File: fastmint_mmx_standard_1.c

package info (click to toggle)
hashcash 1.21-2
  • links: PTS
  • area: main
  • in suites: bullseye, buster, stretch
  • size: 868 kB
  • ctags: 870
  • sloc: ansic: 8,476; perl: 925; sh: 298; makefile: 226; python: 41
file content (638 lines) | stat: -rw-r--r-- 21,311 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
#include <stdio.h>
#include <setjmp.h>
#include "libfastmint.h"

#if (defined(__i386__) || defined(__AMD64__)) && defined(__GNUC__) && defined(__MMX__)
typedef int mmx_d_t __attribute__ ((vector_size (8)));
typedef int mmx_q_t __attribute__ ((vector_size (8)));
#endif

int minter_mmx_standard_1_test(void)
{
  /* This minter runs only on x86 and AMD64 hardware supporting MMX - and will only compile on GCC */
#if (defined(__i386__) || defined(__AMD64__)) && defined(__GNUC__) && defined(__MMX__)
	return (gProcessorSupportFlags & HC_CPU_SUPPORTS_MMX) != 0;
#endif
  
  /* Not an x86 or AMD64, or compiler doesn't support MMX or GNU assembly */
  return 0;
}

/* Define low-level primitives in terms of operations */
/* #define S(n, X) ( ( (X) << (n) ) | ( (X) >> ( 32 - (n) ) ) ) */
#define XOR(a,b) ( (mmx_d_t) __builtin_ia32_pxor( (mmx_q_t) a, (mmx_q_t) b) )
#define AND(a,b) ( (mmx_d_t) __builtin_ia32_pand( (mmx_q_t) a, (mmx_q_t) b) )
#define ANDNOT(a,b) ( (mmx_d_t) __builtin_ia32_pandn( (mmx_q_t) b, (mmx_q_t) a) )
#define OR(a,b) ( (mmx_d_t) __builtin_ia32_por( (mmx_q_t) a, (mmx_q_t) b) )
#define ADD(a,b) ( __builtin_ia32_paddd(a,b) )

#if (defined(__i386__) || defined(__AMD64__)) && defined(__GNUC__) && defined(__MMX__)
static inline mmx_d_t S(int n, mmx_d_t X)
{
  mmx_d_t G = {} ;

  asm ("movq %[x],%[g]\n\t"
       "pslld %[sl],%[x]\n\t"
       "psrld %[sr],%[g]\n\t"
       "por %[g],%[x]"
       : [g] "=y" (G), [x] "=y" (X)
       : "[x]" (X), [sl] "g" (n), [sr] "g" (32-n)
       );

  return X;
}
#endif

/* #define F1( B, C, D ) ( ( (B) & (C) ) | ( ~(B) & (D) ) ) */
/* #define F1( B, C, D ) ( (D) ^ ( (B) & ( (C) ^ (D) ) ) ) */
#define F1( B, C, D ) ( \
	F = AND(B,C), \
	G = ANDNOT(D,B), \
	OR(F,G) )
/* #define F2( B, C, D ) ( (B) ^ (C) ^ (D) ) */
#define F2( B, C, D ) ( \
	F = XOR(B,C), \
	XOR(F,D) )
/* #define F3( B, C, D ) ( (B) & (C) ) | ( (C) & (D) ) | ( (B) & (D) ) */
/* #define F3( B, C, D ) ( ( (B) & ( (C) | (D) )) | ( (C) & (D) ) ) */
#define F3( B, C, D ) ( \
	F = OR(C,D), \
	G = AND(C,D), \
	F = AND(B,F), \
	OR(F,G) )
/* #define F4( B, C, D ) ( (B) ^ (C) ^ (D) ) */
#define F4(B,C,D) F2(B,C,D)

#define K1 0x5A827999  /* constant used for rounds 0..19 */
#define K2 0x6ED9EBA1  /* constant used for rounds 20..39 */
#define K3 0x8F1BBCDC  /* constant used for rounds 40..59 */
#define K4 0xCA62C1D6  /* constant used for rounds 60..79 */

/* #define Wf(t) (W[t] = S(1, W[t-16] ^ W[t-14] ^ W[t-8] ^ W[t-3])) */
#define Wf(W,t) ( \
	F = XOR((W)[t-16], (W)[t-14]), \
	G = XOR((W)[t-8], (W)[t-3]), \
	F = XOR(F,G), \
	(W)[t] = S(1,F) )

#define Wfly(W,t) ( (t) < 16 ? (W)[t] : Wf(W,t) )

/*
	#define ROUND(t,A,B,C,D,E,Func,K,W) \
	E = ADD(E,K); \
	F = S(5,A); \
	E = ADD(F,E); \
	F = Wfly(W,t); \
	E = ADD(F,E); \
	F = Func(B,C,D); \
	E = ADD(F,E); \
	B = S(30,B);
*/

#define ROUND_F1_n(t,A,B,C,D,E,K,W) \
	asm ( \
		"\n\t movq  %[d], %%mm5" /* begin F1(B,C,D) */ \
		"\n\t pxor  %[c], %%mm5" \
		"\n\t movq  %[a], %%mm7" /* begin S(5,A) */ \
		"\n\t pand  %[b], %%mm5" \
		"\n\t pslld $5,   %%mm7" \
		"\n\t pxor  %[d], %%mm5" \
		"\n\t movq  %[a], %%mm6" \
		"\n\t paddd %%mm5,%[e]"  /* sum F1(B,C,D) to E */ \
		"\n\t psrld $27,  %%mm6" \
		"\n\t paddd %[k], %[e]"  /* sum K to E */ \
		"\n\t por   %%mm6,%%mm7" \
		"\n\t movq  %[b], %%mm5" /* begin S(30,B) */ \
		"\n\t paddd %%mm7,%[e]"  /* sum S(5,A) to E */ \
		"\n\t pslld $30,  %[b]" \
		"\n\t psrld $2,   %%mm5" \
		"\n\t paddd %[Wt],%[e]"  /* sum W[t] to E */ \
		"\n\t por   %%mm5,%[b]"  /* complete S(30,B) */ \
		: [a] "+y" (A), [b] "+y" (B), [c] "+y" (C), [d] "+y" (D), [e] "+y" (E) \
		: [Wt] "m" ((W)[t]), [k] "m" (K) \
		: "mm5", "mm6", "mm7" );

#define ROUND_F1_u(t,A,B,C,D,E,K,W) \
	asm ( \
		"\n\t movq  %[d], %%mm5" /* begin F1(B,C,D) */ \
		"\n\t pxor  %[c], %%mm5" \
		"\n\t movq  %[a], %%mm7" /* begin S(5,A) */ \
		"\n\t pand  %[b], %%mm5" \
		"\n\t pslld $5,   %%mm7" \
		"\n\t pxor  %[d], %%mm5" \
		"\n\t movq  %[a], %%mm6" \
		"\n\t paddd %%mm5,%[e]"  /* sum F1(B,C,D) to E */ \
		"\n\t psrld $27,  %%mm6" \
		"\n\t paddd %[k], %[e]"  /* sum K to E */ \
		"\n\t por   %%mm6,%%mm7" \
		"\n\t movq  %[b], %%mm5" /* begin S(30,B) */ \
		"\n\t paddd %%mm7,%[e]"  /* sum S(5,A) to E */ \
		"\n\t movq  %[Wt_3],%%mm7" /* begin Wf(t) */ \
		"\n\t movq  %[Wt_8],%%mm6" \
		"\n\t pxor  %[Wt_14],%%mm7" \
		"\n\t pxor  %[Wt_16],%%mm6" \
		"\n\t pslld $30,  %[b]" \
		"\n\t pxor  %%mm6,%%mm7" \
		"\n\t movq  %%mm7,%%mm6" \
		"\n\t pslld $1,   %%mm7" \
		"\n\t psrld $31,  %%mm6" \
		"\n\t psrld $2,   %%mm5" \
		"\n\t por   %%mm6,%%mm7" \
		"\n\t por   %%mm5,%[b]"  /* complete S(30,B) */ \
		"\n\t paddd %%mm7,%[e]"  /* sum Wf(t) to E */ \
		"\n\t movq  %%mm7,%[Wt]" /* write back Wf(t) to W[t] */ \
		: [a] "+y" (A), [b] "+y" (B), [c] "+y" (C), [d] "+y" (D), [e] "+y" (E), [Wt] "=m" ((W)[t]) \
		: [Wt_3] "m" ((W)[t-3]), [Wt_14] "m" ((W)[t-14]), [Wt_8] "m" ((W)[t-8]), [Wt_16] "m" ((W)[t-16]), [k] "m" (K) \
		: "mm5", "mm6", "mm7" );

#define ROUND_F2_n(t,A,B,C,D,E,K,W) \
	asm ( \
		"\n\t movq  %[b], %%mm5" /* begin F2(B,C,D) */ \
		"\n\t movq  %[a], %%mm7" /* begin S(5,A) */ \
		"\n\t pxor  %[c], %%mm5" \
		"\n\t pslld $5,   %%mm7" \
		"\n\t pxor  %[d], %%mm5" \
		"\n\t movq  %[a], %%mm6" \
		"\n\t paddd %%mm5,%[e]"  /* sum F2(B,C,D) to E */ \
		"\n\t psrld $27,  %%mm6" \
		"\n\t paddd %[k], %[e]"  /* sum K to E */ \
		"\n\t por   %%mm6,%%mm7" \
		"\n\t movq  %[b], %%mm5" /* begin S(30,B) */ \
		"\n\t paddd %%mm7,%[e]"  /* sum S(5,A) to E */ \
		"\n\t pslld $30,  %[b]" \
		"\n\t psrld $2,   %%mm5" \
		"\n\t paddd %[Wt],%[e]"  /* sum W[t] to E */ \
		"\n\t por   %%mm5,%[b]"  /* complete S(30,B) */ \
		: [a] "+y" (A), [b] "+y" (B), [c] "+y" (C), [d] "+y" (D), [e] "+y" (E) \
		: [Wt] "m" ((W)[t]), [k] "m" (K) \
		: "mm5", "mm6", "mm7" );

#define ROUND_F2_u(t,A,B,C,D,E,K,W) \
	asm ( \
		"\n\t movq  %[b], %%mm5" /* begin F2(B,C,D) */ \
		"\n\t movq  %[a], %%mm7" /* begin S(5,A) */ \
		"\n\t pxor  %[c], %%mm5" \
		"\n\t pslld $5,   %%mm7" \
		"\n\t pxor  %[d], %%mm5" \
		"\n\t movq  %[a], %%mm6" \
		"\n\t paddd %%mm5,%[e]"  /* sum F2(B,C,D) to E */ \
		"\n\t psrld $27,  %%mm6" \
		"\n\t paddd %[k], %[e]"  /* sum K to E */ \
		"\n\t por   %%mm6,%%mm7" \
		"\n\t movq  %[b], %%mm5" /* begin S(30,B) */ \
		"\n\t paddd %%mm7,%[e]"  /* sum S(5,A) to E */ \
		"\n\t movq  %[Wt_3],%%mm7" /* begin Wf(t) */ \
		"\n\t movq  %[Wt_8],%%mm6" \
		"\n\t pxor  %[Wt_14],%%mm7" \
		"\n\t pxor  %[Wt_16],%%mm6" \
		"\n\t pslld $30,  %[b]" \
		"\n\t pxor  %%mm6,%%mm7" \
		"\n\t movq  %%mm7,%%mm6" \
		"\n\t pslld $1,   %%mm7" \
		"\n\t psrld $31,  %%mm6" \
		"\n\t psrld $2,   %%mm5" \
		"\n\t por   %%mm6,%%mm7" \
		"\n\t por   %%mm5,%[b]"  /* complete S(30,B) */ \
		"\n\t paddd %%mm7,%[e]"  /* sum Wf(t) to E */ \
		"\n\t movq  %%mm7,%[Wt]" /* write back Wf(t) to W[t] */ \
		: [a] "+y" (A), [b] "+y" (B), [c] "+y" (C), [d] "+y" (D), [e] "+y" (E), [Wt] "=m" ((W)[t]) \
		: [Wt_3] "m" ((W)[t-3]), [Wt_14] "m" ((W)[t-14]), [Wt_8] "m" ((W)[t-8]), [Wt_16] "m" ((W)[t-16]), [k] "m" (K) \
		: "mm5", "mm6", "mm7" );

#define ROUND_F3_n(t,A,B,C,D,E,K,W) \
	asm ( \
		"\n\t movq  %[d], %%mm5" /* begin F3(B,C,D) */ \
		"\n\t movq  %[d], %%mm6" \
		"\n\t por   %[c], %%mm5" \
		"\n\t pand  %[c], %%mm6" \
		"\n\t movq  %[a], %%mm7" /* begin S(5,A) */ \
		"\n\t pand  %[b], %%mm5" \
		"\n\t pslld $5,   %%mm7" \
		"\n\t por   %%mm6,%%mm5" \
		"\n\t movq  %[a], %%mm6" \
		"\n\t paddd %%mm5,%[e]"  /* sum F3(B,C,D) to E */ \
		"\n\t psrld $27,  %%mm6" \
		"\n\t paddd %[k], %[e]"  /* sum K to E */ \
		"\n\t por   %%mm6,%%mm7" \
		"\n\t movq  %[b], %%mm5" /* begin S(30,B) */ \
		"\n\t paddd %%mm7,%[e]"  /* sum S(5,A) to E */ \
		"\n\t pslld $30,  %[b]" \
		"\n\t psrld $2,   %%mm5" \
		"\n\t paddd %[Wt],%[e]"  /* sum W[t] to E */ \
		"\n\t por   %%mm5,%[b]"  /* complete S(30,B) */ \
		: [a] "+y" (A), [b] "+y" (B), [c] "+y" (C), [d] "+y" (D), [e] "+y" (E) \
		: [Wt] "m" ((W)[t]), [k] "m" (K) \
		: "mm5", "mm6", "mm7" );

#define ROUND_F3_u(t,A,B,C,D,E,K,W) \
	asm ( \
		"\n\t movq  %[d], %%mm5" /* begin F3(B,C,D) */ \
		"\n\t movq  %[d], %%mm6" \
		"\n\t por   %[c], %%mm5" \
		"\n\t pand  %[c], %%mm6" \
		"\n\t movq  %[a], %%mm7" /* begin S(5,A) */ \
		"\n\t pand  %[b], %%mm5" \
		"\n\t pslld $5,   %%mm7" \
		"\n\t por   %%mm6,%%mm5" \
		"\n\t movq  %[a], %%mm6" \
		"\n\t paddd %%mm5,%[e]"  /* sum F3(B,C,D) to E */ \
		"\n\t psrld $27,  %%mm6" \
		"\n\t paddd %[k], %[e]"  /* sum K to E */ \
		"\n\t por   %%mm6,%%mm7" \
		"\n\t movq  %[b], %%mm5" /* begin S(30,B) */ \
		"\n\t paddd %%mm7,%[e]"  /* sum S(5,A) to E */ \
		"\n\t movq  %[Wt_3],%%mm7" /* begin Wf(t) */ \
		"\n\t movq  %[Wt_8],%%mm6" \
		"\n\t pxor  %[Wt_14],%%mm7" \
		"\n\t pxor  %[Wt_16],%%mm6" \
		"\n\t pslld $30,  %[b]" \
		"\n\t pxor  %%mm6,%%mm7" \
		"\n\t movq  %%mm7,%%mm6" \
		"\n\t pslld $1,   %%mm7" \
		"\n\t psrld $31,  %%mm6" \
		"\n\t psrld $2,   %%mm5" \
		"\n\t por   %%mm6,%%mm7" \
		"\n\t por   %%mm5,%[b]"  /* complete S(30,B) */ \
		"\n\t paddd %%mm7,%[e]"  /* sum Wf(t) to E */ \
		"\n\t movq  %%mm7,%[Wt]" /* write back Wf(t) to W[t] */ \
		: [a] "+y" (A), [b] "+y" (B), [c] "+y" (C), [d] "+y" (D), [e] "+y" (E), [Wt] "=m" ((W)[t]) \
		: [Wt_3] "m" ((W)[t-3]), [Wt_14] "m" ((W)[t-14]), [Wt_8] "m" ((W)[t-8]), [Wt_16] "m" ((W)[t-16]), [k] "m" (K) \
		: "mm5", "mm6", "mm7" );

#define ROUND_F4_u ROUND_F2_u
#define ROUND_F4_n ROUND_F2_n

#define ROUNDu(t,A,B,C,D,E,Func,K) \
		if((t) < 16) { \
			ROUND_##Func##_n(t,A,B,C,D,E,K,W); \
		} else { \
			ROUND_##Func##_u(t,A,B,C,D,E,K,W); \
		}

#define ROUNDn(t,A,B,C,D,E,Func,K) \
		ROUND_##Func##_n(t,A,B,C,D,E,K,W); \

#define ROUND5( t, Func, K ) \
    ROUNDu( t + 0, A, B, C, D, E, Func, K );\
    ROUNDu( t + 1, E, A, B, C, D, Func, K );\
    ROUNDu( t + 2, D, E, A, B, C, Func, K );\
    ROUNDu( t + 3, C, D, E, A, B, Func, K );\
    ROUNDu( t + 4, B, C, D, E, A, Func, K );

#if defined(MINTER_CALLBACK_CLEANUP_FP)
#undef MINTER_CALLBACK_CLEANUP_FP
#endif
#define MINTER_CALLBACK_CLEANUP_FP __builtin_ia32_emms()

unsigned long minter_mmx_standard_1(int bits, int* best, unsigned char *block, const uInt32 IV[5], int tailIndex, unsigned long maxIter, MINTER_CALLBACK_ARGS)
{
#if (defined(__i386__) || defined(__AMD64__)) && defined(__GNUC__) && defined(__MMX__)
  MINTER_CALLBACK_VARS;
  unsigned long iters = 0 ;
  int n = 0, t = 0, gotBits = 0, maxBits = (bits > 16) ? 16 : bits;
  uInt32 bitMask1Low = 0 , bitMask1High = 0 , s = 0 ;
  mmx_d_t vBitMaskHigh = {} , vBitMaskLow = {} ;
  register mmx_d_t A = {} , B = {} , C = {} , D = {} , E = {} ;
  mmx_d_t MA = {} , MB = {} ;
  mmx_d_t W[80] = {} ;
  mmx_d_t H[5] = {} , pH[5] = {} ;
  mmx_d_t K[4] = {} ;
  uInt32 *Hw = (uInt32*) H;
  uInt32 *pHw = (uInt32*) pH;
  uInt32 IA = 0 , IB = 0 ;
  const char *p = encodeAlphabets[EncodeBase64];
  unsigned char *X = (unsigned char*) W;
  unsigned char *output = (unsigned char*) block;
	
  *best = 0;

  /* Splat Kn constants into MMX-style array */
  ((uInt32*)K)[0] = ((uInt32*)K)[1] = K1;
  ((uInt32*)K)[2] = ((uInt32*)K)[3] = K2;
  ((uInt32*)K)[4] = ((uInt32*)K)[5] = K3;
  ((uInt32*)K)[6] = ((uInt32*)K)[7] = K4;
	
  /* Work out which bits to mask out for test */
  if(maxBits < 32) {
    if ( bits == 0 ) { bitMask1Low = 0; } else {
      bitMask1Low = ~((((uInt32) 1) << (32 - maxBits)) - 1);
    }
    bitMask1High = 0;
  } else {
    bitMask1Low = ~0;
    bitMask1High = ~((((uInt32) 1) << (64 - maxBits)) - 1);
  }
  ((uInt32*) &vBitMaskLow )[0] = bitMask1Low ;
  ((uInt32*) &vBitMaskLow )[1] = bitMask1Low ;
  ((uInt32*) &vBitMaskHigh)[0] = bitMask1High;
  ((uInt32*) &vBitMaskHigh)[1] = bitMask1High;
  maxBits = 0;
	
  /* Copy block and IV to vectorised internal storage */
  /* Assume little-endian order, as we're on x86 or AMD64 */
  for(t=0; t < 16; t++) {
    X[t*8+ 0] = X[t*8+ 4] = output[t*4+3];
    X[t*8+ 1] = X[t*8+ 5] = output[t*4+2];
    X[t*8+ 2] = X[t*8+ 6] = output[t*4+1];
    X[t*8+ 3] = X[t*8+ 7] = output[t*4+0];
  }
  for(t=0; t < 5; t++) {
    Hw[t*2+0] = Hw[t*2+1] =
    pHw[t*2+0] = pHw[t*2+1] = IV[t];
  }
	
  /* The Tight Loop - everything in here should be extra efficient */
  for(iters=0; iters < maxIter-2; iters += 2) {

    /* Encode iteration count into tail */
    /* Iteration count is always 2-aligned, so only least-significant character needs multiple lookup */
    /* Further, we assume we're always little-endian */
    X[(((tailIndex - 1) & ~3) << 1) + (((tailIndex - 1) & 3) ^ 3) +  0] = p[(iters & 0x3e) + 0];
    X[(((tailIndex - 1) & ~3) << 1) + (((tailIndex - 1) & 3) ^ 3) +  4] = p[(iters & 0x3e) + 1];
    if(!(iters & 0x3f)) {
      if ( iters >> 6 ) {
	X[(((tailIndex - 2) & ~3) << 1) + (((tailIndex - 2) & 3) ^ 3) +  0] =
	X[(((tailIndex - 2) & ~3) << 1) + (((tailIndex - 2) & 3) ^ 3) +  4] = p[(iters >>  6) & 0x3f];
      }
      if ( iters >> 12 ) { 
	X[(((tailIndex - 3) & ~3) << 1) + (((tailIndex - 3) & 3) ^ 3) +  0] =
	X[(((tailIndex - 3) & ~3) << 1) + (((tailIndex - 3) & 3) ^ 3) +  4] = p[(iters >> 12) & 0x3f];
      }
      if ( iters >> 18 ) { 
	X[(((tailIndex - 4) & ~3) << 1) + (((tailIndex - 4) & 3) ^ 3) +  0] =
	X[(((tailIndex - 4) & ~3) << 1) + (((tailIndex - 4) & 3) ^ 3) +  4] = p[(iters >> 18) & 0x3f];
      }
      if ( iters >> 24 ) {
	X[(((tailIndex - 5) & ~3) << 1) + (((tailIndex - 5) & 3) ^ 3) +  0] =
	X[(((tailIndex - 5) & ~3) << 1) + (((tailIndex - 5) & 3) ^ 3) +  4] = p[(iters >> 24) & 0x3f];
      }
      if ( iters >> 30 ) { 
	X[(((tailIndex - 6) & ~3) << 1) + (((tailIndex - 6) & 3) ^ 3) +  0] =
	X[(((tailIndex - 6) & ~3) << 1) + (((tailIndex - 6) & 3) ^ 3) +  4] = p[(iters >> 30) & 0x3f];
      }
    }
		
    /* Force compiler to flush and reload MMX registers */
    asm volatile ( "nop" : : : "mm0", "mm1", "mm2", "mm3", "mm4", "mm5", "mm6", "mm7", "memory" );

    /* Bypass shortcuts below on certain iterations */
    if((!(iters & 0xffffff)) && (tailIndex == 52 || tailIndex == 32)) {
			/* Populate W buffer */
			for(t=16; t < 32; t += 4) {
				asm volatile (
											/* Use pairs of adjacent MMX registers to build four nearly-independent chains */
	"\n\t movq -128(%[w]),%%mm0"
	"\n\t movq -120(%[w]),%%mm2"
	"\n\t movq -112(%[w]),%%mm4"
	"\n\t movq -104(%[w]),%%mm6"
	"\n\t pxor  %%mm4,    %%mm0"
	"\n\t pxor  %%mm6,    %%mm2"
	"\n\t pxor  -96(%[w]),%%mm4"
	"\n\t pxor  -88(%[w]),%%mm6"
	"\n\t pxor  -64(%[w]),%%mm0"
	"\n\t pxor  -56(%[w]),%%mm2"
	"\n\t pxor  -48(%[w]),%%mm4"
	"\n\t pxor  -40(%[w]),%%mm6"
	"\n\t pxor  -24(%[w]),%%mm0"
	"\n\t pxor  -16(%[w]),%%mm2"

	/* 0(%[w]) is not yet valid! */

	"\n\t movq  %%mm0, %%mm1"
	"\n\t movq  %%mm2, %%mm3"
	"\n\t pslld $1,    %%mm0"
	"\n\t psrld $31,   %%mm1"
	"\n\t pslld $1,    %%mm2"
	"\n\t psrld $31,   %%mm3"
	"\n\t por   %%mm1, %%mm0"
	"\n\t por   %%mm3, %%mm2"

	/* ...now it is */
	"\n\t pxor   -8(%[w]),%%mm4"
	"\n\t pxor  %%mm0, %%mm6"
	"\n\t movq  %%mm4, %%mm5"
	"\n\t movq  %%mm6, %%mm7"
	"\n\t pslld $1,    %%mm4"
	"\n\t psrld $31,   %%mm5"
	"\n\t pslld $1,    %%mm6"
	"\n\t psrld $31,   %%mm7"
	"\n\t por   %%mm5, %%mm4"
	"\n\t por   %%mm7, %%mm6"
      	
	"\n\t movq  %%mm0, 0(%[w])"
	"\n\t movq  %%mm2, 8(%[w])"
	"\n\t movq  %%mm4,16(%[w])"
	"\n\t movq  %%mm6,24(%[w])"
      	
	: /* no outputs */
	: [w] "r" (W+t)
	: "mm0", "mm1", "mm2", "mm3", "mm4", "mm5", "mm6", "mm7", "memory"
	);
			}

      A = H[0];
      B = H[1];
      C = H[2];
      D = H[3];
      E = H[4];
			
      ROUNDn( 0, A, B, C, D, E, F1, K[0] );
      ROUNDn( 1, E, A, B, C, D, F1, K[0] );
      ROUNDn( 2, D, E, A, B, C, F1, K[0] );
      ROUNDn( 3, C, D, E, A, B, F1, K[0] );
      ROUNDn( 4, B, C, D, E, A, F1, K[0] );
      ROUNDn( 5, A, B, C, D, E, F1, K[0] );
      ROUNDn( 6, E, A, B, C, D, F1, K[0] );
			
      if(tailIndex == 52) {
				ROUNDn( 7, D, E, A, B, C, F1, K[0] );
				ROUNDn( 8, C, D, E, A, B, F1, K[0] );
				ROUNDn( 9, B, C, D, E, A, F1, K[0] );
				ROUNDn(10, A, B, C, D, E, F1, K[0] );
				ROUNDn(11, E, A, B, C, D, F1, K[0] );
      }
			
      pH[0] = A;
      pH[1] = B;
      pH[2] = C;
      pH[3] = D;
      pH[4] = E;
    }
		
    /* Set up working variables */
    A = pH[0];
    B = pH[1];
    C = pH[2];
    D = pH[3];
    E = pH[4];
		
    /* Do the rounds */
    switch(tailIndex) {
    default:
      ROUNDn( 0, A, B, C, D, E, F1, K[0] );
      ROUNDn( 1, E, A, B, C, D, F1, K[0] );
      ROUNDn( 2, D, E, A, B, C, F1, K[0] );
      ROUNDn( 3, C, D, E, A, B, F1, K[0] );
      ROUNDn( 4, B, C, D, E, A, F1, K[0] );
      ROUNDn( 5, A, B, C, D, E, F1, K[0] );
      ROUNDn( 6, E, A, B, C, D, F1, K[0] );
    case 32:
      ROUNDn( 7, D, E, A, B, C, F1, K[0] );
      ROUNDn( 8, C, D, E, A, B, F1, K[0] );
      ROUNDn( 9, B, C, D, E, A, F1, K[0] );
      ROUNDn(10, A, B, C, D, E, F1, K[0] );
      ROUNDn(11, E, A, B, C, D, F1, K[0] );
    case 52:
      ROUNDn(12, D, E, A, B, C, F1, K[0] );
      ROUNDn(13, C, D, E, A, B, F1, K[0] );
      ROUNDn(14, B, C, D, E, A, F1, K[0] );
      ROUNDn(15, A, B, C, D, E, F1, K[0] );
    }

    if(tailIndex == 52) {
      ROUNDn(16, E, A, B, C, D, F1, K[0] );
      ROUNDn(17, D, E, A, B, C, F1, K[0] );
      ROUNDn(18, C, D, E, A, B, F1, K[0] );
      ROUNDn(19, B, C, D, E, A, F1, K[0] );
      ROUNDu(20, A, B, C, D, E, F2, K[1] );
      ROUNDn(21, E, A, B, C, D, F2, K[1] );
      ROUNDn(22, D, E, A, B, C, F2, K[1] );
      ROUNDu(23, C, D, E, A, B, F2, K[1] );
      ROUNDn(24, B, C, D, E, A, F2, K[1] );
      ROUNDn(25, A, B, C, D, E, F2, K[1] );
      ROUNDu(26, E, A, B, C, D, F2, K[1] );
      ROUNDn(27, D, E, A, B, C, F2, K[1] );
      ROUNDu(28, C, D, E, A, B, F2, K[1] );
      ROUNDu(29, B, C, D, E, A, F2, K[1] );
      ROUNDn(30, A, B, C, D, E, F2, K[1] );
    } else if (tailIndex == 32) {
      ROUNDn(16, E, A, B, C, D, F1, K[0] );
      ROUNDn(17, D, E, A, B, C, F1, K[0] );
      ROUNDn(18, C, D, E, A, B, F1, K[0] );
      ROUNDn(19, B, C, D, E, A, F1, K[0] );
      ROUNDn(20, A, B, C, D, E, F2, K[1] );
      ROUNDu(21, E, A, B, C, D, F2, K[1] );
      ROUNDn(22, D, E, A, B, C, F2, K[1] );
      ROUNDu(23, C, D, E, A, B, F2, K[1] );
      ROUNDu(24, B, C, D, E, A, F2, K[1] );
      ROUNDn(25, A, B, C, D, E, F2, K[1] );
      ROUNDu(26, E, A, B, C, D, F2, K[1] );
      ROUNDu(27, D, E, A, B, C, F2, K[1] );
      ROUNDn(28, C, D, E, A, B, F2, K[1] );
      ROUNDu(29, B, C, D, E, A, F2, K[1] );
      ROUNDu(30, A, B, C, D, E, F2, K[1] );
    } else {
      ROUNDu(16, E, A, B, C, D, F1, K[0] );
      ROUNDu(17, D, E, A, B, C, F1, K[0] );
      ROUNDu(18, C, D, E, A, B, F1, K[0] );
      ROUNDu(19, B, C, D, E, A, F1, K[0] );
      ROUNDu(20, A, B, C, D, E, F2, K[1] );
      ROUNDu(21, E, A, B, C, D, F2, K[1] );
      ROUNDu(22, D, E, A, B, C, F2, K[1] );
      ROUNDu(23, C, D, E, A, B, F2, K[1] );
      ROUNDu(24, B, C, D, E, A, F2, K[1] );
      ROUNDu(25, A, B, C, D, E, F2, K[1] );
      ROUNDu(26, E, A, B, C, D, F2, K[1] );
      ROUNDu(27, D, E, A, B, C, F2, K[1] );
      ROUNDu(28, C, D, E, A, B, F2, K[1] );
      ROUNDu(29, B, C, D, E, A, F2, K[1] );
      ROUNDu(30, A, B, C, D, E, F2, K[1] );
    }
	  
    ROUNDu(31, E, A, B, C, D, F2, K[1] );
    ROUNDu(32, D, E, A, B, C, F2, K[1] );
    ROUNDu(33, C, D, E, A, B, F2, K[1] );
    ROUNDu(34, B, C, D, E, A, F2, K[1] );
    ROUNDu(35, A, B, C, D, E, F2, K[1] );
    ROUNDu(36, E, A, B, C, D, F2, K[1] );
    ROUNDu(37, D, E, A, B, C, F2, K[1] );
    ROUNDu(38, C, D, E, A, B, F2, K[1] );
    ROUNDu(39, B, C, D, E, A, F2, K[1] );
		
    ROUND5(40, F3, K[2] );
    ROUND5(45, F3, K[2] );
    ROUND5(50, F3, K[2] );
    ROUND5(55, F3, K[2] );
		
    ROUND5(60, F4, K[3] );
    ROUND5(65, F4, K[3] );
    ROUND5(70, F4, K[3] );
    ROUND5(75, F4, K[3] );
		
    /* Mix in the IV again */
    MA = ADD(A, H[0]);
    MB = ADD(B, H[1]);
					
    /* Go over each vector element in turn */
    for(n=0; n < 2; n++) {
      /* Extract A and B components */
      IA = ((uInt32*) &MA)[n];
      IB = ((uInt32*) &MB)[n];
				
      /* Is this the best bit count so far? */
      if(!(IA & bitMask1Low) && !(IB & bitMask1High)) {
				/* Count bits */
				gotBits = 0;
				if(IA) {
					s = IA;
					while(!(s & 0x80000000)) {
						s <<= 1;
						gotBits++;
					}
				} else {
					gotBits = 32;
					if(IB) {
						s = IB;
						while(!(s & 0x80000000)) {
							s <<= 1;
							gotBits++;
						}
					} else {
						gotBits = 64;
					}
				}
				*best = gotBits;
				/* Regenerate the bit mask */
				maxBits = gotBits+1;
				if(maxBits < 32) {
					bitMask1Low = ~((((uInt32) 1) << (32 - maxBits)) - 1);
					bitMask1High = 0;
				} else {
					bitMask1Low = ~0;
					bitMask1High = ~((((uInt32) 1) << (64 - maxBits)) - 1);
				}
				((uInt32*) &vBitMaskLow )[0] = bitMask1Low ;
				((uInt32*) &vBitMaskLow )[1] = bitMask1Low ;
				((uInt32*) &vBitMaskHigh)[0] = bitMask1High;
				((uInt32*) &vBitMaskHigh)[1] = bitMask1High;
				
				/* Copy this result back to the block buffer, little-endian */
				for(t=0; t < 16; t++) {
					output[t*4+0] = X[t*8+3+n*4];
					output[t*4+1] = X[t*8+2+n*4];
					output[t*4+2] = X[t*8+1+n*4];
					output[t*4+3] = X[t*8+0+n*4];
				}
				
				/* Is it good enough to bail out? */
				if(gotBits >= bits) {
					/* Shut down use of MMX */
					__builtin_ia32_emms();
				  
					return iters+2;
				}
      }
    }
    MINTER_CALLBACK();
  }
	
  /* Shut down use of MMX */
  __builtin_ia32_emms();

  return iters+2;

  /* For other platforms */
#else
  return 0;
#endif
}