block-sha1: re-use the temporary array as we calculate the SHA1
The mozilla-SHA1 code did this 80-word array for the 80 iterations. But the SHA1 state is really just 512 bits, and you can actually keep it in a kind of "circular queue" of just 16 words instead. This requires us to do the xor updates as we go along (rather than as a pre-phase), but that's really what we want to do anyway. This gets me really close to the OpenSSL performance on my Nehalem. Look ma, all C code (ok, there's the rol/ror hack, but that one doesn't strictly even matter on my Nehalem, it's just a local optimization). Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Linus Torvalds committed
Aug 5, 2009 at 20:49 UTC
7b5075fcfb069fc36ba4cfe5567234974793ab58
1 file changed
+16
-12
block-sha1/sha1.c
+16
-12
@@ -96,9 +96,8 @@ void blk_SHA1_Final(unsigned char hashout[20], blk_SHA_CTX *ctx)
96
97
static void blk_SHA1Block(blk_SHA_CTX *ctx, const unsigned int *data)
98
{
99
- int t;
99
unsigned int A,B,C,D,E,TEMP;
101
- unsigned int W[80];
100
+ unsigned int array[16];
101
102
A = ctx->H[0];
103
B = ctx->H[1];
@@ -107,8 +106,8 @@ static void blk_SHA1Block(blk_SHA_CTX *ctx, const unsigned int *data)
106
E = ctx->H[4];
107
108
#define T_0_15(t) \
110
- TEMP = htonl(data[t]); W[t] = TEMP; \
111
- TEMP += SHA_ROL(A,5) + (((C^D)&B)^D) + E + 0x5a827999; \
109
+ TEMP = htonl(data[t]); array[t] = TEMP; \
110
+ TEMP += SHA_ROL(A,5) + (((C^D)&B)^D) + E + 0x5a827999; \
111
E = D; D = C; C = SHA_ROR(B, 2); B = A; A = TEMP; \
112
113
T_0_15( 0); T_0_15( 1); T_0_15( 2); T_0_15( 3); T_0_15( 4);
@@ -116,18 +115,21 @@ static void blk_SHA1Block(blk_SHA_CTX *ctx, const unsigned int *data)
115
T_0_15(10); T_0_15(11); T_0_15(12); T_0_15(13); T_0_15(14);
116
T_0_15(15);
117
119
- /* Unroll it? */
120
- for (t = 16; t <= 79; t++)
121
- W[t] = SHA_ROL(W[t-3] ^ W[t-8] ^ W[t-14] ^ W[t-16], 1);
118
+/* This "rolls" over the 512-bit array */
119
+#define W(x) (array[(x)&15])
120
+#define SHA_XOR(t) \
121
+ TEMP = SHA_ROL(W(t+13) ^ W(t+8) ^ W(t+2) ^ W(t), 1); W(t) = TEMP;
122
123
#define T_16_19(t) \
124
- TEMP = SHA_ROL(A,5) + (((C^D)&B)^D) + E + W[t] + 0x5a827999; \
125
- E = D; D = C; C = SHA_ROR(B, 2); B = A; A = TEMP;
124
+ SHA_XOR(t); \
125
+ TEMP += SHA_ROL(A,5) + (((C^D)&B)^D) + E + 0x5a827999; \
126
+ E = D; D = C; C = SHA_ROR(B, 2); B = A; A = TEMP; \
127
128
T_16_19(16); T_16_19(17); T_16_19(18); T_16_19(19);
129
130
#define T_20_39(t) \
130
- TEMP = SHA_ROL(A,5) + (B^C^D) + E + W[t] + 0x6ed9eba1; \
131
+ SHA_XOR(t); \
132
+ TEMP += SHA_ROL(A,5) + (B^C^D) + E + 0x6ed9eba1; \
133
E = D; D = C; C = SHA_ROR(B, 2); B = A; A = TEMP;
134
135
T_20_39(20); T_20_39(21); T_20_39(22); T_20_39(23); T_20_39(24);
@@ -136,7 +138,8 @@ static void blk_SHA1Block(blk_SHA_CTX *ctx, const unsigned int *data)
138
T_20_39(35); T_20_39(36); T_20_39(37); T_20_39(38); T_20_39(39);
139
140
#define T_40_59(t) \
139
- TEMP = SHA_ROL(A,5) + ((B&C)|(D&(B|C))) + E + W[t] + 0x8f1bbcdc; \
141
+ SHA_XOR(t); \
142
+ TEMP += SHA_ROL(A,5) + ((B&C)|(D&(B|C))) + E + 0x8f1bbcdc; \
143
E = D; D = C; C = SHA_ROR(B, 2); B = A; A = TEMP;
144
145
T_40_59(40); T_40_59(41); T_40_59(42); T_40_59(43); T_40_59(44);
@@ -145,7 +148,8 @@ static void blk_SHA1Block(blk_SHA_CTX *ctx, const unsigned int *data)
148
T_40_59(55); T_40_59(56); T_40_59(57); T_40_59(58); T_40_59(59);
149
150
#define T_60_79(t) \
148
- TEMP = SHA_ROL(A,5) + (B^C^D) + E + W[t] + 0xca62c1d6; \
151
+ SHA_XOR(t); \
152
+ TEMP += SHA_ROL(A,5) + (B^C^D) + E + 0xca62c1d6; \
153
E = D; D = C; C = SHA_ROR(B, 2); B = A; A = TEMP;
154
155
T_60_79(60); T_60_79(61); T_60_79(62); T_60_79(63); T_60_79(64);