--len == 0)
	80	+ break;
	81	+ y = ty;
	82	+ a = ta;
	83	+ b = tb;
	84	+ } while (true);
	85	+
68	86	ctx->x = x;
69	87	ctx->y = y;
70	88	}
71	89
72		-static struct crypto_alg arc4_alg = {
	90	+static void arc4_crypt_one(struct crypto_tfm tfm, u8 out, const u8 *in)
	91	+{
	92	+ arc4_crypt(crypto_tfm_ctx(tfm), out, in, 1);
	93	+}
	94	+
	95	+static int ecb_arc4_crypt(struct blkcipher_desc desc, struct scatterlist dst,
	96	+ struct scatterlist *src, unsigned int nbytes)
	97	+{
	98	+ struct arc4_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
	99	+ struct blkcipher_walk walk;
	100	+ int err;
	101	+
	102	+ blkcipher_walk_init(&walk, dst, src, nbytes);
	103	+
	104	+ err = blkcipher_walk_virt(desc, &walk);
	105	+
	106	+ while (walk.nbytes > 0) {
	107	+ u8 *wsrc = walk.src.virt.addr;
	108	+ u8 *wdst = walk.dst.virt.addr;
	109	+
	110	+ arc4_crypt(ctx, wdst, wsrc, walk.nbytes);
	111	+
	112	+ err = blkcipher_walk_done(desc, &walk, 0);
	113	+ }
	114	+
	115	+ return err;
	116	+}
	117	+
	118	+static struct crypto_alg arc4_algs[2] = { {
73	119	.cra_name = "arc4",
74	120	.cra_flags = CRYPTO_ALG_TYPE_CIPHER,
75	121	.cra_blocksize = ARC4_BLOCK_SIZE,
76	122	.cra_ctxsize = sizeof(struct arc4_ctx),
77	123	.cra_module = THIS_MODULE,
78		- .cra_list = LIST_HEAD_INIT(arc4_alg.cra_list),
79		- .cra_u = { .cipher = {
80		- .cia_min_keysize = ARC4_MIN_KEY_SIZE,
81		- .cia_max_keysize = ARC4_MAX_KEY_SIZE,
82		- .cia_setkey = arc4_set_key,
83		- .cia_encrypt = arc4_crypt,
84		- .cia_decrypt = arc4_crypt } }
85		-};
	124	+ .cra_u = {
	125	+ .cipher = {
	126	+ .cia_min_keysize = ARC4_MIN_KEY_SIZE,
	127	+ .cia_max_keysize = ARC4_MAX_KEY_SIZE,
	128	+ .cia_setkey = arc4_set_key,
	129	+ .cia_encrypt = arc4_crypt_one,
	130	+ .cia_decrypt = arc4_crypt_one,
	131	+ },
	132	+ },
	133	+}, {
	134	+ .cra_name = "ecb(arc4)",
	135	+ .cra_priority = 100,
	136	+ .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER,
	137	+ .cra_blocksize = ARC4_BLOCK_SIZE,
	138	+ .cra_ctxsize = sizeof(struct arc4_ctx),
	139	+ .cra_alignmask = 0,
	140	+ .cra_type = &crypto_blkcipher_type,
	141	+ .cra_module = THIS_MODULE,
	142	+ .cra_u = {
	143	+ .blkcipher = {
	144	+ .min_keysize = ARC4_MIN_KEY_SIZE,
	145	+ .max_keysize = ARC4_MAX_KEY_SIZE,
	146	+ .setkey = arc4_set_key,
	147	+ .encrypt = ecb_arc4_crypt,
	148	+ .decrypt = ecb_arc4_crypt,
	149	+ },
	150	+ },
	151	+} };
86	152
87	153	static int __init arc4_init(void)
88	154	{
89		- return crypto_register_alg(&arc4_alg);
	155	+ return crypto_register_algs(arc4_algs, ARRAY_SIZE(arc4_algs));
90	156	}
91	157
92		-
93	158	static void __exit arc4_exit(void)
94	159	{
95		- crypto_unregister_alg(&arc4_alg);
	160	+ crypto_unregister_algs(arc4_algs, ARRAY_SIZE(arc4_algs));
96	161	}
97	162
98	163	module_init(arc4_init);