-
Notifications
You must be signed in to change notification settings - Fork 20
/
disp.c
3162 lines (2695 loc) · 83.6 KB
/
disp.c
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
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/******************************************************************************
* Name : disp.c
* Title : ISP plane packing routines for PowerSGL Direct
* Author : Jim Page
* Created : 14/10/1996
*
* Copyright : 1995-2022 Imagination Technologies (c)
* License : MIT
*
*
* Description :
*
* Platform : ANSI
*
* Modifications:
* $Log: disp.c,v $
* Revision 1.65 1998/08/13 09:50:34 ls
* NewmanHaas was dropping triangles in the release build but not in
* the debug. (Fine in the release without the -Qip option of the CFLAGS).
* Tracked this down to the (FLOAT_TO_LONG)fMaxVal < LONG_INVMAGIC - a speed
* up for the pentium. fMaxVal was typically 0.00000 so I presume the FLOAT_TO
* LONG has a problem with this. Forcing it to do the float compare once agai
* fixes the game.
*
* Revision 1.64 1997/11/10 14:18:19 gdc
* made depth bias work
*
* Revision 1.63 1997/11/07 15:59:45 gdc
* added depth bias - but not tested yet
*
* Revision 1.62 1997/10/16 11:36:50 sxy
* removed DPFDEVFs, its all in DPFDEV.
*
* Revision 1.61 1997/10/15 11:05:34 sxy
* added some DPFDEVF messages.
*
* Revision 1.60 1997/09/24 14:29:57 sxy
* changed some DPF to DPFDEV.
*
* Revision 1.59 1997/08/28 13:58:25 gdc
* set number back from 5 to 4
*
* Revision 1.58 1997/08/27 16:48:06 gdc
* shelias fix for vingetting
*
* Revision 1.57 1997/08/26 08:46:01 gdc
* fixed a compiler warning and removed uneeded code
*
* Revision 1.56 1997/08/22 16:35:40 sxy
* Fixed a bug in counting TSPSize in PackISPPolygonExtra.
*
* Revision 1.55 1997/08/21 16:32:55 sxy
* added vertex fog in quads routines.
*
* Revision 1.54 1997/08/18 16:24:18 gdc
* started adding stuff for vertex for and smooth highlights,
* PCX2 and triangles only at the moment
*
* Revision 1.53 1997/08/11 14:00:49 gdc
* removed ISPTSP compatibility
*
* Revision 1.52 1997/08/08 19:41:28 mjg
* Added PCX2_003 compatibility.
*
* Revision 1.51 1997/07/16 10:44:33 mjg
* Recalculated the DataSize variable and added the MODULE_ID.
*
* Revision 1.50 1997/06/17 09:28:34 mjg
* Added the FORCE_NO_FPU option.
*
* Revision 1.49 1997/06/09 15:46:59 mjg
* Hopefully fixed some compiler warnings and broken nothing:
* 1) Removed a rather unpleasant l-value cast.
* 2) Added a cast to resolve a signed/unsigned equality.
* 3) For ISPTSP and PCX1 removed long/float mismatch by pre-calculating
* the FLOAT_TO_FIXED * (+/-) 0.5f value.
* 4) Removed some token ^Ms.
*
* Revision 1.48 1997/06/06 10:00:28 erf
* Fixed clipped tested PCX1 & ISPTSP packing routines, when special
* case required.
*
* Revision 1.47 1997/06/04 08:57:12 erf
* Fixed clipping code for PCX1 and ISPTSP. The PackISP clip tested
* routines are in C. These will have to be converted to assembler
* at some stage.
*
* Revision 1.46 1997/06/03 15:55:15 erf
* Added Simon's clipping fix to PCX1 & ISPTSP and for N sided polygons.
*
* Revision 1.45 1997/06/02 14:15:51 sjf
* WARNING CONTAINS UNTESTED CODE!
* ===============================
* Added the "unclipped" version of the pack routines which will
* test the ranges of the Edge parameters before packing them.
* This has been tested fairly well (via tombraider) with PCX2 but
* not with PCX1.
*
* Please also note: It has not been done for the polygon core (only
* the triangles), and it should also be added to SGL.
*
* Revision 1.44 1997/06/02 02:44:06 jop
* Small changed to improve asm intermediate builds
*
* Revision 1.43 1997/05/19 09:02:08 gdc
* moved get sabre limit to parabuff.h
*
* Revision 1.42 1997/05/19 07:35:17 jop
* Fixed sabre buffer limit checks
*
* Revision 1.41 1997/05/02 07:42:44 erf
* Removed UINT32's etc and replaced with sgl_uint32's etc.
*
* Revision 1.40 1997/04/22 18:00:44 mjg
* Removed brdsetup.h and altered the SabreMem functions.
*
* Revision 1.39 1997/04/21 15:24:16 mjg
* Replaced ParamBufferInfo with PVRParamBuffs.
*
* Revision 1.38 1997/04/08 13:52:59 jop
* Cleaned up includes a bit
*
* Revision 1.37 1997/04/02 13:43:17 gdc
* a couple of defines with prepended _'s so dispml.asm funcs link
*
* Revision 1.36 1997/03/31 22:05:55 jop
* rationalised WIN32s, #include "windows.h", and stuck in some DOS32s
*
* Revision 1.35 1997/03/14 18:08:08 jop
* Added pvrosapi.h include!
*
* Revision 1.34 1997/03/14 15:56:32 jwf
* Added HWIO macros for macintosh compatability
*
* Revision 1.33 1997/03/14 13:04:19 jop
* Fixed some // comments fletch put in
*
* Revision 1.32 1997/03/14 11:05:01 jwf
* Made some changes to make it compile on the macintosh
*
* Revision 1.31 1997/02/24 10:40:06 gdc
* removed a PCX2 conditional bit of code from within a !PCX2 bit of code
* and neatened up a bit
*
* Revision 1.30 1997/01/30 09:46:53 erf
* Revised PCX2 optimization mentioned in 1.27.
*
* Revision 1.29 1997/01/22 16:19:06 erf
* Fixed PCX2 floating point hardware bug in software.
* Tidied up PCX2 packing code. ie seperated for PCX1 and ISPTSP.
*
* Revision 1.28 1997/01/21 11:46:06 mjg
* Removed changes made in version 1.26.
*
* Revision 1.27 1997/01/21 11:03:14 erf
* Optimized PCX2 path.
*
* Revision 1.26 1997/01/17 14:42:09 mjg
* Test to stop very complex scenes on midas4 from hanging - effective?
*
* Revision 1.25 1997/01/16 21:26:29 erf
* Fixed PCX2 problem where meshs on edge of screen were not displayed.
*
* Revision 1.24 1997/01/03 14:31:52 rgi
* Replaced the SGLTRI_SPARE_TIME metric with PACK_ISPCORE_TIME
* so that it is a bit more intuitive.
*
* Revision 1.23 1996/12/11 19:37:39 mjg
* Put in a check to ensure the sabre limit for non-PCX builds.
*
* Revision 1.22 1996/12/10 23:21:54 erf
* Changed compile switches:
* 1. PCX1 to PCX1 || PCX2.
* 2. PVR1 to ISPTSP.
* 3. PVR2 to PCX1.
* 4. PVR3 to PCX2.
*
* Revision 1.21 1996/12/06 15:38:56 jop
* Fixed netrace crashing bug (null trans depth struct ptr)
*
* Revision 1.20 1996/12/05 20:55:14 jop
* Removed asm compile flags to dtri.h
*
* Revision 1.19 1996/12/05 18:03:45 jop
* Fixed near-perp bug
*
* Revision 1.18 1996/12/05 13:32:16 erf
* Add #elif to use assembler verions if WIN32 && PVR3.
*
* Revision 1.17 1996/12/03 10:56:42 rgi
* Added assembler PackISPPolygonCore after making the
* C more efficient and easier to rip off.
*
* Revision 1.16 1996/12/02 13:29:28 mjg
* Sorted out the FLOAT_TO_LONG(INVMAGIC) problem.
*
* Revision 1.15 1996/12/02 13:13:32 jop
* Fixed processing of extra edge in poly pack routine
*
* Revision 1.14 1996/12/02 11:20:40 mjg
* Changed a nested #define that wouldn't compile for Midas3.
*
* Revision 1.13 1996/11/30 04:52:48 jop
* Fixed a comment bug and went through poly code to hopefully
* accelerate, and bring the architecture into line with the
* rest of the new regime. Initial results are disappointing.
* Ah well, there's always the Severn bridge.
*
* Revision 1.12 1996/11/30 00:39:34 sjf
* Small amount of optimisation (mainly doing INT compares instead of float)
* Well OK the optimisation is only for pentium.
*
* Revision 1.11 1996/11/29 19:26:58 jop
* Things
*
* Revision 1.10 1996/11/28 17:15:14 gdc
* put define for InPackTo20Bit to PackTo20Bit when pto20b.h is included
* to stop link undef'd extern error
*
* Revision 1.9 1996/11/28 15:53:51 sjf
* Moved pack to 20 bits into its own file so that we are no longer duplicating
* it.
*
* Revision 1.8 1996/11/28 12:11:15 sjf
* 1) Got rid of those ^Ms
* 2) Redid the maths to use the new basis vectors which makes a lot
* of the maths disappear up its own sphincter.
* 3) Ummm Haven't actually compiled it yet SO it may have some
* syntax errors.
*
* Revision 1.7 1996/11/27 19:00:01 mjg
* Started messin' about with translucency sorting.
*
* Revision 1.6 1996/11/27 00:08:27 jop
* Improved pcx buffer efficiency
*
* Revision 1.5 1996/11/23 04:14:50 jop
* Structure changes
*
* Revision 1.4 1996/11/22 02:24:30 jrc
* First attempt at polygons.
*
* Revision 1.3 1996/11/19 11:27:49 rgi
* Modified to work with METRICS and the new assembler
* version done for performance dispml.asm in win32.
*
* Revision 1.2 1996/11/19 00:53:42 jop
* Got it to build
*
* Revision 1.1 1996/11/18 18:29:36 jop
* Initial revision
*
*
*****************************************************************************/
#define MODULE_ID MODID_DISP
#include "sgl_defs.h"
#include "pvrosapi.h"
#include "metrics.h"
#include "pvrlims.h"
#include "hwinterf.h"
#include "parmbuff.h"
#include "sgl_math.h"
#include "dregion.h" /* JWF added for missing prototypes */
#include "pmsabre.h"
#include "dtri.h"
#include "disp.h"
SGL_EXTERN_TIME_REF /* if we are timing code */
#define UPPER_6_OF_TAG 0x3F000UL
#define SFLOAT_20BIT_ZERO (0)
#define FLOAT_TO_FIXED ((float) (((sgl_uint32)1) <<31))
#define MAGIC (1024.0f)
#define INVMAGIC (1.0f/MAGIC)
#if SLOW_FCMP && !MULTI_FP_REG
/* A bit messy, but the pre-processor was unable to deal with the
* value "FLOAT_TO_LONG (INVMAGIC)".
*/
#define LONG_INVMAGIC 0x3A800000
#endif
/*
// Define some dummy Tags for the planes which don't use them, but
// Would be useful for debugging. Note in the non debug, use small
// 16 bit constants as these should be loaded faster.
*/
#if DEBUG
#define DEBUG_CLIP_PLANE_TAG 0xAAAAAAUL
#define DEBUG_INVIS_FP_TAG 0xBBBBBBUL
#define DEBUG_INVIS_RP_TAG 0xCCCCCCUL
#define DEBUG_PERP_TAG 0xDDDDDDUL
#define DEBUG_DUMMY_F_TAG 0xEEEEEEUL
#define DEBUG_DUMMY_R_TAG 0xFFFFFFUL
#else
#define DEBUG_CLIP_PLANE_TAG 0x000000UL
#define DEBUG_INVIS_FP_TAG 0x000000UL
#define DEBUG_INVIS_RP_TAG 0x000000UL
#define DEBUG_PERP_TAG 0x000000UL
#define DEBUG_DUMMY_F_TAG 0x000000UL
#define DEBUG_DUMMY_R_TAG 0x000000UL
#endif
extern float gfDepthBias;
static TRANS_REGION_DEPTHS_STRUCT gDepthInfo[IBUFFERSIZE];
/* this array indicates whether polys have a vfog or gouraud specular plane */
static FOGHIGHLIGHT eExtraStatus[IBUFFERSIZE];
/* ////////////////////////////////////////////////////
// Include the correct pack to 20 Bit
//////////////////////////////////////////////////// */
#if PACKISPCORE || PACKISPPOLYGONCORE || PACKISPCORECLIPTESTED || PACKISPPOLYGONCORECLIPTESTED
#if 1 /* Change if verifying assembler stuff */
/*
// Include the real C Packto20bit code
*/
#define INLINE_P20 static INLINE
#include "pto20b.h"
#define InPackTo20Bit PackTo20Bit /* Try C version */
#undef INLINE_P20
#else
/* For debuggin assembler equivalents */
extern sgl_int32 PackTo20Bit(const float Value); /* pkisp.c version */
extern sgl_int32 PackTo20BitTest(const float Value); /* dispml.asm version */
#if 0
#define InPackTo20Bit PackTo20BitTest /* Try ASM version */
#elif 1
#define InPackTo20Bit PackTo20Bit /* Try C version */
#else
#define InPackTo20Bit PackTo20BitTrial /* Try both? */
static INLINE sgl_int32 PackTo20BitTrial( const float Value )
{
sgl_int32 Res1 = PackTo20Bit(Value);
sgl_int32 Res2 = PackTo20BitTest(Value);
if ( Res1 != Res2 ) __asm int 3;
return (Res1);
}
#endif
#endif
#endif /* #if PACKISPCORE || PACKISPPOLYGONCORE */
/**********************************************************************
**
** Function to generate the pittance of depth information used
** by the list sorting of translucent objects in dregion.c
** At one point some tried to put in dz/dx and dz/dy parameters
** to improve sorting, but I don't think it has been fully implemented
**
**********************************************************************/
void GenerateDepthInfo(PITRI pTri, sgl_uint32 nPolys)
{
TRANS_REGION_DEPTHS_STRUCT* pGDepth;
pGDepth = &gDepthInfo[0];
for ( ; nPolys; nPolys--, pGDepth++)
{
/* static variable should init these to zero */
#if 0
pGDepth->DepthDx = 0.0f;
pGDepth->DepthDy = 0.0f;
#endif
/*
// Find and save the miniumum of the "Z" (i.e. 1/w) values.
// Now, to my knowledge, these Z values are all positive, so
// fcloate compare is nad, use Integer instead.
*/
ASSERT(pTri->fZ[0] >= 0);
ASSERT(pTri->fZ[1] >= 0);
ASSERT(pTri->fZ[2] >= 0);
#if (SLOW_FCMP && !MULTI_FP_REG)
{
long lTemp;
lTemp = MIN( FLOAT_TO_LONG(pTri->fZ[2]),
MIN( FLOAT_TO_LONG(pTri->fZ[0]),
FLOAT_TO_LONG(pTri->fZ[1])));
pGDepth->BaseDepth = LONG_TO_FLOAT(lTemp);
}
#else
pGDepth->BaseDepth = MIN(pTri->fZ[2], MIN(pTri->fZ[0], pTri->fZ[1]));
#endif
pTri ++;
}
}
#if (PCX2 || PCX2_003) && !FORCE_NO_FPU
/**********************************************************************
**
** Hardware: PCX2, PCX2_003
** No assembler equivalent
** Packs the front face and three edges
** This routine is similar to the below except it also checks the Edges to make
** sure they don't go out of range
** It also add the extra planes needed for smooth specular and vertex fog
**
**********************************************************************/
void PackISPCoreClipTestedExtra(sgl_uint32 *pPlaneMem,
sgl_uint32 nPolysInChunk,
sgl_uint32 *rTSPAddr,
PITRI *rpTri,
int nIncrement,
PIMATERIAL *rpMat,
int nMatInc,
sgl_uint32 TSPIncrement,
sgl_uint32 *rTSPOffsetAddr,
sgl_uint32 TSPIncrementExtra)
{
float fA, fB, fC;
float fAe, fBe, fCe;
float fZ0, fZ1, fZ2;
float f1OverDet;
float fMaxVal;
PITRI pTri = *rpTri;
sgl_uint32 TSPAddr = *rTSPAddr;
sgl_uint32 TSPOffsetAddr = *rTSPOffsetAddr;
PIMATERIAL pMat = *rpMat;
PFOGHIGHLIGHT epExtras = eExtraStatus;
while (nPolysInChunk--)
{
fZ2 = pTri->fZ[2];
f1OverDet = pTri->f1OverDet;
fZ0 = (pTri->fZ[0] - fZ2) * f1OverDet;
fZ1 = (pTri->fZ[1] - fZ2) * f1OverDet;
/*
// Compute initial A, B and C values before scaling, using the projected
// vertex positions:
//
// NOTE: I have changed the calculation to use a different basis system
// which means that the new "adjoint" which actually re-uses values from
// the adjoint we pass in except for the last row which is 0,0, Det.
// Simon.
*/
fA = fZ0*pTri->fAdjoint[0][0] + fZ1*pTri->fAdjoint[0][1];
fB = fZ0*pTri->fAdjoint[1][0] + fZ1*pTri->fAdjoint[1][1];
fC = fZ0*pTri->fAdjoint[2][0] + fZ1*pTri->fAdjoint[2][1] + fZ2 + gfDepthBias;
FW( pPlaneMem, 0+0, fA);
FW( pPlaneMem, 1+0, fB);
/* Check C value. PCX2 hardware has problems converting a number in
* the range -9.313e-10 to 0. If C is in this range set to 0.
*/
#if PCX2
if (sfabs(fC) <= MAGIC_FLOAT_FIX)
{
/* If problem set C to 0.
*/
fC = 0.0f;
IW( pPlaneMem, 2+0, 0);
}
else
#endif
/* Continue as normal.
*/
FW( pPlaneMem, 2+0, fC);
/*
// to prevent the 20Bit SFLOATS from overflowing (currently they are 2x
// larger than I originally thought necessary)
*/
fMaxVal = sfabs (fA + fB + (fC * INVMAGIC)) + (sfabs (fA) + sfabs (fB));
/*
// Once again, this float value is always positive. On Pentium, the
// float compare is awful, so do it in INT instead.
*/
/* Doing it the INT way fails when fMaxVal is 0.0 - so have to go back to the
float way. This problem was highlighted by the release build with the -Qip
option of the CFLAGS */
if (fMaxVal > INVMAGIC)
{
fA = 0.0f;
fB = 0.0f;
fC = fZ2;
IW( pPlaneMem, 0+0, 0);
IW( pPlaneMem, 1+0, 0);
FW( pPlaneMem, 2+0, fZ2);
}
/* Pack instruction and TSP address.
*/
IW( pPlaneMem, 3+0, (forw_visib_fp | (TSPAddr << 4)));
/* do first edge plane
*/
fAe = pTri->fAdjoint[0][2];
fBe = pTri->fAdjoint[1][2];
fCe = pTri->fAdjoint[2][2];
/* Check the range*/
if(sfabs((2*INVMAGIC) * fCe + fAe + fBe) > (sfabs(fAe) + sfabs(fBe)))
{
fAe = 0.0f;
fBe = 0.0f;
}
FW( pPlaneMem, 0+4, fAe);
FW( pPlaneMem, 1+4, fBe);
FW( pPlaneMem, 2+4, fCe);
IW( pPlaneMem, 3+4, (forw_perp | (DEBUG_PERP_TAG << 4)));
/* do second edge plane
*/
fAe = pTri->fAdjoint[0][0];
fBe = pTri->fAdjoint[1][0];
fCe = pTri->fAdjoint[2][0];
/* Check the range*/
if(sfabs((2*INVMAGIC) * fCe + fAe + fBe) > (sfabs(fAe) + sfabs(fBe)))
{
fAe = 0.0f;
fBe = 0.0f;
}
FW( pPlaneMem, 0+8, fAe);
FW( pPlaneMem, 1+8, fBe);
FW( pPlaneMem, 2+8, fCe);
IW( pPlaneMem, 3+8, (forw_perp | (DEBUG_PERP_TAG << 4)));
/* do third edge plane
*/
fAe = pTri->fAdjoint[0][1];
fBe = pTri->fAdjoint[1][1];
fCe = pTri->fAdjoint[2][1];
/* Check the range*/
if(sfabs((2*INVMAGIC) * fCe + fAe + fBe) > (sfabs(fAe) + sfabs(fBe)))
{
fAe = 0.0f;
fBe = 0.0f;
}
FW( pPlaneMem, 0+12, fAe);
FW( pPlaneMem, 1+12, fBe);
FW( pPlaneMem, 2+12, fCe);
IW( pPlaneMem, 3+12, (forw_perp | (DEBUG_PERP_TAG << 4)));
TSPAddr += TSPIncrement;
pPlaneMem += WORDS_PER_PLANE * 4;
/* if the smooth specular plane is added ... */
if(pMat->nSmoothSpecularIndex)
{
/* we saved the values that were used for the front plane */
FW( pPlaneMem, 0+0, fA);
FW( pPlaneMem, 1+0, fB);
FW( pPlaneMem, 2+0, fC);
IW( pPlaneMem, 3+0, (forw_visib | (TSPOffsetAddr << 4)));
*epExtras = GOURAUD_HIGHLIGHT;
pPlaneMem += WORDS_PER_PLANE;
TSPOffsetAddr += TSPIncrementExtra;
}
else
{
/* no extra plane */
*epExtras = NO_FOG_AND_HIGHLIGHT;
}
/* if the fog plane is added ... */
if(pMat->nFogIndex)
{
/* we saved the values that were used for the front plane */
FW( pPlaneMem, 0+0, fA);
FW( pPlaneMem, 1+0, fB);
FW( pPlaneMem, 2+0, fC);
IW( pPlaneMem, 3+0, (forw_visib | (TSPOffsetAddr << 4)));
*epExtras |= VERTEX_FOG;
pPlaneMem += WORDS_PER_PLANE;
TSPOffsetAddr += TSPIncrementExtra;
}
*((sgl_uint32 *) &pTri) += nIncrement;
*((sgl_uint32 *) &pMat) += nMatInc;
epExtras++;
}
*rpTri = pTri;
*rpMat = pMat;
*rTSPAddr = TSPAddr;
*rTSPOffsetAddr = TSPOffsetAddr;
}
/**********************************************************************
**
** Hardware: PCX2, PCX2_003
** No assembler equivalent
** Packs the front face and three edges - technically a Triangle
** Also conditionally adds an extra face if there is vertex fog
** on the triangle
**
**********************************************************************/
void PackISPCoreExtra(sgl_uint32 *pPlaneMem,
sgl_uint32 nPolysInChunk,
sgl_uint32 *rTSPAddr,
PITRI *rpTri,
int nIncrement,
PIMATERIAL *rpMat,
int nMatInc,
sgl_uint32 TSPIncrement,
sgl_uint32 *rTSPOffsetAddr,
sgl_uint32 TSPIncrementExtra)
{
float fA, fB, fC;
float fAe, fBe, fCe;
float fZ0, fZ1, fZ2;
float f1OverDet;
float fMaxVal;
PITRI pTri = *rpTri;
PIMATERIAL pMat = *rpMat;
sgl_uint32 TSPAddr = *rTSPAddr;
sgl_uint32 TSPOffsetAddr = *rTSPOffsetAddr;
PFOGHIGHLIGHT epExtras = eExtraStatus;
while (nPolysInChunk--)
{
fZ2 = pTri->fZ[2];
f1OverDet = pTri->f1OverDet;
fZ0 = (pTri->fZ[0] - fZ2) * f1OverDet;
fZ1 = (pTri->fZ[1] - fZ2) * f1OverDet;
/*
// Compute initial A, B and C values before scaling, using the projected
// vertex positions:
//
// NOTE: I have changed the calculation to use a different basis system
// which means that the new "adjoint" which actually re-uses values from
// the adjoint we pass in except for the last row which is 0,0, Det.
// Simon.
*/
fA = fZ0*pTri->fAdjoint[0][0] + fZ1*pTri->fAdjoint[0][1];
fB = fZ0*pTri->fAdjoint[1][0] + fZ1*pTri->fAdjoint[1][1];
fC = fZ0*pTri->fAdjoint[2][0] + fZ1*pTri->fAdjoint[2][1] + fZ2 + gfDepthBias;
FW( pPlaneMem, 0+0, fA);
FW( pPlaneMem, 1+0, fB);
/* Check C value. PCX2 hardware has problems converting a number in
* the range -9.313e-10 to 0. If C is in this range set to 0.
*/
#if PCX2
if (sfabs(fC) <= MAGIC_FLOAT_FIX)
{
/* If problem set C to 0.
*/
IW( pPlaneMem, 2+0, 0);
fC = 0.0f;
}
else
#endif
/* Continue as normal.
*/
FW( pPlaneMem, 2+0, fC);
/*
// to prevent the 20Bit SFLOATS from overflowing (currently they are 2x
// larger than I originally thought necessary)
*/
fMaxVal = sfabs (fA + fB + (fC * INVMAGIC)) + (sfabs (fA) + sfabs (fB));
/*
// Once again, this float value is always positive. On Pentium, the
// float compare is awful, so do it in INT instead.
*/
#if SLOW_FCMP && !MULTI_FP_REG
if (FLOAT_TO_LONG(fMaxVal) > LONG_INVMAGIC )
#else
if (fMaxVal > INVMAGIC)
#endif
{
fA = 0.0f;
fB = 0.0f;
fC = fZ2;
IW( pPlaneMem, 0+0, 0);
IW( pPlaneMem, 1+0, 0);
FW( pPlaneMem, 2+0, fZ2);
}
/* Pack instruction and TSP address.
*/
IW( pPlaneMem, 3+0, (forw_visib_fp | (TSPAddr << 4)));
/* do first edge plane
*/
fAe = pTri->fAdjoint[0][2];
fBe = pTri->fAdjoint[1][2];
fCe = pTri->fAdjoint[2][2];
FW( pPlaneMem, 0+4, fAe);
FW( pPlaneMem, 1+4, fBe);
FW( pPlaneMem, 2+4, fCe);
IW( pPlaneMem, 3+4, (forw_perp | (DEBUG_PERP_TAG << 4)));
/* do second edge plane
*/
fAe = pTri->fAdjoint[0][0];
fBe = pTri->fAdjoint[1][0];
fCe = pTri->fAdjoint[2][0];
FW( pPlaneMem, 0+8, fAe);
FW( pPlaneMem, 1+8, fBe);
FW( pPlaneMem, 2+8, fCe);
IW( pPlaneMem, 3+8, (forw_perp | (DEBUG_PERP_TAG << 4)));
/* do third edge plane
*/
fAe = pTri->fAdjoint[0][1];
fBe = pTri->fAdjoint[1][1];
fCe = pTri->fAdjoint[2][1];
FW( pPlaneMem, 0+12, fAe);
FW( pPlaneMem, 1+12, fBe);
FW( pPlaneMem, 2+12, fCe);
IW( pPlaneMem, 3+12, (forw_perp | (DEBUG_PERP_TAG << 4)));
TSPAddr += TSPIncrement;
pPlaneMem += WORDS_PER_PLANE * 4;
/* if the smooth specular plane is added ... */
if(pMat->nSmoothSpecularIndex)
{
/* we saved the values that were used for the front plane */
FW( pPlaneMem, 0+0, fA);
FW( pPlaneMem, 1+0, fB);
FW( pPlaneMem, 2+0, fC);
IW( pPlaneMem, 3+0, (forw_visib | (TSPOffsetAddr << 4)));
*epExtras = GOURAUD_HIGHLIGHT;
pPlaneMem += WORDS_PER_PLANE;
TSPOffsetAddr += TSPIncrementExtra;
}
else
{
/* no extra plane */
*epExtras = NO_FOG_AND_HIGHLIGHT;
}
/* if the fog plane is added ... */
if(pMat->nFogIndex)
{
/* we saved the values that were used for the front plane */
FW( pPlaneMem, 0+0, fA);
FW( pPlaneMem, 1+0, fB);
FW( pPlaneMem, 2+0, fC);
IW( pPlaneMem, 3+0, (forw_visib | (TSPOffsetAddr << 4)));
*epExtras |= VERTEX_FOG;
pPlaneMem += WORDS_PER_PLANE;
TSPOffsetAddr += TSPIncrementExtra;
}
*((sgl_uint32 *) &pTri) += nIncrement;
*((sgl_uint32 *) &pMat) += nMatInc;
epExtras++;
}
*rpTri = pTri;
*rpMat = pMat;
*rTSPAddr = TSPAddr;
*rTSPOffsetAddr = TSPOffsetAddr;
}
/**********************************************************************
**
** Hardware: PCX2, PCX2_003
** No assembler equivalent
** Packs the front face and three edges - technically a Triangle
**
**********************************************************************/
void PackISPCore(sgl_uint32 *pPlaneMem, sgl_uint32 nPolysInChunk, sgl_uint32 *rTSPAddr, PITRI *rpTri,
int nIncrement, sgl_uint32 TSPIncrement)
{
float fA, fB, fC;
float fZ0, fZ1, fZ2;
float f1OverDet;
float fMaxVal;
PITRI pTri = *rpTri;
sgl_uint32 TSPAddr = *rTSPAddr;
while (nPolysInChunk--)
{
fZ2 = pTri->fZ[2];
f1OverDet = pTri->f1OverDet;
fZ0 = (pTri->fZ[0] - fZ2) * f1OverDet;
fZ1 = (pTri->fZ[1] - fZ2) * f1OverDet;
/*
// Compute initial A, B and C values before scaling, using the projected
// vertex positions:
//
// NOTE: I have changed the calculation to use a different basis system
// which means that the new "adjoint" which actually re-uses values from
// the adjoint we pass in except for the last row which is 0,0, Det.
// Simon.
*/
fA = fZ0*pTri->fAdjoint[0][0] + fZ1*pTri->fAdjoint[0][1];
fB = fZ0*pTri->fAdjoint[1][0] + fZ1*pTri->fAdjoint[1][1];
fC = fZ0*pTri->fAdjoint[2][0] + fZ1*pTri->fAdjoint[2][1] + fZ2 + gfDepthBias;
FW( pPlaneMem, 0+0, fA);
FW( pPlaneMem, 1+0, fB);
/* Check C value. PCX2 hardware has problems converting a number in
* the range -9.313e-10 to 0. If C is in this range set to 0.
*/
#if PCX2
if (sfabs(fC) <= MAGIC_FLOAT_FIX)
/* If problem set C to 0.
*/
IW( pPlaneMem, 2+0, 0);
else
#endif
/* Continue as normal.
*/
FW( pPlaneMem, 2+0, fC);
/*
// to prevent the 20Bit SFLOATS from overflowing (currently they are 2x
// larger than I originally thought necessary)
*/
fMaxVal = sfabs (fA + fB + (fC * INVMAGIC)) + (sfabs (fA) + sfabs (fB));
/*
// Once again, this float value is always positive. On Pentium, the
// float compare is awful, so do it in INT instead.
*/
#if SLOW_FCMP && !MULTI_FP_REG
if (FLOAT_TO_LONG(fMaxVal) > LONG_INVMAGIC )
#else
if (fMaxVal > INVMAGIC)
#endif
{
IW( pPlaneMem, 0+0, 0);
IW( pPlaneMem, 1+0, 0);
FW( pPlaneMem, 2+0, fZ2);
}
/* Pack instruction and TSP address.
*/
IW( pPlaneMem, 3+0, (forw_visib_fp | (TSPAddr << 4)));
/* do first edge plane
*/
fA = pTri->fAdjoint[0][2];
fB = pTri->fAdjoint[1][2];
fC = pTri->fAdjoint[2][2];
FW( pPlaneMem, 0+4, fA);
FW( pPlaneMem, 1+4, fB);
FW( pPlaneMem, 2+4, fC);
IW( pPlaneMem, 3+4, (forw_perp | (DEBUG_PERP_TAG << 4)));
/* do second edge plane
*/
fA = pTri->fAdjoint[0][0];
fB = pTri->fAdjoint[1][0];
fC = pTri->fAdjoint[2][0];
FW( pPlaneMem, 0+8, fA);
FW( pPlaneMem, 1+8, fB);
FW( pPlaneMem, 2+8, fC);
IW( pPlaneMem, 3+8, (forw_perp | (DEBUG_PERP_TAG << 4)));
/* do third edge plane
*/
fA = pTri->fAdjoint[0][1];
fB = pTri->fAdjoint[1][1];
fC = pTri->fAdjoint[2][1];
FW( pPlaneMem, 0+12, fA);
FW( pPlaneMem, 1+12, fB);
FW( pPlaneMem, 2+12, fC);
IW( pPlaneMem, 3+12, (forw_perp | (DEBUG_PERP_TAG << 4)));
pPlaneMem += WORDS_PER_PLANE * 4;
*((sgl_uint32 *) &pTri) += nIncrement;
TSPAddr += TSPIncrement;
}
*rpTri = pTri;
*rTSPAddr = TSPAddr;
}
/**********************************************************************
**
** Hardware: PCX2, PCX2_003
** No assembler equivalent
** Packs the front face and three edges
** This routine is similar to the above except it also checks the Edges to make
** sure they don't go out of range
**
**********************************************************************/
void PackISPCoreClipTested(sgl_uint32 *pPlaneMem, sgl_uint32 nPolysInChunk, sgl_uint32 *rTSPAddr, PITRI *rpTri,
int nIncrement, sgl_uint32 TSPIncrement)
{
float fA, fB, fC;
float fZ0, fZ1, fZ2;
float f1OverDet;
float fMaxVal;
PITRI pTri = *rpTri;
sgl_uint32 TSPAddr = *rTSPAddr;
while (nPolysInChunk--)
{
fZ2 = pTri->fZ[2];
f1OverDet = pTri->f1OverDet;
fZ0 = (pTri->fZ[0] - fZ2) * f1OverDet;
fZ1 = (pTri->fZ[1] - fZ2) * f1OverDet;
/*
// Compute initial A, B and C values before scaling, using the projected
// vertex positions:
//
// NOTE: I have changed the calculation to use a different basis system
// which means that the new "adjoint" which actually re-uses values from
// the adjoint we pass in except for the last row which is 0,0, Det.
// Simon.
*/
fA = fZ0*pTri->fAdjoint[0][0] + fZ1*pTri->fAdjoint[0][1];
fB = fZ0*pTri->fAdjoint[1][0] + fZ1*pTri->fAdjoint[1][1];
fC = fZ0*pTri->fAdjoint[2][0] + fZ1*pTri->fAdjoint[2][1] + fZ2 + gfDepthBias;
FW( pPlaneMem, 0+0, fA);
FW( pPlaneMem, 1+0, fB);
/* Check C value. PCX2 hardware has problems converting a number in
* the range -9.313e-10 to 0. If C is in this range set to 0.
*/
#if PCX2
if (sfabs(fC) <= MAGIC_FLOAT_FIX)
/* If problem set C to 0.
*/
IW( pPlaneMem, 2+0, 0);
else
#endif
/* Continue as normal.
*/
FW( pPlaneMem, 2+0, fC);
/*
// to prevent the 20Bit SFLOATS from overflowing (currently they are 2x
// larger than I originally thought necessary)
*/
fMaxVal = sfabs (fA + fB + (fC * INVMAGIC)) + (sfabs (fA) + sfabs (fB));
/*
// Once again, this float value is always positive. On Pentium, the
// float compare is awful, so do it in INT instead.
*/
#if SLOW_FCMP && !MULTI_FP_REG
if (FLOAT_TO_LONG(fMaxVal) > LONG_INVMAGIC )
#else
if (fMaxVal > INVMAGIC)
#endif
{
IW( pPlaneMem, 0+0, 0);
IW( pPlaneMem, 1+0, 0);
FW( pPlaneMem, 2+0, fZ2);