Source

quake / QW / client / d_draw16.s

Full commit
  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
/*
Copyright (C) 1996-1997 Id Software, Inc.

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  

See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.

*/
//
// d_draw16.s
// x86 assembly-language horizontal 8-bpp span-drawing code, with 16-pixel
// subdivision.
//

#include "asm_i386.h"
#include "quakeasm.h"
#include "asm_draw.h"
#include "d_ifacea.h"

#if	id386

//----------------------------------------------------------------------
// 8-bpp horizontal span drawing code for polygons, with no transparency and
// 16-pixel subdivision.
//
// Assumes there is at least one span in pspans, and that every span
// contains at least one pixel
//----------------------------------------------------------------------

	.data

	.text

// out-of-line, rarely-needed clamping code

LClampHigh0:
	movl	C(bbextents),%esi
	jmp		LClampReentry0
LClampHighOrLow0:
	jg		LClampHigh0
	xorl	%esi,%esi
	jmp		LClampReentry0

LClampHigh1:
	movl	C(bbextentt),%edx
	jmp		LClampReentry1
LClampHighOrLow1:
	jg		LClampHigh1
	xorl	%edx,%edx
	jmp		LClampReentry1

LClampLow2:
	movl	$4096,%ebp
	jmp		LClampReentry2
LClampHigh2:
	movl	C(bbextents),%ebp
	jmp		LClampReentry2

LClampLow3:
	movl	$4096,%ecx
	jmp		LClampReentry3
LClampHigh3:
	movl	C(bbextentt),%ecx
	jmp		LClampReentry3

LClampLow4:
	movl	$4096,%eax
	jmp		LClampReentry4
LClampHigh4:
	movl	C(bbextents),%eax
	jmp		LClampReentry4

LClampLow5:
	movl	$4096,%ebx
	jmp		LClampReentry5
LClampHigh5:
	movl	C(bbextentt),%ebx
	jmp		LClampReentry5


#define pspans	4+16

	.align 4
.globl C(D_DrawSpans16)
C(D_DrawSpans16):
	pushl	%ebp				// preserve caller's stack frame
	pushl	%edi
	pushl	%esi				// preserve register variables
	pushl	%ebx

//
// set up scaled-by-16 steps, for 16-long segments; also set up cacheblock
// and span list pointers
//
// TODO: any overlap from rearranging?
	flds	C(d_sdivzstepu)
	fmuls	fp_16
	movl	C(cacheblock),%edx
	flds	C(d_tdivzstepu)
	fmuls	fp_16
	movl	pspans(%esp),%ebx	// point to the first span descriptor
	flds	C(d_zistepu)
	fmuls	fp_16
	movl	%edx,pbase			// pbase = cacheblock
	fstps	zi16stepu
	fstps	tdivz16stepu
	fstps	sdivz16stepu

LSpanLoop:
//
// set up the initial s/z, t/z, and 1/z on the FP stack, and generate the
// initial s and t values
//
// FIXME: pipeline FILD?
	fildl	espan_t_v(%ebx)
	fildl	espan_t_u(%ebx)

	fld		%st(1)			// dv | du | dv
	fmuls	C(d_sdivzstepv)	// dv*d_sdivzstepv | du | dv
	fld		%st(1)			// du | dv*d_sdivzstepv | du | dv
	fmuls	C(d_sdivzstepu)	// du*d_sdivzstepu | dv*d_sdivzstepv | du | dv
	fld		%st(2)			// du | du*d_sdivzstepu | dv*d_sdivzstepv | du | dv
	fmuls	C(d_tdivzstepu)	// du*d_tdivzstepu | du*d_sdivzstepu |
							//  dv*d_sdivzstepv | du | dv
	fxch	%st(1)			// du*d_sdivzstepu | du*d_tdivzstepu |
							//  dv*d_sdivzstepv | du | dv
	faddp	%st(0),%st(2)	// du*d_tdivzstepu |
							//  du*d_sdivzstepu + dv*d_sdivzstepv | du | dv
	fxch	%st(1)			// du*d_sdivzstepu + dv*d_sdivzstepv |
							//  du*d_tdivzstepu | du | dv
	fld		%st(3)			// dv | du*d_sdivzstepu + dv*d_sdivzstepv |
							//  du*d_tdivzstepu | du | dv
	fmuls	C(d_tdivzstepv)	// dv*d_tdivzstepv |
							//  du*d_sdivzstepu + dv*d_sdivzstepv |
							//  du*d_tdivzstepu | du | dv
	fxch	%st(1)			// du*d_sdivzstepu + dv*d_sdivzstepv |
							//  dv*d_tdivzstepv | du*d_tdivzstepu | du | dv
	fadds	C(d_sdivzorigin)	// sdivz = d_sdivzorigin + dv*d_sdivzstepv +
							//  du*d_sdivzstepu; stays in %st(2) at end
	fxch	%st(4)			// dv | dv*d_tdivzstepv | du*d_tdivzstepu | du |
							//  s/z
	fmuls	C(d_zistepv)		// dv*d_zistepv | dv*d_tdivzstepv |
							//  du*d_tdivzstepu | du | s/z
	fxch	%st(1)			// dv*d_tdivzstepv |  dv*d_zistepv |
							//  du*d_tdivzstepu | du | s/z
	faddp	%st(0),%st(2)	// dv*d_zistepv |
							//  dv*d_tdivzstepv + du*d_tdivzstepu | du | s/z
	fxch	%st(2)			// du | dv*d_tdivzstepv + du*d_tdivzstepu |
							//  dv*d_zistepv | s/z
	fmuls	C(d_zistepu)		// du*d_zistepu |
							//  dv*d_tdivzstepv + du*d_tdivzstepu |
							//  dv*d_zistepv | s/z
	fxch	%st(1)			// dv*d_tdivzstepv + du*d_tdivzstepu |
							//  du*d_zistepu | dv*d_zistepv | s/z
	fadds	C(d_tdivzorigin)	// tdivz = d_tdivzorigin + dv*d_tdivzstepv +
							//  du*d_tdivzstepu; stays in %st(1) at end
	fxch	%st(2)			// dv*d_zistepv | du*d_zistepu | t/z | s/z
	faddp	%st(0),%st(1)	// dv*d_zistepv + du*d_zistepu | t/z | s/z

	flds	fp_64k			// fp_64k | dv*d_zistepv + du*d_zistepu | t/z | s/z
	fxch	%st(1)			// dv*d_zistepv + du*d_zistepu | fp_64k | t/z | s/z
	fadds	C(d_ziorigin)		// zi = d_ziorigin + dv*d_zistepv +
							//  du*d_zistepu; stays in %st(0) at end
							// 1/z | fp_64k | t/z | s/z
//
// calculate and clamp s & t
//
	fdivr	%st(0),%st(1)	// 1/z | z*64k | t/z | s/z

//
// point %edi to the first pixel in the span
//
	movl	C(d_viewbuffer),%ecx
	movl	espan_t_v(%ebx),%eax
	movl	%ebx,pspantemp	// preserve spans pointer

	movl	C(tadjust),%edx
	movl	C(sadjust),%esi
	movl	C(d_scantable)(,%eax,4),%edi	// v * screenwidth
	addl	%ecx,%edi
	movl	espan_t_u(%ebx),%ecx
	addl	%ecx,%edi				// pdest = &pdestspan[scans->u];
	movl	espan_t_count(%ebx),%ecx

//
// now start the FDIV for the end of the span
//
	cmpl	$16,%ecx
	ja		LSetupNotLast1

	decl	%ecx
	jz		LCleanup1		// if only one pixel, no need to start an FDIV
	movl	%ecx,spancountminus1

// finish up the s and t calcs
	fxch	%st(1)			// z*64k | 1/z | t/z | s/z

	fld		%st(0)			// z*64k | z*64k | 1/z | t/z | s/z
	fmul	%st(4),%st(0)	// s | z*64k | 1/z | t/z | s/z
	fxch	%st(1)			// z*64k | s | 1/z | t/z | s/z
	fmul	%st(3),%st(0)	// t | s | 1/z | t/z | s/z
	fxch	%st(1)			// s | t | 1/z | t/z | s/z
	fistpl	s				// 1/z | t | t/z | s/z
	fistpl	t				// 1/z | t/z | s/z

	fildl	spancountminus1

	flds	C(d_tdivzstepu)	// C(d_tdivzstepu) | spancountminus1
	flds	C(d_zistepu)		// C(d_zistepu) | C(d_tdivzstepu) | spancountminus1
	fmul	%st(2),%st(0)	// C(d_zistepu)*scm1 | C(d_tdivzstepu) | scm1
	fxch	%st(1)			// C(d_tdivzstepu) | C(d_zistepu)*scm1 | scm1
	fmul	%st(2),%st(0)	// C(d_tdivzstepu)*scm1 | C(d_zistepu)*scm1 | scm1
	fxch	%st(2)			// scm1 | C(d_zistepu)*scm1 | C(d_tdivzstepu)*scm1
	fmuls	C(d_sdivzstepu)	// C(d_sdivzstepu)*scm1 | C(d_zistepu)*scm1 |
							//  C(d_tdivzstepu)*scm1
	fxch	%st(1)			// C(d_zistepu)*scm1 | C(d_sdivzstepu)*scm1 |
							//  C(d_tdivzstepu)*scm1
	faddp	%st(0),%st(3)	// C(d_sdivzstepu)*scm1 | C(d_tdivzstepu)*scm1
	fxch	%st(1)			// C(d_tdivzstepu)*scm1 | C(d_sdivzstepu)*scm1
	faddp	%st(0),%st(3)	// C(d_sdivzstepu)*scm1
	faddp	%st(0),%st(3)

	flds	fp_64k
	fdiv	%st(1),%st(0)	// this is what we've gone to all this trouble to
							//  overlap
	jmp		LFDIVInFlight1

LCleanup1:
// finish up the s and t calcs
	fxch	%st(1)			// z*64k | 1/z | t/z | s/z

	fld		%st(0)			// z*64k | z*64k | 1/z | t/z | s/z
	fmul	%st(4),%st(0)	// s | z*64k | 1/z | t/z | s/z
	fxch	%st(1)			// z*64k | s | 1/z | t/z | s/z
	fmul	%st(3),%st(0)	// t | s | 1/z | t/z | s/z
	fxch	%st(1)			// s | t | 1/z | t/z | s/z
	fistpl	s				// 1/z | t | t/z | s/z
	fistpl	t				// 1/z | t/z | s/z
	jmp		LFDIVInFlight1

	.align	4
LSetupNotLast1:
// finish up the s and t calcs
	fxch	%st(1)			// z*64k | 1/z | t/z | s/z

	fld		%st(0)			// z*64k | z*64k | 1/z | t/z | s/z
	fmul	%st(4),%st(0)	// s | z*64k | 1/z | t/z | s/z
	fxch	%st(1)			// z*64k | s | 1/z | t/z | s/z
	fmul	%st(3),%st(0)	// t | s | 1/z | t/z | s/z
	fxch	%st(1)			// s | t | 1/z | t/z | s/z
	fistpl	s				// 1/z | t | t/z | s/z
	fistpl	t				// 1/z | t/z | s/z

	fadds	zi16stepu
	fxch	%st(2)
	fadds	sdivz16stepu
	fxch	%st(2)
	flds	tdivz16stepu
	faddp	%st(0),%st(2)
	flds	fp_64k
	fdiv	%st(1),%st(0)	// z = 1/1/z
							// this is what we've gone to all this trouble to
							//  overlap
LFDIVInFlight1:

	addl	s,%esi
	addl	t,%edx
	movl	C(bbextents),%ebx
	movl	C(bbextentt),%ebp
	cmpl	%ebx,%esi
	ja		LClampHighOrLow0
LClampReentry0:
	movl	%esi,s
	movl	pbase,%ebx
	shll	$16,%esi
	cmpl	%ebp,%edx
	movl	%esi,sfracf
	ja		LClampHighOrLow1
LClampReentry1:
	movl	%edx,t
	movl	s,%esi					// sfrac = scans->sfrac;
	shll	$16,%edx
	movl	t,%eax					// tfrac = scans->tfrac;
	sarl	$16,%esi
	movl	%edx,tfracf

//
// calculate the texture starting address
//
	sarl	$16,%eax
	movl	C(cachewidth),%edx
	imull	%edx,%eax				// (tfrac >> 16) * cachewidth
	addl	%ebx,%esi
	addl	%eax,%esi				// psource = pbase + (sfrac >> 16) +
									//           ((tfrac >> 16) * cachewidth);
//
// determine whether last span or not
//
	cmpl	$16,%ecx
	jna		LLastSegment

//
// not the last segment; do full 16-wide segment
//
LNotLastSegment:

//
// advance s/z, t/z, and 1/z, and calculate s & t at end of span and steps to
// get there
//

// pick up after the FDIV that was left in flight previously

	fld		%st(0)			// duplicate it
	fmul	%st(4),%st(0)	// s = s/z * z
	fxch	%st(1)
	fmul	%st(3),%st(0)	// t = t/z * z
	fxch	%st(1)
	fistpl	snext
	fistpl	tnext
	movl	snext,%eax
	movl	tnext,%edx

	movb	(%esi),%bl	// get first source texel
	subl	$16,%ecx		// count off this segments' pixels
	movl	C(sadjust),%ebp
	movl	%ecx,counttemp	// remember count of remaining pixels

	movl	C(tadjust),%ecx
	movb	%bl,(%edi)	// store first dest pixel

	addl	%eax,%ebp
	addl	%edx,%ecx

	movl	C(bbextents),%eax
	movl	C(bbextentt),%edx

	cmpl	$4096,%ebp
	jl		LClampLow2
	cmpl	%eax,%ebp
	ja		LClampHigh2
LClampReentry2:

	cmpl	$4096,%ecx
	jl		LClampLow3
	cmpl	%edx,%ecx
	ja		LClampHigh3
LClampReentry3:

	movl	%ebp,snext
	movl	%ecx,tnext

	subl	s,%ebp
	subl	t,%ecx
	
//
// set up advancetable
//
	movl	%ecx,%eax
	movl	%ebp,%edx
	sarl	$20,%eax			// tstep >>= 16;
	jz		LZero
	sarl	$20,%edx			// sstep >>= 16;
	movl	C(cachewidth),%ebx
	imull	%ebx,%eax
	jmp		LSetUp1

LZero:
	sarl	$20,%edx			// sstep >>= 16;
	movl	C(cachewidth),%ebx

LSetUp1:

	addl	%edx,%eax			// add in sstep
								// (tstep >> 16) * cachewidth + (sstep >> 16);
	movl	tfracf,%edx
	movl	%eax,advancetable+4	// advance base in t
	addl	%ebx,%eax			// ((tstep >> 16) + 1) * cachewidth +
								//  (sstep >> 16);
	shll	$12,%ebp			// left-justify sstep fractional part
	movl	sfracf,%ebx
	shll	$12,%ecx			// left-justify tstep fractional part
	movl	%eax,advancetable	// advance extra in t

	movl	%ecx,tstep
	addl	%ecx,%edx			// advance tfrac fractional part by tstep frac

	sbbl	%ecx,%ecx			// turn tstep carry into -1 (0 if none)
	addl	%ebp,%ebx			// advance sfrac fractional part by sstep frac
	adcl	advancetable+4(,%ecx,4),%esi	// point to next source texel

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	(%esi),%al
	addl	%ebp,%ebx
	movb	%al,1(%edi)
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,2(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,3(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,4(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,5(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,6(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,7(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi


//
// start FDIV for end of next segment in flight, so it can overlap
//
	movl	counttemp,%ecx
	cmpl	$16,%ecx			// more than one segment after this?
	ja		LSetupNotLast2	// yes

	decl	%ecx
	jz		LFDIVInFlight2	// if only one pixel, no need to start an FDIV
	movl	%ecx,spancountminus1
	fildl	spancountminus1

	flds	C(d_zistepu)		// C(d_zistepu) | spancountminus1
	fmul	%st(1),%st(0)	// C(d_zistepu)*scm1 | scm1
	flds	C(d_tdivzstepu)	// C(d_tdivzstepu) | C(d_zistepu)*scm1 | scm1
	fmul	%st(2),%st(0)	// C(d_tdivzstepu)*scm1 | C(d_zistepu)*scm1 | scm1
	fxch	%st(1)			// C(d_zistepu)*scm1 | C(d_tdivzstepu)*scm1 | scm1
	faddp	%st(0),%st(3)	// C(d_tdivzstepu)*scm1 | scm1
	fxch	%st(1)			// scm1 | C(d_tdivzstepu)*scm1
	fmuls	C(d_sdivzstepu)	// C(d_sdivzstepu)*scm1 | C(d_tdivzstepu)*scm1
	fxch	%st(1)			// C(d_tdivzstepu)*scm1 | C(d_sdivzstepu)*scm1
	faddp	%st(0),%st(3)	// C(d_sdivzstepu)*scm1
	flds	fp_64k			// 64k | C(d_sdivzstepu)*scm1
	fxch	%st(1)			// C(d_sdivzstepu)*scm1 | 64k
	faddp	%st(0),%st(4)	// 64k

	fdiv	%st(1),%st(0)	// this is what we've gone to all this trouble to
							//  overlap
	jmp		LFDIVInFlight2

	.align	4
LSetupNotLast2:
	fadds	zi16stepu
	fxch	%st(2)
	fadds	sdivz16stepu
	fxch	%st(2)
	flds	tdivz16stepu
	faddp	%st(0),%st(2)
	flds	fp_64k
	fdiv	%st(1),%st(0)	// z = 1/1/z
							// this is what we've gone to all this trouble to
							//  overlap
LFDIVInFlight2:
	movl	%ecx,counttemp

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,8(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,9(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,10(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,11(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,12(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,13(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,14(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi

	addl	$16,%edi
	movl	%edx,tfracf
	movl	snext,%edx
	movl	%ebx,sfracf
	movl	tnext,%ebx
	movl	%edx,s
	movl	%ebx,t

	movl	counttemp,%ecx		// retrieve count

//
// determine whether last span or not
//
	cmpl	$16,%ecx				// are there multiple segments remaining?
	movb	%al,-1(%edi)
	ja		LNotLastSegment		// yes

//
// last segment of scan
//
LLastSegment:

//
// advance s/z, t/z, and 1/z, and calculate s & t at end of span and steps to
// get there. The number of pixels left is variable, and we want to land on the
// last pixel, not step one past it, so we can't run into arithmetic problems
//
	testl	%ecx,%ecx
	jz		LNoSteps		// just draw the last pixel and we're done

// pick up after the FDIV that was left in flight previously


	fld		%st(0)			// duplicate it
	fmul	%st(4),%st(0)	// s = s/z * z
	fxch	%st(1)
	fmul	%st(3),%st(0)	// t = t/z * z
	fxch	%st(1)
	fistpl	snext
	fistpl	tnext

	movb	(%esi),%al		// load first texel in segment
	movl	C(tadjust),%ebx
	movb	%al,(%edi)		// store first pixel in segment
	movl	C(sadjust),%eax

	addl	snext,%eax
	addl	tnext,%ebx

	movl	C(bbextents),%ebp
	movl	C(bbextentt),%edx

	cmpl	$4096,%eax
	jl		LClampLow4
	cmpl	%ebp,%eax
	ja		LClampHigh4
LClampReentry4:
	movl	%eax,snext

	cmpl	$4096,%ebx
	jl		LClampLow5
	cmpl	%edx,%ebx
	ja		LClampHigh5
LClampReentry5:

	cmpl	$1,%ecx			// don't bother 
	je		LOnlyOneStep	// if two pixels in segment, there's only one step,
							//  of the segment length
	subl	s,%eax
	subl	t,%ebx

	addl	%eax,%eax		// convert to 15.17 format so multiply by 1.31
	addl	%ebx,%ebx		//  reciprocal yields 16.48

	imull	reciprocal_table_16-8(,%ecx,4)	// sstep = (snext - s) /
											//  (spancount-1)
	movl	%edx,%ebp

	movl	%ebx,%eax
	imull	reciprocal_table_16-8(,%ecx,4)	// tstep = (tnext - t) /
											//  (spancount-1)
LSetEntryvec:
//
// set up advancetable
//
	movl	entryvec_table_16(,%ecx,4),%ebx
	movl	%edx,%eax
	movl	%ebx,jumptemp		// entry point into code for RET later
	movl	%ebp,%ecx
	sarl	$16,%edx			// tstep >>= 16;
	movl	C(cachewidth),%ebx
	sarl	$16,%ecx			// sstep >>= 16;
	imull	%ebx,%edx

	addl	%ecx,%edx			// add in sstep
								// (tstep >> 16) * cachewidth + (sstep >> 16);
	movl	tfracf,%ecx
	movl	%edx,advancetable+4	// advance base in t
	addl	%ebx,%edx			// ((tstep >> 16) + 1) * cachewidth +
								//  (sstep >> 16);
	shll	$16,%ebp			// left-justify sstep fractional part
	movl	sfracf,%ebx
	shll	$16,%eax			// left-justify tstep fractional part
	movl	%edx,advancetable	// advance extra in t

	movl	%eax,tstep
	movl	%ecx,%edx
	addl	%eax,%edx
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi

	jmp		*jumptemp			// jump to the number-of-pixels handler

//----------------------------------------

LNoSteps:
	movb	(%esi),%al		// load first texel in segment
	subl	$15,%edi			// adjust for hardwired offset
	jmp		LEndSpan


LOnlyOneStep:
	subl	s,%eax
	subl	t,%ebx
	movl	%eax,%ebp
	movl	%ebx,%edx
	jmp		LSetEntryvec

//----------------------------------------

.globl	Entry2_16, Entry3_16, Entry4_16, Entry5_16
.globl	Entry6_16, Entry7_16, Entry8_16, Entry9_16
.globl	Entry10_16, Entry11_16, Entry12_16, Entry13_16
.globl	Entry14_16, Entry15_16, Entry16_16

Entry2_16:
	subl	$14,%edi		// adjust for hardwired offsets
	movb	(%esi),%al
	jmp		LEntry2_16

//----------------------------------------

Entry3_16:
	subl	$13,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	jmp		LEntry3_16

//----------------------------------------

Entry4_16:
	subl	$12,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry4_16

//----------------------------------------

Entry5_16:
	subl	$11,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry5_16

//----------------------------------------

Entry6_16:
	subl	$10,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry6_16

//----------------------------------------

Entry7_16:
	subl	$9,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry7_16

//----------------------------------------

Entry8_16:
	subl	$8,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry8_16

//----------------------------------------

Entry9_16:
	subl	$7,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry9_16

//----------------------------------------

Entry10_16:
	subl	$6,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry10_16

//----------------------------------------

Entry11_16:
	subl	$5,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry11_16

//----------------------------------------

Entry12_16:
	subl	$4,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry12_16

//----------------------------------------

Entry13_16:
	subl	$3,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry13_16

//----------------------------------------

Entry14_16:
	subl	$2,%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry14_16

//----------------------------------------

Entry15_16:
	decl	%edi		// adjust for hardwired offsets
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
	jmp		LEntry15_16

//----------------------------------------

Entry16_16:
	addl	%eax,%edx
	movb	(%esi),%al
	sbbl	%ecx,%ecx
	addl	%ebp,%ebx
	adcl	advancetable+4(,%ecx,4),%esi

	addl	tstep,%edx
	sbbl	%ecx,%ecx
	movb	%al,1(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry15_16:
	sbbl	%ecx,%ecx
	movb	%al,2(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry14_16:
	sbbl	%ecx,%ecx
	movb	%al,3(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry13_16:
	sbbl	%ecx,%ecx
	movb	%al,4(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry12_16:
	sbbl	%ecx,%ecx
	movb	%al,5(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry11_16:
	sbbl	%ecx,%ecx
	movb	%al,6(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry10_16:
	sbbl	%ecx,%ecx
	movb	%al,7(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry9_16:
	sbbl	%ecx,%ecx
	movb	%al,8(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry8_16:
	sbbl	%ecx,%ecx
	movb	%al,9(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry7_16:
	sbbl	%ecx,%ecx
	movb	%al,10(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry6_16:
	sbbl	%ecx,%ecx
	movb	%al,11(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry5_16:
	sbbl	%ecx,%ecx
	movb	%al,12(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
	addl	tstep,%edx
LEntry4_16:
	sbbl	%ecx,%ecx
	movb	%al,13(%edi)
	addl	%ebp,%ebx
	movb	(%esi),%al
	adcl	advancetable+4(,%ecx,4),%esi
LEntry3_16:
	movb	%al,14(%edi)
	movb	(%esi),%al
LEntry2_16:

LEndSpan:

//
// clear s/z, t/z, 1/z from FP stack
//
	fstp %st(0)
	fstp %st(0)
	fstp %st(0)

	movl	pspantemp,%ebx				// restore spans pointer
	movl	espan_t_pnext(%ebx),%ebx	// point to next span
	testl	%ebx,%ebx			// any more spans?
	movb	%al,15(%edi)
	jnz		LSpanLoop			// more spans

	popl	%ebx				// restore register variables
	popl	%esi
	popl	%edi
	popl	%ebp				// restore the caller's stack frame
	ret

#endif	// id386