-
Notifications
You must be signed in to change notification settings - Fork 13
/
OpArith.cpp
843 lines (671 loc) · 22.5 KB
/
OpArith.cpp
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
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
// This file is part of the Cyclone 68000 Emulator
// Copyright (c) 2004,2011 FinalDave (emudave (at) gmail.com)
// Copyright (c) 2005-2011 Gražvydas "notaz" Ignotas (notasas (at) gmail.com)
// This code is licensed under the GNU General Public License version 2.0 and the MAME License.
// You can choose the license that has the most advantages for you.
// SVN repository can be found at http://code.google.com/p/cyclone68000/
#include "app.h"
// --------------------- Opcodes 0x0000+ ---------------------
// Emit an Ori/And/Sub/Add/Eor/Cmp Immediate opcode, 0000ttt0 ssaaaaaa
int OpArith(int op)
{
int type=0,size=0;
int sea=0,tea=0;
int use=0;
const char *shiftstr="";
// Get source and target EA
type=(op>>9)&7; if (type==4 || type>=7) return 1;
size=(op>>6)&3; if (size>=3) return 1;
sea= 0x003c;
tea=op&0x003f;
// See if we can do this opcode:
if (EaCanRead(tea,size)==0) return 1;
if (EaCanWrite(tea)==0 || EaAn(tea)) return 1;
use=OpBase(op,size);
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op, sea, tea); Cycles=4;
// imm must be read first
EaCalcRead(-1,10,sea,size,0,earwt_msb_dont_care);
EaCalcRead((type!=6)?11:-1,0,tea,size,0x003f,earwt_msb_dont_care);
if (size<2) shiftstr=(char *)(size?",asl #16":",asl #24");
if (size<2) ot(" mov r10,r10,asl #%i\n",size?16:24);
ot(";@ Do arithmetic:\n");
if (type==0) ot(" orrs r1,r10,r0%s\n",shiftstr);
if (type==1) ot(" ands r1,r10,r0%s\n",shiftstr);
if (type==2||type==6)
ot(" rsbs r1,r10,r0%s ;@ Defines NZCV\n",shiftstr);
if (type==3) ot(" adds r1,r10,r0%s ;@ Defines NZCV\n",shiftstr);
if (type==5) ot(" eors r1,r10,r0%s\n",shiftstr);
if (type< 2) OpGetFlagsNZ(1); // Ori/And
if (type==2) OpGetFlags(1,1); // Sub: Subtract/X-bit
if (type==3) OpGetFlags(0,1); // Add: X-bit
if (type==5) OpGetFlagsNZ(1); // Eor
if (type==6) OpGetFlags(1,0); // Cmp: Subtract
ot("\n");
if (type!=6)
{
EaWrite(11, 1, tea,size,0x003f,earwt_shifted_up);
}
// Correct cycles:
if (type==6)
{
if (size>=2 && tea<0x10) Cycles+=2;
}
else
{
if (size>=2) Cycles+=4;
if (tea>=8) Cycles+=4;
}
OpEnd(sea,tea);
return 0;
}
// --------------------- Opcodes 0x5000+ ---------------------
int OpAddq(int op)
{
// 0101nnnt xxeeeeee (nnn=#8,1-7 t=addq/subq xx=size, eeeeee=EA)
int num=0,type=0,size=0,ea=0;
int use=0;
char count[16]="";
int shift=0;
num =(op>>9)&7; if (num==0) num=8;
type=(op>>8)&1;
size=(op>>6)&3; if (size>=3) return 1;
ea = op&0x3f;
// See if we can do this opcode:
if (EaCanRead (ea,size)==0) return 1;
if (EaCanWrite(ea) ==0) return 1;
if (size == 0 && EaAn(ea) ) return 1;
use=OpBase(op,size,1);
if (num!=8) use|=0x0e00; // If num is not 8, use same handler
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,ea);
Cycles=ea<8?4:8;
if(size>=2) Cycles=ea<0x10?8:12;
if (size>0 && (ea&0x38)==0x08) size=2; // addq.w #n,An is also 32-bit
EaCalcRead(11,0,ea,size,0x003f,earwt_msb_dont_care);
shift=32-(8<<size);
if (num!=8)
{
int lsr=9-shift;
ot(" and r2,r8,#0x0e00 ;@ Get quick value\n");
if (lsr>=0) sprintf(count,"r2,lsr #%d", lsr);
else sprintf(count,"r2,lsl #%d", -lsr);
ot("\n");
}
else
{
sprintf(count,"#0x%.4x",8<<shift);
}
if (size<2) ot(" mov r0,r0,asl #%d\n\n",size?16:24);
if (type==0) ot(" adds r1,r0,%s\n",count);
if (type==1) ot(" subs r1,r0,%s\n",count);
if ((ea&0x38)!=0x08) OpGetFlags(type,1);
ot("\n");
EaWrite(11, 1, ea,size,0x003f,earwt_shifted_up);
OpEnd(ea);
return 0;
}
// --------------------- Opcodes 0x8000+ ---------------------
// 1t0tnnnd xxeeeeee (tt=type:or/sub/and/add xx=size, eeeeee=EA)
int OpArithReg(int op)
{
int use=0;
int type=0,size=0,dir=0,rea=0,ea=0;
const char *asl="";
const char *strop=0;
type=(op>>12)&5;
rea =(op>> 9)&7;
dir =(op>> 8)&1; // er,re
size=(op>> 6)&3; if (size>=3) return 1;
ea = op&0x3f;
if (dir && ea<0x10) return 1; // addx/subx opcode
// See if we can do this opcode:
if (dir==0 && EaCanRead (ea,size)==0) return 1;
if (dir && EaCanWrite(ea)==0) return 1;
if ((size==0||!(type&1))&&EaAn(ea)) return 1;
use=OpBase(op,size);
use&=~0x0e00; // Use same opcode for Dn
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,ea); Cycles=4;
EaCalcRead(dir?11:-1,0,ea,size,0x003f,earwt_msb_dont_care);
EaCalcRead(dir?-1:11,1,rea,size,0x0e00,earwt_msb_dont_care);
ot(";@ Do arithmetic:\n");
if (type==0) strop = "orrs";
if (type==1) strop = (char *) (dir ? "subs" : "rsbs");
if (type==4) strop = "ands";
if (type==5) strop = "adds";
if (size==0) asl=",asl #24";
if (size==1) asl=",asl #16";
if (size<2) ot(" mov r0,r0%s\n",asl);
ot(" %s r1,r0,r1%s\n",strop,asl);
if (type&1) OpGetFlags(type==1,type&1); // add/subtract
else OpGetFlagsNZ(1);
ot("\n");
ot(";@ Save result:\n");
if (size<2) ot(" mov r1,r1,asr #%d\n",size?16:24);
if (dir) EaWrite(11, 1, ea,size,0x003f,earwt_msb_dont_care);
else EaWrite(11, 1,rea,size,0x0e00,earwt_msb_dont_care);
if(rea==ea) {
if(ea<8) Cycles=(size>=2)?8:4; else Cycles+=(size>=2)?26:14;
} else if(dir) {
Cycles+=4;
if(size>=2) Cycles+=4;
} else {
if(size>=2) {
Cycles+=2;
if(ea<0x10||ea==0x3c) Cycles+=2;
}
}
OpEnd(ea);
return 0;
}
// --------------------- Opcodes 0x80c0+ ---------------------
int OpMul(int op)
{
// Div/Mul: 1m00nnns 11eeeeee (m=Mul, nnn=Register Dn, s=signed, eeeeee=EA)
int type=0,rea=0,sign=0,ea=0;
int use=0;
type=(op>>14)&1; // div/mul
rea =(op>> 9)&7;
sign=(op>> 8)&1;
ea = op&0x3f;
// See if we can do this opcode:
if (EaCanRead(ea,1)==0||EaAn(ea)) return 1;
use=OpBase(op,1);
use&=~0x0e00; // Use same for all registers
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,ea,0,1);
if(type) Cycles=38;
else Cycles=sign?16:10;
EaCalcRead(-1,0,ea,1,0x003f,earwt_msb_dont_care);
EaCalc(11,0x0e00,rea, 2);
EaRead(11, 2,rea, 2,0x0e00);
ot(" movs r1,r0,asl #16\n");
if (type==0) // div
{
// the manual says C is always cleared, but neither Musashi nor FAME do that
//ot(" bic r10,r10,#0x20000000 ;@ always clear C\n");
ot(" beq divzero%.4x ;@ division by zero\n",op);
ot("\n");
if (sign)
{
ot(" mov r12,#0 ;@ r12 = 1 or 2 if the result is negative\n");
ot(" tst r2,r2\n");
ot(" orrmi r12,r12,#2\n");
ot(" submi r5,r5,#2\n");
ot(" rsbmi r2,r2,#0 ;@ Make r2 positive\n");
ot("\n");
ot(" movs r0,r1,asr #16\n");
ot(" orrmi r12,r12,#1\n");
ot(" rsbmi r0,r0,#0 ;@ Make r0 positive\n");
ot("\n");
ot(";@ detect the nasty 0x80000000 / -1 situation\n");
ot(" subs r3,r2,#0x80000000\n");
ot(" addeqs r3,r1,#0x00010000\n");
ot(" beq wrendofop%.4x\n",op);
}
else
{
ot(" mov r0,r1,lsr #16 ;@ use only 16 bits of divisor\n");
}
ot("\n");
ot(";@ Overflow?\n");
ot(" cmp r0,r2,lsr #16\n");
ot(" orrls r10,r10,#0x10000000 ;@ set overflow flag\n");
ot(" bls endofop%.4x ;@ overflow!\n",op);
ot("\n");
ot(";@ Divide r2 by r0\n");
ot(" mov r3,#0\n");
ot(" mov r1,r0\n");
ot("\n");
ot(";@ Shift up divisor till it's just less than numerator\n");
ot("Shift%.4x%s\n",op,ms?"":":");
ot(" cmp r1,r2,lsr #1\n");
ot(" movls r1,r1,lsl #1\n");
ot(" bcc Shift%.4x\n",op);
ot("\n");
ot(" sub r5,r5,#8*16 ;@ Maximum cycles divide loop can take\n");
ot("Divide%.4x%s\n",op,ms?"":":");
ot(" cmp r2,r1\n");
ot(" adc r3,r3,r3 ;@ Double r3 and add 1 if carry set\n");
ot(" subcs r2,r2,r1\n");
ot(" addcs r5,r5,#2\n");
ot(" teq r1,r0\n");
ot(" movne r1,r1,lsr #1\n");
ot(" bne Divide%.4x\n",op);
ot("\n");
ot(";@r3==quotient,r2==remainder\n");
if (sign)
{
// sign correction
ot(" sub r5,r5,#8\n");
ot(" and r1,r12,#1\n");
ot(" teq r1,r12,lsr #1\n");
ot(" rsbne r3,r3,#0 ;@ negate if quotient is negative\n");
ot(" subne r5,r5,#2\n");
ot(" tst r12,#2\n");
ot(" rsbne r2,r2,#0 ;@ negate the remainder if divident was negative\n");
ot(" subne r5,r5,#2\n");
ot("\n");
// signed overflow check
ot(" mov r1,r3,asl #16\n");
ot(" cmp r3,r1,asr #16 ;@ signed overflow?\n");
ot(" orrne r10,r10,#0x10000000 ;@ set overflow flag\n");
ot(" bne endofop%.4x ;@ overflow!\n",op);
ot("\n");
ot("wrendofop%.4x%s\n",op,ms?"":":");
}
ot(" movs r1,r3,lsl #16 ;@ Clip to 16-bits\n");
OpGetFlagsNZ(1);
ot(" mov r1,r1,lsr #16\n");
ot(" orr r1,r1,r2,lsl #16 ;@ Insert remainder\n");
}
if (type==1)
{
ot(";@ Calculate cycles needed: 2*(#bits set in multiplier engine mask)\n");
if (sign) {
ot(" eor r0,r1,r1,asr #1\n");
ot(" tst r0,#0x8000 ;@ handle 17th bit\n");
ot(" subne r5,r5,#2\n");
} else ot(" mov r0,r1\n");
ot(" mov r3,#0x55000000 ;@ count top 16 bits, the O(1) way\n");
ot(" orr r3,r3,r3,lsr #8\n");
ot(" and r3,r3,r0,lsr #1\n");
ot(" sub r0,r0,r3\n");
ot(" mov r3,#0x33000000\n");
ot(" orr r3,r3,r3,lsr #8\n");
ot(" and r12,r3,r0,lsr #2\n");
ot(" and r0,r3,r0\n");
ot(" add r0,r0,r12\n");
ot(" mov r3,#0x0f000000\n");
ot(" orr r3,r3,r3,lsr #8\n");
ot(" add r0,r0,r0,lsr #4\n");
ot(" and r0,r3,r0\n");
ot(" add r0,r0,r0,lsl #8\n");
ot(" sub r5,r5,r0,lsr #23 ;@ cycles -= 2*bitcount(mask)\n");
ot(";@ Get 16-bit signs right:\n");
ot(" mov r0,r1,%s #16\n",sign?"asr":"lsr");
ot(" mov r2,r2,lsl #16\n");
ot(" mov r2,r2,%s #16\n",sign?"asr":"lsr");
ot("\n");
ot(" muls r1,r2,r0\n");
OpGetFlagsNZ(1);
}
ot("\n");
EaWrite(11, 1,rea, 2,0x0e00,earwt_shifted_up);
if (type==0) ot("endofop%.4x%s\n",op,ms?"":":");
opend_op_changes_cycles=1;
OpEnd(ea);
if (type==0) // div
{
ot("divzero%.4x%s\n",op,ms?"":":");
ot(" mov r0,#5 ;@ Divide by zero\n");
ot(" bl Exception\n");
Cycles+=34-6;
OpEnd(ea);
ot("\n");
}
return 0;
}
// Get X Bit into carry - trashes r2
int GetXBit(int subtract)
{
ot(";@ Get X bit:\n");
ot(" ldr r2,[r7,#0x4c]\n");
if (subtract) ot(" mvn r2,r2 ;@ Invert it\n");
ot(" tst r2,r2,lsl #3 ;@ Get into Carry\n");
ot("\n");
return 0;
}
// --------------------- Opcodes 0x8100+ ---------------------
// 1t00ddd1 0000asss - sbcd/abcd Ds,Dd or -(As),-(Ad)
int OpAbcd(int op)
{
int use=0;
int type=0,sea=0,mem=0,dea=0;
type=(op>>14)&1; // sbcd/abcd
dea =(op>> 9)&7;
mem =(op>> 3)&1;
sea = op &7;
if (mem) { sea|=0x20; dea|=0x20; }
use=op&~0x0e07; // Use same opcode for all registers..
if (sea==0x27) use|=0x0007; // ___x.b -(a7)
if (dea==0x27) use|=0x0e00; // ___x.b -(a7)
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,sea,dea); Cycles=6;
if (mem)
{
ot(";@ Get src/dest EA vals\n");
EaCalc (0,0x000f, sea,0,earwt_msb_dont_care);
EaRead (0, 6, sea,0,0x000f,earwt_msb_dont_care);
EaCalcRead(11,0,dea,0,0x0e00,earwt_msb_dont_care);
}
else
{
ot(";@ Get src/dest reg vals\n");
EaCalcRead(-1,6,sea,0,0x0007,earwt_msb_dont_care);
EaCalcRead(11,0,dea,0,0x0e00,earwt_msb_dont_care);
}
ot(" bic r10,r10,#0xb1000000 ;@ clear all flags except old Z\n");
ot(" ldr r1,[r7,#0x4c] ;@ Get X bit\n");
ot(" and r2,r0,#0x0f\n");
ot(" movs r12,r1,lsl #3 ;@ X into carry\n");
ot(" and r1,r6,#0x0f\n");
if (type)
{
// abcd
ot(" adc r1,r2,r1\n");
ot(" cmp r1,#9\n");
ot(" and r0,r0,#0xf0\n");
ot(" and r6,r6,#0xf0\n");
ot(" add r1,r1,r0\n");
ot(" add r1,r1,r6\n");
ot(" mov r12,r1\n");
ot(" addhi r12,r12,#6 ;@ Decimal adjust units\n");
ot(" tst r1,#0x80\n");
ot(" orreq r10,r10,#0x10000000 ;@ Undefined V behavior\n");
ot(" cmp r12,#0x9f\n");
ot(" orrhi r10,r10,#0x20000000 ;@ C\n");
ot(" subhi r12,r12,#0xa0\n");
ot(" movs r0,r12,lsl #24\n");
ot(" bicpl r10,r10,#0x10000000 ;@ Undefined V behavior part II\n");
}
else
{
// sbcd
ot(" mov r12,#0 ;@ corf\n");
ot(" adc r1,r1,#0\n");
ot(" sub r1,r2,r1\n");
ot(" cmp r1,#0x0f\n");
ot(" movhi r12,#6\n");
ot(" and r0,r0,#0xf0\n");
ot(" and r6,r6,#0xf0\n");
ot(" add r1,r1,r0\n");
ot(" sub r1,r1,r6\n");
ot(" tst r1,#0x80\n");
ot(" orrne r10,r10,#0x10000000 ;@ Undefined V behavior\n");
ot(" cmp r1,r12\n");
ot(" orrlt r10,r10,#0x20000000 ;@ C\n");
ot(" cmp r1,#0xff\n");
ot(" addhi r1,r1,#0xa0\n");
ot(" sub r12,r1,r12\n");
ot(" movs r0,r12,lsl #24\n");
ot(" bicmi r10,r10,#0x10000000 ;@ Undefined V behavior part II\n");
}
ot(" orrmi r10,r10,#0x80000000 ;@ Undefined N behavior\n");
ot(" bicne r10,r10,#0x40000000 ;@ Z flag\n");
ot(" str r10,[r7,#0x4c] ;@ Save X bit\n");
ot("\n");
EaWrite(11, 0, dea,0,0x0e00,earwt_shifted_up);
ot(" ldr r6,[r7,#0x54]\n");
OpEnd(sea,dea);
return 0;
}
// 01001000 00eeeeee - nbcd <ea>
int OpNbcd(int op)
{
int use=0;
int ea=0;
ea=op&0x3f;
if(EaCanWrite(ea)==0||EaAn(ea)) return 1;
use=OpBase(op,0);
if(op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,ea); Cycles=6;
if(ea >= 8) Cycles+=2;
EaCalcRead(11,0,ea,0,0x003f,earwt_zero_extend);
// this is rewrite of Musashi's code
ot(" ldr r2,[r7,#0x4c]\n");
ot(" bic r10,r10,#0xb0000000 ;@ clear all flags, except Z\n");
ot(" and r2,r2,#0x20000000\n");
ot(" rsb r1,r0,#0 ;@ do arithmetic\n");
ot(" subs r1,r1,r2,lsr #29 ;@ X\n");
ot(" beq finish%.4x\n",op);
ot("\n");
ot(" movs r1,r1,lsl #24\n");
ot(" orrmi r10,r10,#0x10000000 ;@ Undefined V behavior\n");
ot(" orr r2,r1,r0,lsl #24\n");
ot(" tst r2,#0x0f000000\n");
ot(" andeq r1,r1,#0xf0000000\n");
ot(" orreq r1,r1,#0x06000000\n");
ot(" adds r1,r1,#0x9a000000\n");
ot(" bicmi r10,r10,#0x10000000 ;@ Undefined V behavior part II\n");
ot(" orrmi r10,r10,#0x80000000 ;@ Undefined N behavior\n");
ot(" bicne r10,r10,#0x40000000 ;@ Z\n");
ot(" orr r10,r10,#0x20000000 ;@ C\n");
ot("\n");
EaWrite(11, 1, ea,0,0x3f,earwt_shifted_up);
ot("finish%.4x%s\n",op,ms?"":":");
ot(" str r10,[r7,#0x4c] ;@ Save X\n");
ot("\n");
ot(" ldr r6,[r7,#0x54]\n");
OpEnd(ea);
return 0;
}
// --------------------- Opcodes 0x90c0+ ---------------------
// Suba/Cmpa/Adda 1tt1nnnx 11eeeeee (tt=type, x=size, eeeeee=Source EA)
int OpAritha(int op)
{
int use=0;
int type=0,size=0,sea=0,dea=0;
const char *asr="";
// Suba/Cmpa/Adda/(invalid):
type=(op>>13)&3; if (type>=3) return 1;
size=(op>>8)&1; size++;
dea=(op>>9)&7; dea|=8; // Dest=An
sea=op&0x003f; // Source
// See if we can do this opcode:
if (EaCanRead(sea,size)==0) return 1;
use=OpBase(op,size);
use&=~0x0e00; // Use same opcode for An
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,sea); Cycles=(size==2)?6:8;
if(size==2&&(sea<0x10||sea==0x3c)) Cycles+=2;
if(type==1) Cycles=6;
// EA calculation order defines how situations like suba.w (A0)+, A0 get handled.
// different emus act differently in this situation, I couldn't fugure which is right behaviour.
//if (type == 1)
{
EaCalcRead(-1,0,sea,size,0x003f,earwt_msb_dont_care);
EaCalcRead(type!=1?11:-1,1,dea,2,0x0e00,earwt_msb_dont_care);
}
#if 0
else
{
EaCalcRead(type!=1?11:-1,1,dea,2,0x0e00,earwt_msb_dont_care);
EaCalcRead(-1,0,sea,size,0x003f,earwt_msb_dont_care);
}
#endif
if (size<2) ot(" mov r0,r0,asl #%d\n\n",size?16:24);
if (size<2) asr=(char *)(size?",asr #16":",asr #24");
if (type==0) ot(" sub r1,r1,r0%s\n",asr);
if (type==1) ot(" cmp r1,r0%s ;@ Defines NZCV\n",asr);
if (type==1) OpGetFlags(1,0); // Get Cmp flags
if (type==2) ot(" add r1,r1,r0%s\n",asr);
ot("\n");
if (type!=1) EaWrite(11, 1, dea,2,0x0e00);
OpEnd(sea);
return 0;
}
// --------------------- Opcodes 0x9100+ ---------------------
// Emit a Subx/Addx opcode, 1t01ddd1 zz00rsss addx.z Ds,Dd
int OpAddx(int op)
{
int use=0;
int type=0,size=0,dea=0,sea=0,mem=0;
const char *asl="";
type=(op>>14)&1;
dea =(op>> 9)&7;
size=(op>> 6)&3; if (size>=3) return 1;
sea = op&7;
mem =(op>> 3)&1;
// See if we can do this opcode:
if (EaCanRead(sea,size)==0) return 1;
if (EaCanWrite(dea)==0) return 1;
if (mem) { sea+=0x20; dea+=0x20; }
use=op&~0x0e07; // Use same opcode for Dn
if (size==0&&sea==0x27) use|=0x0007; // ___x.b -(a7)
if (size==0&&dea==0x27) use|=0x0e00; // ___x.b -(a7)
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,sea,dea); Cycles=4;
if(size>=2) Cycles+=4;
if(sea>=0x10) Cycles+=2;
if (mem)
{
ot(";@ Get src/dest EA vals\n");
EaCalc (0,0x000f, sea,size,earwt_shifted_up);
EaRead (0, 6, sea,size,0x000f,earwt_shifted_up);
EaCalcRead(11,0,dea,size,0x0e00,earwt_msb_dont_care);
}
else
{
ot(";@ Get src/dest reg vals\n");
EaCalcRead(-1,6,sea,size,0x0007,earwt_msb_dont_care);
EaCalcRead(11,0,dea,size,0x0e00,earwt_msb_dont_care);
if (size<2) ot(" mov r6,r6,asl #%d\n\n",size?16:24);
}
if (size<2) asl=(char *)(size?",asl #16":",asl #24");
ot(";@ Do arithmetic:\n");
GetXBit(type==0);
if (type==1 && size<2)
{
ot(";@ Make sure the carry bit will tip the balance:\n");
ot(" mvn r2,#0\n");
ot(" orr r6,r6,r2,lsr #%i\n",(size==0)?8:16);
ot("\n");
}
if (type==0) ot(" rscs r1,r6,r0%s\n",asl);
if (type==1) ot(" adcs r1,r6,r0%s\n",asl);
ot(" orr r3,r10,#0xb0000000 ;@ for old Z\n");
OpGetFlags(type==0,1,0); // subtract
if (size<2) {
ot(" movs r2,r1,lsr #%i\n", size?16:24);
ot(" orreq r10,r10,#0x40000000 ;@ add potentially missed Z\n");
}
ot(" andeq r10,r10,r3 ;@ fix Z\n");
ot("\n");
ot(";@ Save result:\n");
EaWrite(11, 1, dea,size,0x0e00,earwt_shifted_up);
ot(" ldr r6,[r7,#0x54]\n");
OpEnd(sea,dea);
return 0;
}
// --------------------- Opcodes 0xb000+ ---------------------
// Emit a Cmp/Eor opcode, 1011rrrt xxeeeeee (rrr=Dn, t=cmp/eor, xx=size extension, eeeeee=ea)
int OpCmpEor(int op)
{
int rea=0,eor=0;
int size=0,ea=0,use=0;
const char *asl="";
// Get EA and register EA
rea=(op>>9)&7;
eor=(op>>8)&1;
size=(op>>6)&3; if (size>=3) return 1;
ea=op&0x3f;
if (eor && (ea>>3) == 1) return 1; // not a valid mode for eor
// See if we can do this opcode:
if (EaCanRead(ea,size)==0) return 1;
if (eor && EaCanWrite(ea)==0) return 1;
if (EaAn(ea)&&(eor||size==0)) return 1;
use=OpBase(op,size);
use&=~0x0e00; // Use 1 handler for register d0-7
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,ea); Cycles=4;
if(eor) {
if(ea>8) Cycles+=4;
if(size>=2) Cycles+=4;
} else {
if(size>=2) Cycles+=2;
}
ot(";@ Get EA into r11 and value into r0:\n");
EaCalcRead(eor?11:-1,0,ea,size,0x003f,earwt_msb_dont_care);
ot(";@ Get register operand into r1:\n");
EaCalcRead(-1,1,rea,size,0x0e00,earwt_msb_dont_care);
if (size<2) ot(" mov r0,r0,asl #%d\n\n",size?16:24);
if (size<2) asl=(char *)(size?",asl #16":",asl #24");
ot(";@ Do arithmetic:\n");
if (eor)
{
ot(" eors r1,r0,r1%s\n",asl);
OpGetFlagsNZ(1);
}
else
{
ot(" rsbs r1,r0,r1%s\n",asl);
OpGetFlags(1,0); // Cmp like subtract
}
ot("\n");
if (eor) EaWrite(11, 1,ea,size,0x003f,earwt_shifted_up);
OpEnd(ea);
return 0;
}
// Emit a Cmpm opcode, 1011ddd1 xx001sss (rrr=Adst, xx=size extension, sss=Asrc)
int OpCmpm(int op)
{
int size=0,sea=0,dea=0,use=0;
const char *asl="";
// get size, get EAs
size=(op>>6)&3; if (size>=3) return 1;
sea=(op&7)|0x18;
dea=(op>>9)&0x3f;
use=op&~0x0e07; // Use 1 handler for all registers..
if (size==0&&sea==0x1f) use|=0x0007; // ..except (a7)+
if (size==0&&dea==0x1f) use|=0x0e00;
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,sea); Cycles=4;
ot(";@ Get src operand into r11:\n");
EaCalc (0,0x0007, sea,size,earwt_shifted_up);
EaRead (0, 11, sea,size,0x0007,earwt_shifted_up);
ot(";@ Get dst operand into r0:\n");
EaCalcRead(-1,0,dea,size,0x0e00,earwt_msb_dont_care);
if (size<2) asl=(char *)(size?",asl #16":",asl #24");
ot(" rsbs r0,r11,r0%s\n",asl);
OpGetFlags(1,0); // Cmp like subtract
ot("\n");
OpEnd(sea);
return 0;
}
// Emit a Chk opcode, 0100ddd1 x0eeeeee (rrr=Dn, x=size extension, eeeeee=ea)
int OpChk(int op)
{
int rea=0;
int size=0,ea=0,use=0;
// Get EA and register EA
rea=(op>>9)&7;
if((op>>7)&1)
size=1; // word operation
else size=2; // long
ea=op&0x3f;
if (EaAn(ea)) return 1; // not a valid mode
if (size!=1) return 1; // 000 variant only supports word
// See if we can do this opcode:
if (EaCanRead(ea,size)==0) return 1;
use=OpBase(op,size);
use&=~0x0e00; // Use 1 handler for register d0-7
if (op!=use) { OpUse(op,use); return 0; } // Use existing handler
OpStart(op,ea); Cycles=10;
ot(";@ Get value into r0:\n");
EaCalcRead(-1,0,ea,size,0x003f,earwt_msb_dont_care);
ot(";@ Get register operand into r1:\n");
EaCalcRead(-1,1,rea,size,0x0e00,earwt_msb_dont_care);
if (size<2) ot(" mov r0,r0,asl #%d\n",size?16:24);
if (size<2) ot(" movs r1,r1,asl #%d\n\n",size?16:24);
else ot(" adds r1,r1,#0 ;@ Define flags\n");
ot(";@ get flags, including undocumented ones\n");
ot(" and r3,r10,#0x80000000\n");
OpGetFlagsNZ(1);
ot(";@ is reg negative?\n");
ot(" bmi chktrap%.4x\n",op);
ot(";@ Do arithmetic:\n");
ot(" cmp r1,r0\n");
ot(" bgt chktrap%.4x\n",op);
ot(";@ old N remains\n");
ot(" orr r10,r10,r3\n");
OpEnd(ea);
ot("chktrap%.4x%s ;@ CHK exception:\n",op,ms?"":":");
ot(" mov r0,#6\n");
ot(" bl Exception\n");
Cycles+=34-6;
OpEnd(ea);
return 0;
}
// vim:ts=2:sw=2:expandtab