block-sha1: perform register rotation using cpp
Instead of letting the compiler to figure out the optimal way to rotate register usage, explicitly rotate the register names with cpp. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Linus Torvalds committed
Aug 6, 2009 at 12:41 UTC
30d12d4c16abc052e8961c07651f97bea2c061bd
1 file changed
+90
-27
block-sha1/sha1.c
+90
-27
@@ -101,20 +101,20 @@ void blk_SHA1_Final(unsigned char hashout[20], blk_SHA_CTX *ctx)
101
#define SHA_SRC(t) htonl(data[t])
102
#define SHA_MIX(t) SHA_ROL(W(t+13) ^ W(t+8) ^ W(t+2) ^ W(t), 1)
103
104
-#define SHA_ROUND(t, input, fn, constant) \
105
- TEMP = input(t); W(t) = TEMP; \
106
- TEMP += SHA_ROL(A,5) + (fn) + E + (constant); \
107
- E = D; D = C; C = SHA_ROR(B, 2); B = A; A = TEMP
104
+#define SHA_ROUND(t, input, fn, constant, A, B, C, D, E) do { \
105
+ unsigned int TEMP = input(t); W(t) = TEMP; \
106
+ TEMP += E + SHA_ROL(A,5) + (fn) + (constant); \
107
+ B = SHA_ROR(B, 2); E = TEMP; } while (0)
108
109
-#define T_0_15(t) SHA_ROUND(t, SHA_SRC, (((C^D)&B)^D) , 0x5a827999 )
110
-#define T_16_19(t) SHA_ROUND(t, SHA_MIX, (((C^D)&B)^D) , 0x5a827999 )
111
-#define T_20_39(t) SHA_ROUND(t, SHA_MIX, (B^C^D) , 0x6ed9eba1 )
112
-#define T_40_59(t) SHA_ROUND(t, SHA_MIX, ((B&C)+(D&(B^C))) , 0x8f1bbcdc )
113
-#define T_60_79(t) SHA_ROUND(t, SHA_MIX, (B^C^D) , 0xca62c1d6 )
109
+#define T_0_15(t, A, B, C, D, E) SHA_ROUND(t, SHA_SRC, (((C^D)&B)^D) , 0x5a827999, A, B, C, D, E )
110
+#define T_16_19(t, A, B, C, D, E) SHA_ROUND(t, SHA_MIX, (((C^D)&B)^D) , 0x5a827999, A, B, C, D, E )
111
+#define T_20_39(t, A, B, C, D, E) SHA_ROUND(t, SHA_MIX, (B^C^D) , 0x6ed9eba1, A, B, C, D, E )
112
+#define T_40_59(t, A, B, C, D, E) SHA_ROUND(t, SHA_MIX, ((B&C)+(D&(B^C))) , 0x8f1bbcdc, A, B, C, D, E )
113
+#define T_60_79(t, A, B, C, D, E) SHA_ROUND(t, SHA_MIX, (B^C^D) , 0xca62c1d6, A, B, C, D, E )
114
115
static void blk_SHA1Block(blk_SHA_CTX *ctx, const unsigned int *data)
116
{
117
- unsigned int A,B,C,D,E,TEMP;
117
+ unsigned int A,B,C,D,E;
118
unsigned int array[16];
119
120
A = ctx->H[0];
@@ -124,31 +124,94 @@ static void blk_SHA1Block(blk_SHA_CTX *ctx, const unsigned int *data)
124
E = ctx->H[4];
125
126
/* Round 1 - iterations 0-16 take their input from 'data' */
127
- T_0_15( 0); T_0_15( 1); T_0_15( 2); T_0_15( 3); T_0_15( 4);
128
- T_0_15( 5); T_0_15( 6); T_0_15( 7); T_0_15( 8); T_0_15( 9);
129
- T_0_15(10); T_0_15(11); T_0_15(12); T_0_15(13); T_0_15(14);
130
- T_0_15(15);
127
+ T_0_15( 0, A, B, C, D, E);
128
+ T_0_15( 1, E, A, B, C, D);
129
+ T_0_15( 2, D, E, A, B, C);
130
+ T_0_15( 3, C, D, E, A, B);
131
+ T_0_15( 4, B, C, D, E, A);
132
+ T_0_15( 5, A, B, C, D, E);
133
+ T_0_15( 6, E, A, B, C, D);
134
+ T_0_15( 7, D, E, A, B, C);
135
+ T_0_15( 8, C, D, E, A, B);
136
+ T_0_15( 9, B, C, D, E, A);
137
+ T_0_15(10, A, B, C, D, E);
138
+ T_0_15(11, E, A, B, C, D);
139
+ T_0_15(12, D, E, A, B, C);
140
+ T_0_15(13, C, D, E, A, B);
141
+ T_0_15(14, B, C, D, E, A);
142
+ T_0_15(15, A, B, C, D, E);
143
144
/* Round 1 - tail. Input from 512-bit mixing array */
133
- T_16_19(16); T_16_19(17); T_16_19(18); T_16_19(19);
145
+ T_16_19(16, E, A, B, C, D);
146
+ T_16_19(17, D, E, A, B, C);
147
+ T_16_19(18, C, D, E, A, B);
148
+ T_16_19(19, B, C, D, E, A);
149
150
/* Round 2 */
136
- T_20_39(20); T_20_39(21); T_20_39(22); T_20_39(23); T_20_39(24);
137
- T_20_39(25); T_20_39(26); T_20_39(27); T_20_39(28); T_20_39(29);
138
- T_20_39(30); T_20_39(31); T_20_39(32); T_20_39(33); T_20_39(34);
139
- T_20_39(35); T_20_39(36); T_20_39(37); T_20_39(38); T_20_39(39);
151
+ T_20_39(20, A, B, C, D, E);
152
+ T_20_39(21, E, A, B, C, D);
153
+ T_20_39(22, D, E, A, B, C);
154
+ T_20_39(23, C, D, E, A, B);
155
+ T_20_39(24, B, C, D, E, A);
156
+ T_20_39(25, A, B, C, D, E);
157
+ T_20_39(26, E, A, B, C, D);
158
+ T_20_39(27, D, E, A, B, C);
159
+ T_20_39(28, C, D, E, A, B);
160
+ T_20_39(29, B, C, D, E, A);
161
+ T_20_39(30, A, B, C, D, E);
162
+ T_20_39(31, E, A, B, C, D);
163
+ T_20_39(32, D, E, A, B, C);
164
+ T_20_39(33, C, D, E, A, B);
165
+ T_20_39(34, B, C, D, E, A);
166
+ T_20_39(35, A, B, C, D, E);
167
+ T_20_39(36, E, A, B, C, D);
168
+ T_20_39(37, D, E, A, B, C);
169
+ T_20_39(38, C, D, E, A, B);
170
+ T_20_39(39, B, C, D, E, A);
171
172
/* Round 3 */
142
- T_40_59(40); T_40_59(41); T_40_59(42); T_40_59(43); T_40_59(44);
143
- T_40_59(45); T_40_59(46); T_40_59(47); T_40_59(48); T_40_59(49);
144
- T_40_59(50); T_40_59(51); T_40_59(52); T_40_59(53); T_40_59(54);
145
- T_40_59(55); T_40_59(56); T_40_59(57); T_40_59(58); T_40_59(59);
173
+ T_40_59(40, A, B, C, D, E);
174
+ T_40_59(41, E, A, B, C, D);
175
+ T_40_59(42, D, E, A, B, C);
176
+ T_40_59(43, C, D, E, A, B);
177
+ T_40_59(44, B, C, D, E, A);
178
+ T_40_59(45, A, B, C, D, E);
179
+ T_40_59(46, E, A, B, C, D);
180
+ T_40_59(47, D, E, A, B, C);
181
+ T_40_59(48, C, D, E, A, B);
182
+ T_40_59(49, B, C, D, E, A);
183
+ T_40_59(50, A, B, C, D, E);
184
+ T_40_59(51, E, A, B, C, D);
185
+ T_40_59(52, D, E, A, B, C);
186
+ T_40_59(53, C, D, E, A, B);
187
+ T_40_59(54, B, C, D, E, A);
188
+ T_40_59(55, A, B, C, D, E);
189
+ T_40_59(56, E, A, B, C, D);
190
+ T_40_59(57, D, E, A, B, C);
191
+ T_40_59(58, C, D, E, A, B);
192
+ T_40_59(59, B, C, D, E, A);
193
194
/* Round 4 */
148
- T_60_79(60); T_60_79(61); T_60_79(62); T_60_79(63); T_60_79(64);
149
- T_60_79(65); T_60_79(66); T_60_79(67); T_60_79(68); T_60_79(69);
150
- T_60_79(70); T_60_79(71); T_60_79(72); T_60_79(73); T_60_79(74);
151
- T_60_79(75); T_60_79(76); T_60_79(77); T_60_79(78); T_60_79(79);
195
+ T_60_79(60, A, B, C, D, E);
196
+ T_60_79(61, E, A, B, C, D);
197
+ T_60_79(62, D, E, A, B, C);
198
+ T_60_79(63, C, D, E, A, B);
199
+ T_60_79(64, B, C, D, E, A);
200
+ T_60_79(65, A, B, C, D, E);
201
+ T_60_79(66, E, A, B, C, D);
202
+ T_60_79(67, D, E, A, B, C);
203
+ T_60_79(68, C, D, E, A, B);
204
+ T_60_79(69, B, C, D, E, A);
205
+ T_60_79(70, A, B, C, D, E);
206
+ T_60_79(71, E, A, B, C, D);
207
+ T_60_79(72, D, E, A, B, C);
208
+ T_60_79(73, C, D, E, A, B);
209
+ T_60_79(74, B, C, D, E, A);
210
+ T_60_79(75, A, B, C, D, E);
211
+ T_60_79(76, E, A, B, C, D);
212
+ T_60_79(77, D, E, A, B, C);
213
+ T_60_79(78, C, D, E, A, B);
214
+ T_60_79(79, B, C, D, E, A);
215
216
ctx->H[0] += A;
217
ctx->H[1] += B;