MAME SVN History

199869 Revisions

r32870 Wednesday 22nd October, 2014 at 13:09:22 UTC by Robbbert
These line endings will be the death of me.

[src/mame/video]

n64.c

trunk/src/mame/video/n64.c
r241381	r241382
1		/******************************************************************************
2
3
4		SGI/Nintendo Reality Display Processor
5		-------------------
6
7		by MooglyGuy
8		based on initial C code by Ville Linde
9		contains additional improvements from angrylion, Ziggy, Gonetz and Orkin
10
11
12		*******************************************************************************
13
14		STATUS:
15
16		Much behavior needs verification against real hardware. Many edge cases must
17		be verified on real hardware as well.
18
19		TODO:
20
21		- Further re-work class structure to avoid dependencies
22
23		*******************************************************************************/
24
25		#include "emu.h"
26		//#include "includes/n64.h"
27		#include "video/n64.h"
28
29		#define LOG_RDP_EXECUTION 0
30
31		static FILE *rdp_exec;
32
33		UINT32 n64_rdp::s_special_9bit_clamptable[512];
34
35		bool n64_rdp::rdp_range_check(UINT32 addr)
36		{
37		if(MiscState.FBSize == 0) return false;
38
39		int fbcount = ((MiscState.FBWidth * Scissor.m_yl) << (MiscState.FBSize - 1)) * 3;
40		int fbaddr = MiscState.FBAddress & 0x007fffff;
41		if ((addr >= fbaddr) && (addr < (fbaddr + fbcount)))
42		{
43		return false;
44		}
45
46		int zbcount = MiscState.FBWidth * Scissor.m_yl * 2;
47		int zbaddr = MiscState.ZBAddress & 0x007fffff;
48		if ((addr >= zbaddr) && (addr < (zbaddr + zbcount)))
49		{
50		return false;
51		}
52
53		printf("Check failed: %08x vs. %08x-%08x, %08x-%08x (%d, %d)\n", addr, fbaddr, fbaddr + fbcount, zbaddr, zbaddr + zbcount, MiscState.FBWidth, Scissor.m_yl);
54		fflush(stdout);
55		return true;
56		}
57
58		/*****************************************************************************/
59
60		// The functions in this file should be moved into the parent Processor class.
61		#include "rdpfiltr.inc"
62
63		void n64_rdp::GetAlphaCvg(UINT8 comb_alpha, rdp_span_aux userdata, const rdp_poly_state &object)
64		{
65		INT32 temp = *comb_alpha;
66		INT32 temp2 = userdata->CurrentPixCvg;
67		INT32 temp3 = 0;
68
69		if (object.OtherModes.cvg_times_alpha)
70		{
71		temp3 = (temp * temp2) + 4;
72		userdata->CurrentPixCvg = (temp3 >> 8) & 0xf;
73		}
74		if (object.OtherModes.alpha_cvg_select)
75		{
76		temp = (OtherModes.cvg_times_alpha) ? (temp3 >> 3) : (temp2 << 5);
77		}
78		if (temp > 0xff)
79		{
80		temp = 0xff;
81		}
82		*comb_alpha = temp;
83		}
84
85		/*****************************************************************************/
86
87		void n64_rdp::VideoUpdate(n64_periphs *n64, bitmap_rgb32 &bitmap)
88		{
89		switch(n64->vi_control & 0x3)
90		{
91		case PIXEL_SIZE_16BIT:
92		VideoUpdate16(n64, bitmap);
93		break;
94
95		case PIXEL_SIZE_32BIT:
96		VideoUpdate32(n64, bitmap);
97		break;
98
99		default:
100		//fatalerror("Unsupported framebuffer depth: m_fb_size=%d\n", MiscState.FBSize);
101		break;
102		}
103		}
104
105		void n64_rdp::VideoUpdate16(n64_periphs *n64, bitmap_rgb32 &bitmap)
106		{
107		//int fsaa = (((n64->vi_control >> 8) & 3) < 2);
108		//int divot = (n64->vi_control >> 4) & 1;
109
110		//UINT32 prev_cvg = 0;
111		//UINT32 next_cvg = 0;
112		//int dither_filter = (n64->vi_control >> 16) & 1;
113		//int vibuffering = ((n64->vi_control & 2) && fsaa && divot);
114
115		UINT16 frame_buffer = (UINT16)&rdram[(n64->vi_origin & 0xffffff) >> 2];
116		//UINT32 hb = ((n64->vi_origin & 0xffffff) >> 2) >> 1;
117		//UINT8* hidden_buffer = &HiddenBits[hb];
118
119		INT32 hdiff = (n64->vi_hstart & 0x3ff) - ((n64->vi_hstart >> 16) & 0x3ff);
120		float hcoeff = ((float)(n64->vi_xscale & 0xfff) / (1 << 10));
121		UINT32 hres = ((float)hdiff * hcoeff);
122		INT32 invisiblewidth = n64->vi_width - hres;
123
124		INT32 vdiff = ((n64->vi_vstart & 0x3ff) - ((n64->vi_vstart >> 16) & 0x3ff)) >> 1;
125		float vcoeff = ((float)(n64->vi_yscale & 0xfff) / (1 << 10));
126		UINT32 vres = ((float)vdiff * vcoeff);
127
128		if (vdiff <= 0 \|\| hdiff <= 0)
129		{
130		return;
131		}
132
133		if (hres > 640) // Needed by Top Gear Overdrive (E)
134		{
135		invisiblewidth += (hres - 640);
136		hres = 640;
137		}
138
139		if (vres > bitmap.height()) // makes Perfect Dark boot w/o crashing
140		{
141		vres = bitmap.height();
142		}
143
144		UINT32 pixels = 0;
145
146		if (frame_buffer)
147		{
148		for(int j = 0; j < vres; j++)
149		{
150		UINT32 *d = &bitmap.pix32(j);
151
152		for(int i = 0; i < hres; i++)
153		{
154		Color c;
155		//int r, g, b;
156
157		UINT16 pix = frame_buffer[pixels ^ WORD_ADDR_XOR];
158		//MiscState.CurrentPixCvg = ((pix & 1) << 2) \| (hidden_buffer[pixels ^ BYTE_ADDR_XOR] & 3);
159
160		//if(divot)
161		//{
162		// if(i > 0 && i < (hres - 1))
163		// {
164		// prev_cvg = ((frame_buffer[(pixels - 1)^WORD_ADDR_XOR] & 1) << 2) \| (hidden_buffer[(pixels - 1)^BYTE_ADDR_XOR] & 3);
165		// next_cvg = ((frame_buffer[(pixels + 1)^WORD_ADDR_XOR] & 1) << 2) \| (hidden_buffer[(pixels + 1)^BYTE_ADDR_XOR] & 3);
166		// }
167		//}
168		c.i.r = ((pix >> 8) & 0xf8) \| (pix >> 13);
169		c.i.g = ((pix >> 3) & 0xf8) \| ((pix >> 8) & 0x07);
170		c.i.b = ((pix << 2) & 0xf8) \| ((pix >> 3) & 0x07);
171
172		//if(fsaa)
173		//{
174		//if (/!vibuffering &&/ state->m_rdp.MiscState.CurrentPixCvg < 7 && i > 1 && j > 1 && i < (hres - 2) && j < (vres - 2))
175		//{
176		//video_filter16(&c.i.r, &c.i.g, &c.i.b, &frame_buffer[pixels ^ WORD_ADDR_XOR],&hidden_buffer[pixels ^ BYTE_ADDR_XOR], n64->vi_width);
177		//}
178		//}
179		//else if (dither_filter && state->m_rdp.MiscState.CurrentPixCvg == 7 && i > 0 && j > 0 && i < (hres - 1) && j < (vres - 1))
180		//{
181		//if (vibuffering)
182		//{
183		// restore_filter16_buffer(&r, &g, &b, &ViBuffer[i][j], n64->vi_width);
184		//}
185		//else
186		//{
187		//restore_filter16(&c.i.r, &c.i.g, &c.i.b, &frame_buffer[pixels ^ WORD_ADDR_XOR], pixels ^ WORD_ADDR_XOR, n64->vi_width);
188		//}
189		//}
190		//if(divot)
191		//{
192		//if (i > 0 && i < (hres - 1) && (MiscState.CurrentPixCvg != 7 \|\| prev_cvg != 7 \|\| next_cvg != 7))
193		//{
194		//if (vibuffering)
195		//{
196		// divot_filter16_buffer(&r, &g, &b, &ViBuffer[i][j]);
197		//}
198		//else
199		//{
200		//divot_filter16(&c.i.r, &c.i.g, &c.i.b, &frame_buffer[pixels ^ WORD_ADDR_XOR], pixels ^ WORD_ADDR_XOR);
201		//}
202		//}
203		//}
204
205		/*
206		if (gamma_dither)
207		{
208		dith = screen.machine().rand() & 0x3f;
209		}
210		if (gamma)
211		{
212		if (gamma_dither)
213		{
214		r = m_gamma_dither_table[(r << 6)\|dith];
215		g = m_gamma_dither_table[(g << 6)\|dith];
216		b = m_gamma_dither_table[(b << 6)\|dith];
217		}
218		else
219		{
220		r = m_gamma_table[r];
221		g = m_gamma_table[g];
222		b = m_gamma_table[b];
223		}
224		}
225		else if (gamma_dither)
226		{
227		if (r < 255)
228		r += (dith & 1);
229		if (g < 255)
230		g += (dith & 1);
231		if (b < 255)
232		b += (dith & 1);
233		}
234		*/
235		pixels++;
236
237		d[i] = c.c >> 8;//(r << 16) \| (g << 8) \| b; // Fix me for endianness
238		}
239		pixels +=invisiblewidth;
240		}
241		}
242		}
243
244		void n64_rdp::VideoUpdate32(n64_periphs *n64, bitmap_rgb32 &bitmap)
245		{
246		int gamma = (n64->vi_control >> 3) & 1;
247		int gamma_dither = (n64->vi_control >> 2) & 1;
248		//int vibuffering = ((n64->vi_control & 2) && fsaa && divot);
249
250		UINT32 frame_buffer32 = (UINT32)&rdram[(n64->vi_origin & 0xffffff) >> 2];
251
252		const INT32 hdiff = (n64->vi_hstart & 0x3ff) - ((n64->vi_hstart >> 16) & 0x3ff);
253		const float hcoeff = ((float)(n64->vi_xscale & 0xfff) / (1 << 10));
254		UINT32 hres = ((float)hdiff * hcoeff);
255		INT32 invisiblewidth = n64->vi_width - hres;
256
257		const INT32 vdiff = ((n64->vi_vstart & 0x3ff) - ((n64->vi_vstart >> 16) & 0x3ff)) >> 1;
258		const float vcoeff = ((float)(n64->vi_yscale & 0xfff) / (1 << 10));
259		const UINT32 vres = ((float)vdiff * vcoeff);
260
261		if (vdiff <= 0 \|\| hdiff <= 0)
262		{
263		return;
264		}
265
266		if (hres > 640) // Needed by Top Gear Overdrive (E)
267		{
268		invisiblewidth += (hres - 640);
269		hres = 640;
270		}
271
272		if (frame_buffer32)
273		{
274		for (int j = 0; j < vres; j++)
275		{
276		UINT32 *d = &bitmap.pix32(j);
277		for (int i = 0; i < hres; i++)
278		{
279		UINT32 pix = *frame_buffer32++;
280		if (gamma \|\| gamma_dither)
281		{
282		int r = (pix >> 24) & 0xff;
283		int g = (pix >> 16) & 0xff;
284		int b = (pix >> 8) & 0xff;
285		int dith = 0;
286		if (gamma_dither)
287		{
288		dith = GetRandom() & 0x3f;
289		}
290		if (gamma)
291		{
292		if (gamma_dither)
293		{
294		r = m_gamma_dither_table[(r << 6)\| dith];
295		g = m_gamma_dither_table[(g << 6)\| dith];
296		b = m_gamma_dither_table[(b << 6)\| dith];
297		}
298		else
299		{
300		r = m_gamma_table[r];
301		g = m_gamma_table[g];
302		b = m_gamma_table[b];
303		}
304		}
305		else if (gamma_dither)
306		{
307		if (r < 255)
308		r += (dith & 1);
309		if (g < 255)
310		g += (dith & 1);
311		if (b < 255)
312		b += (dith & 1);
313		}
314		pix = (r << 24) \| (g << 16) \| (b << 8);
315		}
316
317
318		d[i] = (pix >> 8);
319		}
320		frame_buffer32 += invisiblewidth;
321		}
322		}
323		}
324
325		/*****************************************************************************/
326
327		void n64_rdp::TCDivNoPersp(INT32 ss, INT32 st, INT32 sw, INT32* sss, INT32* sst)
328		{
329		*sss = (SIGN16(ss)) & 0x1ffff;
330		*sst = (SIGN16(st)) & 0x1ffff;
331		}
332
333		void n64_rdp::TCDiv(INT32 ss, INT32 st, INT32 sw, INT32* sss, INT32* sst)
334		{
335		int w_carry = 0;
336		if ((sw & 0x8000) \|\| !(sw & 0x7fff))
337		{
338		w_carry = 1;
339		}
340
341		sw &= 0x7fff;
342
343		int shift;
344		for (shift = 1; shift <= 14 && !((sw << shift) & 0x8000); shift++);
345		shift -= 1;
346
347		int normout = (sw << shift) & 0x3fff;
348		int wnorm = (normout & 0xff) << 2;
349		normout >>= 8;
350
351		int temppoint = m_norm_point_rom[normout];
352		int tempslope = m_norm_slope_rom[normout];
353
354		int tlu_rcp = ((-(tempslope * wnorm)) >> 10) + temppoint;
355
356		int sprod = SIGN16(ss) * tlu_rcp;
357		int tprod = SIGN16(st) * tlu_rcp;
358		int tempmask = ((1 << (shift + 1)) - 1) << (29 - shift);
359		int shift_value = 13 - shift;
360
361		int outofbounds_s = sprod & tempmask;
362		int outofbounds_t = tprod & tempmask;
363		if (shift == 0xe)
364		{
365		*sss = sprod << 1;
366		*sst = tprod << 1;
367		}
368		else
369		{
370		*sss = sprod = (sprod >> shift_value);
371		*sst = tprod = (tprod >> shift_value);
372		}
373		//compute clamp flags
374		int under_s = 0;
375		int under_t = 0;
376		int over_s = 0;
377		int over_t = 0;
378
379		if (outofbounds_s != tempmask && outofbounds_s != 0)
380		{
381		if (sprod & (1 << 29))
382		{
383		under_s = 1;
384		}
385		else
386		{
387		over_s = 1;
388		}
389		}
390
391		if (outofbounds_t != tempmask && outofbounds_t != 0)
392		{
393		if (tprod & (1 << 29))
394		{
395		under_t = 1;
396		}
397		else
398		{
399		over_t = 1;
400		}
401		}
402
403		over_s \|= w_carry;
404		over_t \|= w_carry;
405
406		sss = (sss & 0x1ffff) \| (over_s << 18) \| (under_s << 17);
407		sst = (sst & 0x1ffff) \| (over_t << 18) \| (under_t << 17);
408		}
409
410		INT32 n64_rdp::ColorCombinerEquation(INT32 a, INT32 b, INT32 c, INT32 d)
411		{
412		a = KURT_AKELEY_SIGN9(a);
413		b = KURT_AKELEY_SIGN9(b);
414		c = SIGN9(c);
415		d = KURT_AKELEY_SIGN9(d);
416		a = (((a - b) * c) + (d << 8) + 0x80);
417		a = SIGN17(a) >> 8;
418		a = s_special_9bit_clamptable[a & 0x1ff];
419		return a;
420		}
421
422		INT32 n64_rdp::AlphaCombinerEquation(INT32 a, INT32 b, INT32 c, INT32 d)
423		{
424		a = KURT_AKELEY_SIGN9(a);
425		b = KURT_AKELEY_SIGN9(b);
426		c = SIGN9(c);
427		d = KURT_AKELEY_SIGN9(d);
428		a = (((a - b) * c) + (d << 8) + 0x80) >> 8;
429		a = SIGN9(a);
430		a = s_special_9bit_clamptable[a & 0x1ff];
431		return a;
432		}
433
434		void n64_rdp::SetSubAInputRGB(UINT8 input_r, UINT8 input_g, UINT8 *input_b, int code, rdp_span_aux userdata)
435		{
436		switch (code & 0xf)
437		{
438		case 0: input_r = &userdata->CombinedColor.i.r; input_g = &userdata->CombinedColor.i.g; *input_b = &userdata->CombinedColor.i.b; break;
439		case 1: input_r = &userdata->Texel0Color.i.r; input_g = &userdata->Texel0Color.i.g; *input_b = &userdata->Texel0Color.i.b; break;
440		case 2: input_r = &userdata->Texel1Color.i.r; input_g = &userdata->Texel1Color.i.g; *input_b = &userdata->Texel1Color.i.b; break;
441		case 3: input_r = &userdata->PrimColor.i.r; input_g = &userdata->PrimColor.i.g; *input_b = &userdata->PrimColor.i.b; break;
442		case 4: input_r = &userdata->ShadeColor.i.r; input_g = &userdata->ShadeColor.i.g; *input_b = &userdata->ShadeColor.i.b; break;
443		case 5: input_r = &userdata->EnvColor.i.r; input_g = &userdata->EnvColor.i.g; *input_b = &userdata->EnvColor.i.b; break;
444		case 6: input_r = &OneColor.i.r; input_g = &OneColor.i.g; *input_b = &OneColor.i.b; break;
445		case 7: input_r = &userdata->NoiseColor.i.r; input_g = &userdata->NoiseColor.i.g; *input_b = &userdata->NoiseColor.i.b; break;
446		case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
447		{
448		input_r = &ZeroColor.i.r; input_g = &ZeroColor.i.g; *input_b = &ZeroColor.i.b; break;
449		}
450		}
451		}
452
453		void n64_rdp::SetSubBInputRGB(UINT8 input_r, UINT8 input_g, UINT8 *input_b, int code, rdp_span_aux userdata)
454		{
455		switch (code & 0xf)
456		{
457		case 0: input_r = &userdata->CombinedColor.i.r; input_g = &userdata->CombinedColor.i.g; *input_b = &userdata->CombinedColor.i.b; break;
458		case 1: input_r = &userdata->Texel0Color.i.r; input_g = &userdata->Texel0Color.i.g; *input_b = &userdata->Texel0Color.i.b; break;
459		case 2: input_r = &userdata->Texel1Color.i.r; input_g = &userdata->Texel1Color.i.g; *input_b = &userdata->Texel1Color.i.b; break;
460		case 3: input_r = &userdata->PrimColor.i.r; input_g = &userdata->PrimColor.i.g; *input_b = &userdata->PrimColor.i.b; break;
461		case 4: input_r = &userdata->ShadeColor.i.r; input_g = &userdata->ShadeColor.i.g; *input_b = &userdata->ShadeColor.i.b; break;
462		case 5: input_r = &userdata->EnvColor.i.r; input_g = &userdata->EnvColor.i.g; *input_b = &userdata->EnvColor.i.b; break;
463		case 6: fatalerror("SET_SUBB_RGB_INPUT: key_center\n");
464		case 7: input_r = (UINT8)&m_k4; input_g = (UINT8)&m_k4; input_b = (UINT8)&m_k4; break;
465		case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
466		{
467		input_r = &ZeroColor.i.r; input_g = &ZeroColor.i.g; *input_b = &ZeroColor.i.b; break;
468		}
469		}
470		}
471
472		void n64_rdp::SetMulInputRGB(UINT8 input_r, UINT8 input_g, UINT8 *input_b, int code, rdp_span_aux userdata)
473		{
474		switch (code & 0x1f)
475		{
476		case 0: input_r = &userdata->CombinedColor.i.r; input_g = &userdata->CombinedColor.i.g; *input_b = &userdata->CombinedColor.i.b; break;
477		case 1: input_r = &userdata->Texel0Color.i.r; input_g = &userdata->Texel0Color.i.g; *input_b = &userdata->Texel0Color.i.b; break;
478		case 2: input_r = &userdata->Texel1Color.i.r; input_g = &userdata->Texel1Color.i.g; *input_b = &userdata->Texel1Color.i.b; break;
479		case 3: input_r = &userdata->PrimColor.i.r; input_g = &userdata->PrimColor.i.g; *input_b = &userdata->PrimColor.i.b; break;
480		case 4: input_r = &userdata->ShadeColor.i.r; input_g = &userdata->ShadeColor.i.g; *input_b = &userdata->ShadeColor.i.b; break;
481		case 5: input_r = &userdata->EnvColor.i.r; input_g = &userdata->EnvColor.i.g; *input_b = &userdata->EnvColor.i.b; break;
482		case 6: input_r = &userdata->KeyScale.i.r; input_g = &userdata->KeyScale.i.g; *input_b = &userdata->KeyScale.i.b; break;
483		case 7: input_r = &userdata->CombinedColor.i.a; input_g = &userdata->CombinedColor.i.a; *input_b = &userdata->CombinedColor.i.a; break;
484		case 8: input_r = &userdata->Texel0Color.i.a; input_g = &userdata->Texel0Color.i.a; *input_b = &userdata->Texel0Color.i.a; break;
485		case 9: input_r = &userdata->Texel1Color.i.a; input_g = &userdata->Texel1Color.i.a; *input_b = &userdata->Texel1Color.i.a; break;
486		case 10: input_r = &userdata->PrimColor.i.a; input_g = &userdata->PrimColor.i.a; *input_b = &userdata->PrimColor.i.a; break;
487		case 11: input_r = &userdata->ShadeColor.i.a; input_g = &userdata->ShadeColor.i.a; *input_b = &userdata->ShadeColor.i.a; break;
488		case 12: input_r = &userdata->EnvColor.i.a; input_g = &userdata->EnvColor.i.a; *input_b = &userdata->EnvColor.i.a; break;
489		case 13: input_r = &userdata->LODFraction; input_g = &userdata->LODFraction; *input_b = &userdata->LODFraction; break;
490		case 14: input_r = &userdata->PrimLODFraction; input_g = &userdata->PrimLODFraction; *input_b = &userdata->PrimLODFraction; break;
491		case 15: input_r = (UINT8)&m_k5; input_g = (UINT8)&m_k5; input_b = (UINT8)&m_k5; break;
492		case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
493		case 24: case 25: case 26: case 27: case 28: case 29: case 30: case 31:
494		{
495		input_r = &ZeroColor.i.r; input_g = &ZeroColor.i.g; *input_b = &ZeroColor.i.b; break;
496		}
497		}
498		}
499
500		void n64_rdp::SetAddInputRGB(UINT8 input_r, UINT8 input_g, UINT8 *input_b, int code, rdp_span_aux userdata)
501		{
502		switch (code & 0x7)
503		{
504		case 0: input_r = &userdata->CombinedColor.i.r; input_g = &userdata->CombinedColor.i.g; *input_b = &userdata->CombinedColor.i.b; break;
505		case 1: input_r = &userdata->Texel0Color.i.r; input_g = &userdata->Texel0Color.i.g; *input_b = &userdata->Texel0Color.i.b; break;
506		case 2: input_r = &userdata->Texel1Color.i.r; input_g = &userdata->Texel1Color.i.g; *input_b = &userdata->Texel1Color.i.b; break;
507		case 3: input_r = &userdata->PrimColor.i.r; input_g = &userdata->PrimColor.i.g; *input_b = &userdata->PrimColor.i.b; break;
508		case 4: input_r = &userdata->ShadeColor.i.r; input_g = &userdata->ShadeColor.i.g; *input_b = &userdata->ShadeColor.i.b; break;
509		case 5: input_r = &userdata->EnvColor.i.r; input_g = &userdata->EnvColor.i.g; *input_b = &userdata->EnvColor.i.b; break;
510		case 6: input_r = &OneColor.i.r; input_g = &OneColor.i.g; *input_b = &OneColor.i.b; break;
511		case 7: input_r = &ZeroColor.i.r; input_g = &ZeroColor.i.g; *input_b = &ZeroColor.i.b; break;
512		}
513		}
514
515		void n64_rdp::SetSubInputAlpha(UINT8 *input, int code, rdp_span_aux userdata)
516		{
517		switch (code & 0x7)
518		{
519		case 0: *input = &userdata->CombinedColor.i.a; break;
520		case 1: *input = &userdata->Texel0Color.i.a; break;
521		case 2: *input = &userdata->Texel1Color.i.a; break;
522		case 3: *input = &userdata->PrimColor.i.a; break;
523		case 4: *input = &userdata->ShadeColor.i.a; break;
524		case 5: *input = &userdata->EnvColor.i.a; break;
525		case 6: *input = &OneColor.i.a; break;
526		case 7: *input = &ZeroColor.i.a; break;
527		}
528		}
529
530		void n64_rdp::SetMulInputAlpha(UINT8 *input, int code, rdp_span_aux userdata)
531		{
532		switch (code & 0x7)
533		{
534		case 0: *input = &userdata->LODFraction; break;
535		case 1: *input = &userdata->Texel0Color.i.a; break;
536		case 2: *input = &userdata->Texel1Color.i.a; break;
537		case 3: *input = &userdata->PrimColor.i.a; break;
538		case 4: *input = &userdata->ShadeColor.i.a; break;
539		case 5: *input = &userdata->EnvColor.i.a; break;
540		case 6: *input = &userdata->PrimLODFraction; break;
541		case 7: *input = &ZeroColor.i.a; break;
542		}
543		}
544
545		void n64_rdp::SetBlenderInput(int cycle, int which, UINT8 input_r, UINT8 input_g, UINT8 input_b, UINT8 input_a, int a, int b, rdp_span_aux *userdata)
546		{
547		switch (a & 0x3)
548		{
549		case 0:
550		{
551		if (cycle == 0)
552		{
553		*input_r = &userdata->PixelColor.i.r;
554		*input_g = &userdata->PixelColor.i.g;
555		*input_b = &userdata->PixelColor.i.b;
556		}
557		else
558		{
559		*input_r = &userdata->BlendedPixelColor.i.r;
560		*input_g = &userdata->BlendedPixelColor.i.g;
561		*input_b = &userdata->BlendedPixelColor.i.b;
562		}
563		break;
564		}
565
566		case 1:
567		{
568		*input_r = &userdata->MemoryColor.i.r;
569		*input_g = &userdata->MemoryColor.i.g;
570		*input_b = &userdata->MemoryColor.i.b;
571		break;
572		}
573
574		case 2:
575		{
576		*input_r = &userdata->BlendColor.i.r;
577		*input_g = &userdata->BlendColor.i.g;
578		*input_b = &userdata->BlendColor.i.b;
579		break;
580		}
581
582		case 3:
583		{
584		*input_r = &userdata->FogColor.i.r;
585		*input_g = &userdata->FogColor.i.g;
586		*input_b = &userdata->FogColor.i.b;
587		break;
588		}
589		}
590
591		if (which == 0)
592		{
593		switch (b & 0x3)
594		{
595		case 0: *input_a = &userdata->PixelColor.i.a; break;
596		case 1: *input_a = &userdata->FogColor.i.a; break;
597		case 2: *input_a = &userdata->ShadeColor.i.a; break;
598		case 3: *input_a = &ZeroColor.i.a; break;
599		}
600		}
601		else
602		{
603		switch (b & 0x3)
604		{
605		case 0: *input_a = &userdata->InvPixelColor.i.a; break;
606		case 1: *input_a = &userdata->MemoryColor.i.a; break;
607		case 2: *input_a = &OneColor.i.a; break;
608		case 3: *input_a = &ZeroColor.i.a; break;
609		}
610		}
611		}
612
613		const UINT8 n64_rdp::s_bayer_matrix[16] =
614		{ /* Bayer matrix */
615		0, 4, 1, 5,
616		6, 2, 7, 3,
617		1, 5, 0, 4,
618		7, 3, 6, 2
619		};
620
621		const UINT8 n64_rdp::s_magic_matrix[16] =
622		{ /* Magic square matrix */
623		0, 6, 1, 7,
624		4, 2, 5, 3,
625		3, 5, 2, 4,
626		7, 1, 6, 0
627		};
628
629		const n64_rdp::ZDecompressEntry n64_rdp::z_dec_table[8] =
630		{
631		{ 6, 0x00000 },
632		{ 5, 0x20000 },
633		{ 4, 0x30000 },
634		{ 3, 0x38000 },
635		{ 2, 0x3c000 },
636		{ 1, 0x3e000 },
637		{ 0, 0x3f000 },
638		{ 0, 0x3f800 },
639		};
640
641		/*****************************************************************************/
642
643		void n64_rdp::z_build_com_table(void)
644		{
645		UINT16 altmem = 0;
646		for(int z = 0; z < 0x40000; z++)
647		{
648		switch((z >> 11) & 0x7f)
649		{
650		case 0x00:
651		case 0x01:
652		case 0x02:
653		case 0x03:
654		case 0x04:
655		case 0x05:
656		case 0x06:
657		case 0x07:
658		case 0x08:
659		case 0x09:
660		case 0x0a:
661		case 0x0b:
662		case 0x0c:
663		case 0x0d:
664		case 0x0e:
665		case 0x0f:
666		case 0x10:
667		case 0x11:
668		case 0x12:
669		case 0x13:
670		case 0x14:
671		case 0x15:
672		case 0x16:
673		case 0x17:
674		case 0x18:
675		case 0x19:
676		case 0x1a:
677		case 0x1b:
678		case 0x1c:
679		case 0x1d:
680		case 0x1e:
681		case 0x1f:
682		case 0x20:
683		case 0x21:
684		case 0x22:
685		case 0x23:
686		case 0x24:
687		case 0x25:
688		case 0x26:
689		case 0x27:
690		case 0x28:
691		case 0x29:
692		case 0x2a:
693		case 0x2b:
694		case 0x2c:
695		case 0x2d:
696		case 0x2e:
697		case 0x2f:
698		case 0x30:
699		case 0x31:
700		case 0x32:
701		case 0x33:
702		case 0x34:
703		case 0x35:
704		case 0x36:
705		case 0x37:
706		case 0x38:
707		case 0x39:
708		case 0x3a:
709		case 0x3b:
710		case 0x3c:
711		case 0x3d:
712		case 0x3e:
713		case 0x3f:
714		altmem = (z >> 4) & 0x1ffc;
715		break;
716		case 0x40:
717		case 0x41:
718		case 0x42:
719		case 0x43:
720		case 0x44:
721		case 0x45:
722		case 0x46:
723		case 0x47:
724		case 0x48:
725		case 0x49:
726		case 0x4a:
727		case 0x4b:
728		case 0x4c:
729		case 0x4d:
730		case 0x4e:
731		case 0x4f:
732		case 0x50:
733		case 0x51:
734		case 0x52:
735		case 0x53:
736		case 0x54:
737		case 0x55:
738		case 0x56:
739		case 0x57:
740		case 0x58:
741		case 0x59:
742		case 0x5a:
743		case 0x5b:
744		case 0x5c:
745		case 0x5d:
746		case 0x5e:
747		case 0x5f:
748		altmem = ((z >> 3) & 0x1ffc) \| 0x2000;
749		break;
750		case 0x60:
751		case 0x61:
752		case 0x62:
753		case 0x63:
754		case 0x64:
755		case 0x65:
756		case 0x66:
757		case 0x67:
758		case 0x68:
759		case 0x69:
760		case 0x6a:
761		case 0x6b:
762		case 0x6c:
763		case 0x6d:
764		case 0x6e:
765		case 0x6f:
766		altmem = ((z >> 2) & 0x1ffc) \| 0x4000;
767		break;
768		case 0x70:
769		case 0x71:
770		case 0x72:
771		case 0x73:
772		case 0x74:
773		case 0x75:
774		case 0x76:
775		case 0x77:
776		altmem = ((z >> 1) & 0x1ffc) \| 0x6000;
777		break;
778		case 0x78://uncompressed z = 0x3c000
779		case 0x79:
780		case 0x7a:
781		case 0x7b:
782		altmem = (z & 0x1ffc) \| 0x8000;
783		break;
784		case 0x7c://uncompressed z = 0x3e000
785		case 0x7d:
786		altmem = ((z << 1) & 0x1ffc) \| 0xa000;
787		break;
788		case 0x7e://uncompressed z = 0x3f000
789		altmem = ((z << 2) & 0x1ffc) \| 0xc000;
790		break;
791		case 0x7f://uncompressed z = 0x3f000
792		altmem = ((z << 2) & 0x1ffc) \| 0xe000;
793		break;
794		}
795
796		z_com_table[z] = altmem;
797
798		}
799		}
800
801		void n64_rdp::precalc_cvmask_derivatives(void)
802		{
803		const UINT8 yarray[16] = {0, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0};
804		const UINT8 xarray[16] = {0, 3, 2, 2, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0};
805
806		for (int i = 0; i < 0x10000; i++)
807		{
808		compressed_cvmasks[i] = (i & 1) \| ((i & 4) >> 1) \| ((i & 0x20) >> 3) \| ((i & 0x80) >> 4) \|
809		((i & 0x100) >> 4) \| ((i & 0x400) >> 5) \| ((i & 0x2000) >> 7) \| ((i & 0x8000) >> 8);
810		}
811
812		for (int i = 0; i < 0x100; i++)
813		{
814		UINT16 mask = decompress_cvmask_frombyte(i);
815		cvarray[i].cvg = cvarray[i].cvbit = 0;
816		cvarray[i].cvbit = (i >> 7) & 1;
817		for (int k = 0; k < 8; k++)
818		{
819		cvarray[i].cvg += ((i >> k) & 1);
820		}
821
822		UINT16 masky = 0;
823		for (int k = 0; k < 4; k++)
824		{
825		masky \|= ((mask & (0xf000 >> (k << 2))) > 0) << k;
826		}
827		UINT8 offy = yarray[masky];
828
829		UINT16 maskx = (mask & (0xf000 >> (offy << 2))) >> ((offy ^ 3) << 2);
830		UINT8 offx = xarray[maskx];
831
832		cvarray[i].xoff = offx;
833		cvarray[i].yoff = offy;
834		}
835		}
836
837		UINT16 n64_rdp::decompress_cvmask_frombyte(UINT8 x)
838		{
839		UINT16 y = (x & 1) \| ((x & 2) << 1) \| ((x & 4) << 3) \| ((x & 8) << 4) \|
840		((x & 0x10) << 4) \| ((x & 0x20) << 5) \| ((x & 0x40) << 7) \| ((x & 0x80) << 8);
841		return y;
842		}
843
844		void n64_rdp::lookup_cvmask_derivatives(UINT32 mask, UINT8* offx, UINT8* offy, rdp_span_aux *userdata)
845		{
846		UINT32 index = compressed_cvmasks[mask];
847		userdata->CurrentPixCvg = cvarray[index].cvg;
848		userdata->CurrentCvgBit = cvarray[index].cvbit;
849		*offx = cvarray[index].xoff;
850		*offy = cvarray[index].yoff;
851		}
852
853		void n64_rdp::ZStore(const rdp_poly_state &object, UINT32 zcurpixel, UINT32 dzcurpixel, UINT32 z, UINT32 enc)
854		{
855		UINT16 zval = z_com_table[z & 0x3ffff]\|(enc >> 2);
856		if(zcurpixel <= MEM16_LIMIT)
857		{
858		((UINT16*)rdram)[zcurpixel ^ WORD_ADDR_XOR] = zval;
859		}
860		if(dzcurpixel <= MEM8_LIMIT)
861		{
862		HiddenBits[dzcurpixel ^ BYTE_ADDR_XOR] = enc & 3;
863		}
864		}
865
866		INT32 n64_rdp::NormalizeDZPix(INT32 sum)
867		{
868		if (sum & 0xc000)
869		{
870		return 0x8000;
871		}
872		if (!(sum & 0xffff))
873		{
874		return 1;
875		}
876		for(int count = 0x2000; count > 0; count >>= 1)
877		{
878		if (sum & count)
879		{
880		return(count << 1);
881		}
882		}
883		return 0;
884		}
885
886		UINT32 n64_rdp::ZDecompress(UINT32 zcurpixel)
887		{
888		UINT32 zb = RREADIDX16(zcurpixel);
889		CHECK16(zcurpixel);
890		return z_complete_dec_table[(zb >> 2) & 0x3fff];
891		}
892
893		UINT32 n64_rdp::DZDecompress(UINT32 zcurpixel, UINT32 dzcurpixel)
894		{
895		UINT16 zval = RREADIDX16(zcurpixel);
896		CHECK16(zcurpixel);
897		UINT8 dzval = (((dzcurpixel) <= 0x7fffff) ? (HiddenBits[(dzcurpixel) ^ BYTE_ADDR_XOR]) : 0);
898		UINT32 dz_compressed = ((zval & 3) << 2) \| (dzval & 3);
899		return (1 << dz_compressed);
900		}
901
902		UINT32 n64_rdp::DZCompress(UINT32 value)
903		{
904		INT32 j = 0;
905		for (; value > 1; j++, value >>= 1);
906		return j;
907		}
908
909		void n64_rdp::GetDitherValues(int x, int y, int* cdith, int* adith, const rdp_poly_state& object)
910		{
911		int dithindex = ((y & 3) << 2) \| (x & 3);
912		switch((object.OtherModes.rgb_dither_sel << 2) \| object.OtherModes.alpha_dither_sel)
913		{
914		case 0:
915		adith = cdith = s_magic_matrix[dithindex];
916		break;
917		case 1:
918		*cdith = s_magic_matrix[dithindex];
919		adith = (~(cdith)) & 7;
920		break;
921		case 2:
922		*cdith = s_magic_matrix[dithindex];
923		*adith = machine().rand() & 7;
924		break;
925		case 3:
926		*cdith = s_magic_matrix[dithindex];
927		*adith = 0;
928		break;
929		case 4:
930		adith = cdith = s_bayer_matrix[dithindex];
931		break;
932		case 5:
933		*cdith = s_bayer_matrix[dithindex];
934		adith = (~(cdith)) & 7;
935		break;
936		case 6:
937		*cdith = s_bayer_matrix[dithindex];
938		*adith = machine().rand() & 7;
939		break;
940		case 7:
941		*cdith = s_bayer_matrix[dithindex];
942		*adith = 0;
943		break;
944		case 8:
945		*cdith = machine().rand() & 7;
946		*adith = s_magic_matrix[dithindex];
947		break;
948		case 9:
949		*cdith = machine().rand() & 7;
950		*adith = (~s_magic_matrix[dithindex]) & 7;
951		break;
952		case 10:
953		*cdith = machine().rand() & 7;
954		adith = (cdith + 17) & 7;
955		break;
956		case 11:
957		*cdith = machine().rand() & 7;
958		*adith = 0;
959		break;
960		case 12:
961		*cdith = 0;
962		*adith = s_bayer_matrix[dithindex];
963		break;
964		case 13:
965		*cdith = 0;
966		*adith = (~s_bayer_matrix[dithindex]) & 7;
967		break;
968		case 14:
969		*cdith = 0;
970		*adith = m_machine->rand() & 7;
971		break;
972		case 15:
973		adith = cdith = 0;
974		break;
975		}
976		}
977
978		INT32 CLAMP(INT32 in, INT32 min, INT32 max)
979		{
980		if(in < min) return min;
981		if(in > max) return max;
982		return in;
983		}
984
985		bool n64_rdp::ZCompare(UINT32 zcurpixel, UINT32 dzcurpixel, UINT32 sz, UINT16 dzpix, rdp_span_aux *userdata, const rdp_poly_state &object)
986		{
987		bool force_coplanar = false;
988		sz &= 0x3ffff;
989
990		UINT32 oz;
991		UINT32 dzmem;
992		UINT32 zval;
993		INT32 rawdzmem;
994
995		if (object.OtherModes.z_compare_en)
996		{
997		oz = ZDecompress(zcurpixel);
998		dzmem = DZDecompress(zcurpixel, dzcurpixel);
999		zval = RREADIDX16(zcurpixel);
1000		CHECK16(zcurpixel);
1001		rawdzmem = ((zval & 3) << 2) \| ((((dzcurpixel) <= 0x3fffff) ? (HiddenBits[(dzcurpixel) ^ BYTE_ADDR_XOR]) : 0) & 3);
1002		}
1003		else
1004		{
1005		oz = 0;
1006		dzmem = 1 << 0xf;
1007		zval = 0x3;
1008		rawdzmem = 0xf;
1009		}
1010
1011		userdata->m_dzpix_enc = DZCompress(dzpix & 0xffff);
1012		userdata->ShiftA = CLAMP(userdata->m_dzpix_enc - rawdzmem, 0, 4);
1013		userdata->ShiftB = CLAMP(rawdzmem - userdata->m_dzpix_enc, 0, 4);
1014
1015		int precision_factor = (zval >> 13) & 0xf;
1016		if (precision_factor < 3)
1017		{
1018		int dzmemmodifier = 16 >> precision_factor;
1019		if (dzmem == 0x8000)
1020		{
1021		force_coplanar = true;
1022		}
1023		dzmem <<= 1;
1024		if (dzmem <= dzmemmodifier)
1025		{
1026		dzmem = dzmemmodifier;
1027		}
1028		if (!dzmem)
1029		{
1030		dzmem = 0xffff;
1031		}
1032		}
1033		if (dzmem > 0x8000)
1034		{
1035		dzmem = 0xffff;
1036		}
1037
1038		UINT32 dznew = (dzmem > dzpix) ? dzmem : (UINT32)dzpix;
1039		UINT32 dznotshift = dznew;
1040		dznew <<= 3;
1041
1042		bool farther = (sz + dznew) >= oz;
1043		bool infront = sz < oz;
1044
1045		if (force_coplanar)
1046		{
1047		farther = true;
1048		}
1049
1050		bool overflow = ((userdata->CurrentMemCvg + userdata->CurrentPixCvg) & 8) > 0;
1051		userdata->BlendEnable = (object.OtherModes.force_blend \|\| (!overflow && object.OtherModes.antialias_en && farther)) ? 1 : 0;
1052		userdata->PreWrap = overflow;
1053
1054		int cvgcoeff = 0;
1055		UINT32 dzenc = 0;
1056
1057		if (object.OtherModes.z_mode == 1 && infront && farther && overflow)
1058		{
1059		dzenc = DZCompress(dznotshift & 0xffff);
1060		cvgcoeff = ((oz >> dzenc) - (sz >> dzenc)) & 0xf;
1061		userdata->CurrentPixCvg = ((cvgcoeff * userdata->CurrentPixCvg) >> 3) & 0xf;
1062		}
1063
1064		if (!object.OtherModes.z_compare_en)
1065		{
1066		return true;
1067		}
1068
1069		INT32 diff = (INT32)sz - (INT32)dznew;
1070		bool nearer = diff <= (INT32)oz;
1071		bool max = (oz == 0x3ffff);
1072		if (force_coplanar)
1073		{
1074		nearer = true;
1075		}
1076
1077		switch(object.OtherModes.z_mode)
1078		{
1079		case 0:
1080		return (max \|\| (overflow ? infront : nearer));
1081		case 1:
1082		return (max \|\| (overflow ? infront : nearer));
1083		case 2:
1084		return (infront \|\| max);
1085		case 3:
1086		return (farther && nearer && !max);
1087		}
1088
1089		return false;
1090		}
1091
1092		UINT32 n64_rdp::GetLog2(UINT32 lod_clamp)
1093		{
1094		if (lod_clamp < 2)
1095		{
1096		return 0;
1097		}
1098		else
1099		{
1100		for (int i = 7; i > 0; i--)
1101		{
1102		if ((lod_clamp >> i) & 1)
1103		{
1104		return i;
1105		}
1106		}
1107		}
1108
1109		return 0;
1110		}
1111
1112		/*****************************************************************************/
1113
1114		UINT32 n64_rdp::ReadData(UINT32 address)
1115		{
1116		if (m_status & 0x1) // XBUS_DMEM_DMA enabled
1117		{
1118		return rsp_dmem[(address & 0xfff) / 4];
1119		}
1120		else
1121		{
1122		return rdram[((address & 0xffffff) / 4)];
1123		}
1124		}
1125
1126		static const char *const image_format[] = { "RGBA", "YUV", "CI", "IA", "I", "???", "???", "???" };
1127		static const char *const image_size[] = { "4-bit", "8-bit", "16-bit", "32-bit" };
1128
1129		static const int rdp_command_length[64] =
1130		{
1131		8, // 0x00, No Op
1132		8, // 0x01, ???
1133		8, // 0x02, ???
1134		8, // 0x03, ???
1135		8, // 0x04, ???
1136		8, // 0x05, ???
1137		8, // 0x06, ???
1138		8, // 0x07, ???
1139		32, // 0x08, Non-Shaded Triangle
1140		32+16, // 0x09, Non-Shaded, Z-Buffered Triangle
1141		32+64, // 0x0a, Textured Triangle
1142		32+64+16, // 0x0b, Textured, Z-Buffered Triangle
1143		32+64, // 0x0c, Shaded Triangle
1144		32+64+16, // 0x0d, Shaded, Z-Buffered Triangle
1145		32+64+64, // 0x0e, Shaded+Textured Triangle
1146		32+64+64+16,// 0x0f, Shaded+Textured, Z-Buffered Triangle
1147		8, // 0x10, ???
1148		8, // 0x11, ???
1149		8, // 0x12, ???
1150		8, // 0x13, ???
1151		8, // 0x14, ???
1152		8, // 0x15, ???
1153		8, // 0x16, ???
1154		8, // 0x17, ???
1155		8, // 0x18, ???
1156		8, // 0x19, ???
1157		8, // 0x1a, ???
1158		8, // 0x1b, ???
1159		8, // 0x1c, ???
1160		8, // 0x1d, ???
1161		8, // 0x1e, ???
1162		8, // 0x1f, ???
1163		8, // 0x20, ???
1164		8, // 0x21, ???
1165		8, // 0x22, ???
1166		8, // 0x23, ???
1167		16, // 0x24, Texture_Rectangle
1168		16, // 0x25, Texture_Rectangle_Flip
1169		8, // 0x26, Sync_Load
1170		8, // 0x27, Sync_Pipe
1171		8, // 0x28, Sync_Tile
1172		8, // 0x29, Sync_Full
1173		8, // 0x2a, Set_Key_GB
1174		8, // 0x2b, Set_Key_R
1175		8, // 0x2c, Set_Convert
1176		8, // 0x2d, Set_Scissor
1177		8, // 0x2e, Set_Prim_Depth
1178		8, // 0x2f, Set_Other_Modes
1179		8, // 0x30, Load_TLUT
1180		8, // 0x31, ???
1181		8, // 0x32, Set_Tile_Size
1182		8, // 0x33, Load_Block
1183		8, // 0x34, Load_Tile
1184		8, // 0x35, Set_Tile
1185		8, // 0x36, Fill_Rectangle
1186		8, // 0x37, Set_Fill_Color
1187		8, // 0x38, Set_Fog_Color
1188		8, // 0x39, Set_Blend_Color
1189		8, // 0x3a, Set_Prim_Color
1190		8, // 0x3b, Set_Env_Color
1191		8, // 0x3c, Set_Combine
1192		8, // 0x3d, Set_Texture_Image
1193		8, // 0x3e, Set_Mask_Image
1194		8 // 0x3f, Set_Color_Image
1195		};
1196
1197		void n64_rdp::Dasm(char *buffer)
1198		{
1199		int i;
1200		int tile;
1201		const char format, size;
1202		char sl[32], tl[32], sh[32], th[32];
1203		char s[32], t[32], w[32];
1204		char dsdx[32], dtdx[32], dwdx[32];
1205		char dsdy[32], dtdy[32], dwdy[32];
1206		char dsde[32], dtde[32], dwde[32];
1207		char yl[32], yh[32], ym[32], xl[32], xh[32], xm[32];
1208		char dxldy[32], dxhdy[32], dxmdy[32];
1209		char rt[32], gt[32], bt[32], at[32];
1210		char drdx[32], dgdx[32], dbdx[32], dadx[32];
1211		char drdy[32], dgdy[32], dbdy[32], dady[32];
1212		char drde[32], dgde[32], dbde[32], dade[32];
1213		UINT32 r,g,b,a;
1214
1215		UINT32 cmd[64];
1216		UINT32 length;
1217		UINT32 command;
1218
1219		length = m_cmd_ptr * 4;
1220		if (length < 8)
1221		{
1222		sprintf(buffer, "ERROR: length = %d\n", length);
1223		return;
1224		}
1225
1226		cmd[0] = m_cmd_data[m_cmd_cur+0];
1227		cmd[1] = m_cmd_data[m_cmd_cur+1];
1228
1229		tile = (cmd[1] >> 24) & 0x7;
1230		sprintf(sl, "%4.2f", (float)((cmd[0] >> 12) & 0xfff) / 4.0f);
1231		sprintf(tl, "%4.2f", (float)((cmd[0] >> 0) & 0xfff) / 4.0f);
1232		sprintf(sh, "%4.2f", (float)((cmd[1] >> 12) & 0xfff) / 4.0f);
1233		sprintf(th, "%4.2f", (float)((cmd[1] >> 0) & 0xfff) / 4.0f);
1234
1235		format = image_format[(cmd[0] >> 21) & 0x7];
1236		size = image_size[(cmd[0] >> 19) & 0x3];
1237
1238		r = (cmd[1] >> 24) & 0xff;
1239		g = (cmd[1] >> 16) & 0xff;
1240		b = (cmd[1] >> 8) & 0xff;
1241		a = (cmd[1] >> 0) & 0xff;
1242
1243		command = (cmd[0] >> 24) & 0x3f;
1244		switch (command)
1245		{
1246		case 0x00: sprintf(buffer, "No Op"); break;
1247		case 0x08: // Tri_NoShade
1248		{
1249		int lft = (command >> 23) & 0x1;
1250
1251		if (length != rdp_command_length[command])
1252		{
1253		sprintf(buffer, "ERROR: Tri_NoShade length = %d\n", length);
1254		return;
1255		}
1256
1257		cmd[2] = m_cmd_data[m_cmd_cur+2];
1258		cmd[3] = m_cmd_data[m_cmd_cur+3];
1259		cmd[4] = m_cmd_data[m_cmd_cur+4];
1260		cmd[5] = m_cmd_data[m_cmd_cur+5];
1261		cmd[6] = m_cmd_data[m_cmd_cur+6];
1262		cmd[7] = m_cmd_data[m_cmd_cur+7];
1263
1264		sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
1265		sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
1266		sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
1267		sprintf(xl, "%4.4f", (float)(cmd[2] / 65536.0f));
1268		sprintf(dxldy, "%4.4f", (float)(cmd[3] / 65536.0f));
1269		sprintf(xh, "%4.4f", (float)(cmd[4] / 65536.0f));
1270		sprintf(dxhdy, "%4.4f", (float)(cmd[5] / 65536.0f));
1271		sprintf(xm, "%4.4f", (float)(cmd[6] / 65536.0f));
1272		sprintf(dxmdy, "%4.4f", (float)(cmd[7] / 65536.0f));
1273
1274		sprintf(buffer, "Tri_NoShade %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
1275		break;
1276		}
1277		case 0x09: // Tri_NoShadeZ
1278		{
1279		int lft = (command >> 23) & 0x1;
1280
1281		if (length != rdp_command_length[command])
1282		{
1283		sprintf(buffer, "ERROR: Tri_NoShadeZ length = %d\n", length);
1284		return;
1285		}
1286
1287		cmd[2] = m_cmd_data[m_cmd_cur+2];
1288		cmd[3] = m_cmd_data[m_cmd_cur+3];
1289		cmd[4] = m_cmd_data[m_cmd_cur+4];
1290		cmd[5] = m_cmd_data[m_cmd_cur+5];
1291		cmd[6] = m_cmd_data[m_cmd_cur+6];
1292		cmd[7] = m_cmd_data[m_cmd_cur+7];
1293
1294		sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
1295		sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
1296		sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
1297		sprintf(xl, "%4.4f", (float)(cmd[2] / 65536.0f));
1298		sprintf(dxldy, "%4.4f", (float)(cmd[3] / 65536.0f));
1299		sprintf(xh, "%4.4f", (float)(cmd[4] / 65536.0f));
1300		sprintf(dxhdy, "%4.4f", (float)(cmd[5] / 65536.0f));
1301		sprintf(xm, "%4.4f", (float)(cmd[6] / 65536.0f));
1302		sprintf(dxmdy, "%4.4f", (float)(cmd[7] / 65536.0f));
1303
1304		sprintf(buffer, "Tri_NoShadeZ %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
1305		break;
1306		}
1307		case 0x0a: // Tri_Tex
1308		{
1309		int lft = (command >> 23) & 0x1;
1310
1311		if (length < rdp_command_length[command])
1312		{
1313		sprintf(buffer, "ERROR: Tri_Tex length = %d\n", length);
1314		return;
1315		}
1316
1317		for (i=2; i < 24; i++)
1318		{
1319		cmd[i] = m_cmd_data[m_cmd_cur+i];
1320		}
1321
1322		sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
1323		sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
1324		sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
1325		sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
1326		sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
1327		sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
1328		sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
1329		sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
1330		sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
1331
1332		sprintf(s, "%4.4f", (float)(INT32)((cmd[ 8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
1333		sprintf(t, "%4.4f", (float)(INT32)(((cmd[ 8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
1334		sprintf(w, "%4.4f", (float)(INT32)((cmd[ 9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
1335		sprintf(dsdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
1336		sprintf(dtdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
1337		sprintf(dwdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
1338		sprintf(dsde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
1339		sprintf(dtde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
1340		sprintf(dwde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
1341		sprintf(dsdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
1342		sprintf(dtdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
1343		sprintf(dwdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
1344
1345
1346		buffer+=sprintf(buffer, "Tri_Tex %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
1347		buffer+=sprintf(buffer, " ");
1348		buffer+=sprintf(buffer, " S: %s, T: %s, W: %s\n", s, t, w);
1349		buffer+=sprintf(buffer, " ");
1350		buffer+=sprintf(buffer, " DSDX: %s, DTDX: %s, DWDX: %s\n", dsdx, dtdx, dwdx);
1351		buffer+=sprintf(buffer, " ");
1352		buffer+=sprintf(buffer, " DSDE: %s, DTDE: %s, DWDE: %s\n", dsde, dtde, dwde);
1353		buffer+=sprintf(buffer, " ");
1354		buffer+=sprintf(buffer, " DSDY: %s, DTDY: %s, DWDY: %s\n", dsdy, dtdy, dwdy);
1355		break;
1356		}
1357		case 0x0b: // Tri_TexZ
1358		{
1359		int lft = (command >> 23) & 0x1;
1360
1361		if (length < rdp_command_length[command])
1362		{
1363		sprintf(buffer, "ERROR: Tri_TexZ length = %d\n", length);
1364		return;
1365		}
1366
1367		for (i=2; i < 24; i++)
1368		{
1369		cmd[i] = m_cmd_data[m_cmd_cur+i];
1370		}
1371
1372		sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
1373		sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
1374		sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
1375		sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
1376		sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
1377		sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
1378		sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
1379		sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
1380		sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
1381
1382		sprintf(s, "%4.4f", (float)(INT32)((cmd[ 8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
1383		sprintf(t, "%4.4f", (float)(INT32)(((cmd[ 8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
1384		sprintf(w, "%4.4f", (float)(INT32)((cmd[ 9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
1385		sprintf(dsdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
1386		sprintf(dtdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
1387		sprintf(dwdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
1388		sprintf(dsde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
1389		sprintf(dtde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
1390		sprintf(dwde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
1391		sprintf(dsdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
1392		sprintf(dtdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
1393		sprintf(dwdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
1394
1395
1396		buffer+=sprintf(buffer, "Tri_TexZ %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
1397		buffer+=sprintf(buffer, " ");
1398		buffer+=sprintf(buffer, " S: %s, T: %s, W: %s\n", s, t, w);
1399		buffer+=sprintf(buffer, " ");
1400		buffer+=sprintf(buffer, " DSDX: %s, DTDX: %s, DWDX: %s\n", dsdx, dtdx, dwdx);
1401		buffer+=sprintf(buffer, " ");
1402		buffer+=sprintf(buffer, " DSDE: %s, DTDE: %s, DWDE: %s\n", dsde, dtde, dwde);
1403		buffer+=sprintf(buffer, " ");
1404		buffer+=sprintf(buffer, " DSDY: %s, DTDY: %s, DWDY: %s\n", dsdy, dtdy, dwdy);
1405		break;
1406		}
1407		case 0x0c: // Tri_Shade
1408		{
1409		int lft = (command >> 23) & 0x1;
1410
1411		if (length != rdp_command_length[command])
1412		{
1413		sprintf(buffer, "ERROR: Tri_Shade length = %d\n", length);
1414		return;
1415		}
1416
1417		for (i=2; i < 24; i++)
1418		{
1419		cmd[i] = m_cmd_data[i];
1420		}
1421
1422		sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
1423		sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
1424		sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
1425		sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
1426		sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
1427		sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
1428		sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
1429		sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
1430		sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
1431		sprintf(rt, "%4.4f", (float)(INT32)((cmd[8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
1432		sprintf(gt, "%4.4f", (float)(INT32)(((cmd[8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
1433		sprintf(bt, "%4.4f", (float)(INT32)((cmd[9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
1434		sprintf(at, "%4.4f", (float)(INT32)(((cmd[9] & 0xffff) << 16) \| (cmd[13] & 0xffff)) / 65536.0f);
1435		sprintf(drdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
1436		sprintf(dgdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
1437		sprintf(dbdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
1438		sprintf(dadx, "%4.4f", (float)(INT32)(((cmd[11] & 0xffff) << 16) \| (cmd[15] & 0xffff)) / 65536.0f);
1439		sprintf(drde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
1440		sprintf(dgde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
1441		sprintf(dbde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
1442		sprintf(dade, "%4.4f", (float)(INT32)(((cmd[17] & 0xffff) << 16) \| (cmd[21] & 0xffff)) / 65536.0f);
1443		sprintf(drdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
1444		sprintf(dgdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
1445		sprintf(dbdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
1446		sprintf(dady, "%4.4f", (float)(INT32)(((cmd[19] & 0xffff) << 16) \| (cmd[23] & 0xffff)) / 65536.0f);
1447
1448		buffer+=sprintf(buffer, "Tri_Shade %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
1449		buffer+=sprintf(buffer, " ");
1450		buffer+=sprintf(buffer, " R: %s, G: %s, B: %s, A: %s\n", rt, gt, bt, at);
1451		buffer+=sprintf(buffer, " ");
1452		buffer+=sprintf(buffer, " DRDX: %s, DGDX: %s, DBDX: %s, DADX: %s\n", drdx, dgdx, dbdx, dadx);
1453		buffer+=sprintf(buffer, " ");
1454		buffer+=sprintf(buffer, " DRDE: %s, DGDE: %s, DBDE: %s, DADE: %s\n", drde, dgde, dbde, dade);
1455		buffer+=sprintf(buffer, " ");
1456		buffer+=sprintf(buffer, " DRDY: %s, DGDY: %s, DBDY: %s, DADY: %s\n", drdy, dgdy, dbdy, dady);
1457		break;
1458		}
1459		case 0x0d: // Tri_ShadeZ
1460		{
1461		int lft = (command >> 23) & 0x1;
1462
1463		if (length != rdp_command_length[command])
1464		{
1465		sprintf(buffer, "ERROR: Tri_ShadeZ length = %d\n", length);
1466		return;
1467		}
1468
1469		for (i=2; i < 24; i++)
1470		{
1471		cmd[i] = m_cmd_data[i];
1472		}
1473
1474		sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
1475		sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
1476		sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
1477		sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
1478		sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
1479		sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
1480		sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
1481		sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
1482		sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
1483		sprintf(rt, "%4.4f", (float)(INT32)((cmd[8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
1484		sprintf(gt, "%4.4f", (float)(INT32)(((cmd[8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
1485		sprintf(bt, "%4.4f", (float)(INT32)((cmd[9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
1486		sprintf(at, "%4.4f", (float)(INT32)(((cmd[9] & 0xffff) << 16) \| (cmd[13] & 0xffff)) / 65536.0f);
1487		sprintf(drdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
1488		sprintf(dgdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
1489		sprintf(dbdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
1490		sprintf(dadx, "%4.4f", (float)(INT32)(((cmd[11] & 0xffff) << 16) \| (cmd[15] & 0xffff)) / 65536.0f);
1491		sprintf(drde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
1492		sprintf(dgde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
1493		sprintf(dbde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
1494		sprintf(dade, "%4.4f", (float)(INT32)(((cmd[17] & 0xffff) << 16) \| (cmd[21] & 0xffff)) / 65536.0f);
1495		sprintf(drdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
1496		sprintf(dgdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
1497		sprintf(dbdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
1498		sprintf(dady, "%4.4f", (float)(INT32)(((cmd[19] & 0xffff) << 16) \| (cmd[23] & 0xffff)) / 65536.0f);
1499
1500		buffer+=sprintf(buffer, "Tri_ShadeZ %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
1501		buffer+=sprintf(buffer, " ");
1502		buffer+=sprintf(buffer, " R: %s, G: %s, B: %s, A: %s\n", rt, gt, bt, at);
1503		buffer+=sprintf(buffer, " ");
1504		buffer+=sprintf(buffer, " DRDX: %s, DGDX: %s, DBDX: %s, DADX: %s\n", drdx, dgdx, dbdx, dadx);
1505		buffer+=sprintf(buffer, " ");
1506		buffer+=sprintf(buffer, " DRDE: %s, DGDE: %s, DBDE: %s, DADE: %s\n", drde, dgde, dbde, dade);
1507		buffer+=sprintf(buffer, " ");
1508		buffer+=sprintf(buffer, " DRDY: %s, DGDY: %s, DBDY: %s, DADY: %s\n", drdy, dgdy, dbdy, dady);
1509		break;
1510		}
1511		case 0x0e: // Tri_TexShade
1512		{
1513		int lft = (command >> 23) & 0x1;
1514
1515		if (length < rdp_command_length[command])
1516		{
1517		sprintf(buffer, "ERROR: Tri_TexShade length = %d\n", length);
1518		return;
1519		}
1520
1521		for (i=2; i < 40; i++)
1522		{
1523		cmd[i] = m_cmd_data[m_cmd_cur+i];
1524		}
1525
1526		sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
1527		sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
1528		sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
1529		sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
1530		sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
1531		sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
1532		sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
1533		sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
1534		sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
1535		sprintf(rt, "%4.4f", (float)(INT32)((cmd[8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
1536		sprintf(gt, "%4.4f", (float)(INT32)(((cmd[8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
1537		sprintf(bt, "%4.4f", (float)(INT32)((cmd[9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
1538		sprintf(at, "%4.4f", (float)(INT32)(((cmd[9] & 0xffff) << 16) \| (cmd[13] & 0xffff)) / 65536.0f);
1539		sprintf(drdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
1540		sprintf(dgdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
1541		sprintf(dbdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
1542		sprintf(dadx, "%4.4f", (float)(INT32)(((cmd[11] & 0xffff) << 16) \| (cmd[15] & 0xffff)) / 65536.0f);
1543		sprintf(drde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
1544		sprintf(dgde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
1545		sprintf(dbde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
1546		sprintf(dade, "%4.4f", (float)(INT32)(((cmd[17] & 0xffff) << 16) \| (cmd[21] & 0xffff)) / 65536.0f);
1547		sprintf(drdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
1548		sprintf(dgdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
1549		sprintf(dbdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
1550		sprintf(dady, "%4.4f", (float)(INT32)(((cmd[19] & 0xffff) << 16) \| (cmd[23] & 0xffff)) / 65536.0f);
1551
1552		sprintf(s, "%4.4f", (float)(INT32)((cmd[24] & 0xffff0000) \| ((cmd[28] >> 16) & 0xffff)) / 65536.0f);
1553		sprintf(t, "%4.4f", (float)(INT32)(((cmd[24] & 0xffff) << 16) \| (cmd[28] & 0xffff)) / 65536.0f);
1554		sprintf(w, "%4.4f", (float)(INT32)((cmd[25] & 0xffff0000) \| ((cmd[29] >> 16) & 0xffff)) / 65536.0f);
1555		sprintf(dsdx, "%4.4f", (float)(INT32)((cmd[26] & 0xffff0000) \| ((cmd[30] >> 16) & 0xffff)) / 65536.0f);
1556		sprintf(dtdx, "%4.4f", (float)(INT32)(((cmd[26] & 0xffff) << 16) \| (cmd[30] & 0xffff)) / 65536.0f);
1557		sprintf(dwdx, "%4.4f", (float)(INT32)((cmd[27] & 0xffff0000) \| ((cmd[31] >> 16) & 0xffff)) / 65536.0f);
1558		sprintf(dsde, "%4.4f", (float)(INT32)((cmd[32] & 0xffff0000) \| ((cmd[36] >> 16) & 0xffff)) / 65536.0f);
1559		sprintf(dtde, "%4.4f", (float)(INT32)(((cmd[32] & 0xffff) << 16) \| (cmd[36] & 0xffff)) / 65536.0f);
1560		sprintf(dwde, "%4.4f", (float)(INT32)((cmd[33] & 0xffff0000) \| ((cmd[37] >> 16) & 0xffff)) / 65536.0f);
1561		sprintf(dsdy, "%4.4f", (float)(INT32)((cmd[34] & 0xffff0000) \| ((cmd[38] >> 16) & 0xffff)) / 65536.0f);
1562		sprintf(dtdy, "%4.4f", (float)(INT32)(((cmd[34] & 0xffff) << 16) \| (cmd[38] & 0xffff)) / 65536.0f);
1563		sprintf(dwdy, "%4.4f", (float)(INT32)((cmd[35] & 0xffff0000) \| ((cmd[39] >> 16) & 0xffff)) / 65536.0f);
1564
1565
1566		buffer+=sprintf(buffer, "Tri_TexShade %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
1567		buffer+=sprintf(buffer, " ");
1568		buffer+=sprintf(buffer, " R: %s, G: %s, B: %s, A: %s\n", rt, gt, bt, at);
1569		buffer+=sprintf(buffer, " ");
1570		buffer+=sprintf(buffer, " DRDX: %s, DGDX: %s, DBDX: %s, DADX: %s\n", drdx, dgdx, dbdx, dadx);
1571		buffer+=sprintf(buffer, " ");
1572		buffer+=sprintf(buffer, " DRDE: %s, DGDE: %s, DBDE: %s, DADE: %s\n", drde, dgde, dbde, dade);
1573		buffer+=sprintf(buffer, " ");
1574		buffer+=sprintf(buffer, " DRDY: %s, DGDY: %s, DBDY: %s, DADY: %s\n", drdy, dgdy, dbdy, dady);
1575
1576		buffer+=sprintf(buffer, " ");
1577		buffer+=sprintf(buffer, " S: %s, T: %s, W: %s\n", s, t, w);
1578		buffer+=sprintf(buffer, " ");
1579		buffer+=sprintf(buffer, " DSDX: %s, DTDX: %s, DWDX: %s\n", dsdx, dtdx, dwdx);
1580		buffer+=sprintf(buffer, " ");
1581		buffer+=sprintf(buffer, " DSDE: %s, DTDE: %s, DWDE: %s\n", dsde, dtde, dwde);
1582		buffer+=sprintf(buffer, " ");
1583		buffer+=sprintf(buffer, " DSDY: %s, DTDY: %s, DWDY: %s\n", dsdy, dtdy, dwdy);
1584		break;
1585		}
1586		case 0x0f: // Tri_TexShadeZ
1587		{
1588		int lft = (command >> 23) & 0x1;
1589
1590		if (length < rdp_command_length[command])
1591		{
1592		sprintf(buffer, "ERROR: Tri_TexShadeZ length = %d\n", length);
1593		return;
1594		}
1595
1596		for (i=2; i < 40; i++)
1597		{
1598		cmd[i] = m_cmd_data[m_cmd_cur+i];
1599		}
1600
1601		sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
1602		sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
1603		sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
1604		sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
1605		sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
1606		sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
1607		sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
1608		sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
1609		sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
1610		sprintf(rt, "%4.4f", (float)(INT32)((cmd[8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
1611		sprintf(gt, "%4.4f", (float)(INT32)(((cmd[8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
1612		sprintf(bt, "%4.4f", (float)(INT32)((cmd[9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
1613		sprintf(at, "%4.4f", (float)(INT32)(((cmd[9] & 0xffff) << 16) \| (cmd[13] & 0xffff)) / 65536.0f);
1614		sprintf(drdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
1615		sprintf(dgdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
1616		sprintf(dbdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
1617		sprintf(dadx, "%4.4f", (float)(INT32)(((cmd[11] & 0xffff) << 16) \| (cmd[15] & 0xffff)) / 65536.0f);
1618		sprintf(drde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
1619		sprintf(dgde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
1620		sprintf(dbde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
1621		sprintf(dade, "%4.4f", (float)(INT32)(((cmd[17] & 0xffff) << 16) \| (cmd[21] & 0xffff)) / 65536.0f);
1622		sprintf(drdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
1623		sprintf(dgdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
1624		sprintf(dbdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
1625		sprintf(dady, "%4.4f", (float)(INT32)(((cmd[19] & 0xffff) << 16) \| (cmd[23] & 0xffff)) / 65536.0f);
1626
1627		sprintf(s, "%4.4f", (float)(INT32)((cmd[24] & 0xffff0000) \| ((cmd[28] >> 16) & 0xffff)) / 65536.0f);
1628		sprintf(t, "%4.4f", (float)(INT32)(((cmd[24] & 0xffff) << 16) \| (cmd[28] & 0xffff)) / 65536.0f);
1629		sprintf(w, "%4.4f", (float)(INT32)((cmd[25] & 0xffff0000) \| ((cmd[29] >> 16) & 0xffff)) / 65536.0f);
1630		sprintf(dsdx, "%4.4f", (float)(INT32)((cmd[26] & 0xffff0000) \| ((cmd[30] >> 16) & 0xffff)) / 65536.0f);
1631		sprintf(dtdx, "%4.4f", (float)(INT32)(((cmd[26] & 0xffff) << 16) \| (cmd[30] & 0xffff)) / 65536.0f);
1632		sprintf(dwdx, "%4.4f", (float)(INT32)((cmd[27] & 0xffff0000) \| ((cmd[31] >> 16) & 0xffff)) / 65536.0f);
1633		sprintf(dsde, "%4.4f", (float)(INT32)((cmd[32] & 0xffff0000) \| ((cmd[36] >> 16) & 0xffff)) / 65536.0f);
1634		sprintf(dtde, "%4.4f", (float)(INT32)(((cmd[32] & 0xffff) << 16) \| (cmd[36] & 0xffff)) / 65536.0f);
1635		sprintf(dwde, "%4.4f", (float)(INT32)((cmd[33] & 0xffff0000) \| ((cmd[37] >> 16) & 0xffff)) / 65536.0f);
1636		sprintf(dsdy, "%4.4f", (float)(INT32)((cmd[34] & 0xffff0000) \| ((cmd[38] >> 16) & 0xffff)) / 65536.0f);
1637		sprintf(dtdy, "%4.4f", (float)(INT32)(((cmd[34] & 0xffff) << 16) \| (cmd[38] & 0xffff)) / 65536.0f);
1638		sprintf(dwdy, "%4.4f", (float)(INT32)((cmd[35] & 0xffff0000) \| ((cmd[39] >> 16) & 0xffff)) / 65536.0f);
1639
1640
1641		buffer+=sprintf(buffer, "Tri_TexShadeZ %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
1642		buffer+=sprintf(buffer, " ");
1643		buffer+=sprintf(buffer, " R: %s, G: %s, B: %s, A: %s\n", rt, gt, bt, at);
1644		buffer+=sprintf(buffer, " ");
1645		buffer+=sprintf(buffer, " DRDX: %s, DGDX: %s, DBDX: %s, DADX: %s\n", drdx, dgdx, dbdx, dadx);
1646		buffer+=sprintf(buffer, " ");
1647		buffer+=sprintf(buffer, " DRDE: %s, DGDE: %s, DBDE: %s, DADE: %s\n", drde, dgde, dbde, dade);
1648		buffer+=sprintf(buffer, " ");
1649		buffer+=sprintf(buffer, " DRDY: %s, DGDY: %s, DBDY: %s, DADY: %s\n", drdy, dgdy, dbdy, dady);
1650
1651		buffer+=sprintf(buffer, " ");
1652		buffer+=sprintf(buffer, " S: %s, T: %s, W: %s\n", s, t, w);
1653		buffer+=sprintf(buffer, " ");
1654		buffer+=sprintf(buffer, " DSDX: %s, DTDX: %s, DWDX: %s\n", dsdx, dtdx, dwdx);
1655		buffer+=sprintf(buffer, " ");
1656		buffer+=sprintf(buffer, " DSDE: %s, DTDE: %s, DWDE: %s\n", dsde, dtde, dwde);
1657		buffer+=sprintf(buffer, " ");
1658		buffer+=sprintf(buffer, " DSDY: %s, DTDY: %s, DWDY: %s\n", dsdy, dtdy, dwdy);
1659		break;
1660		}
1661		case 0x24:
1662		case 0x25:
1663		{
1664		if (length < 16)
1665		{
1666		sprintf(buffer, "ERROR: Texture_Rectangle length = %d\n", length);
1667		return;
1668		}
1669		cmd[2] = m_cmd_data[m_cmd_cur+2];
1670		cmd[3] = m_cmd_data[m_cmd_cur+3];
1671		sprintf(s, "%4.4f", (float)(INT16)((cmd[2] >> 16) & 0xffff) / 32.0f);
1672		sprintf(t, "%4.4f", (float)(INT16)((cmd[2] >> 0) & 0xffff) / 32.0f);
1673		sprintf(dsdx, "%4.4f", (float)(INT16)((cmd[3] >> 16) & 0xffff) / 1024.0f);
1674		sprintf(dtdy, "%4.4f", (float)(INT16)((cmd[3] >> 16) & 0xffff) / 1024.0f);
1675
1676		if (command == 0x24)
1677		sprintf(buffer, "Texture_Rectangle %d, %s, %s, %s, %s, %s, %s, %s, %s", tile, sh, th, sl, tl, s, t, dsdx, dtdy);
1678		else
1679		sprintf(buffer, "Texture_Rectangle_Flip %d, %s, %s, %s, %s, %s, %s, %s, %s", tile, sh, th, sl, tl, s, t, dsdx, dtdy);
1680
1681		break;
1682		}
1683		case 0x26: sprintf(buffer, "Sync_Load"); break;
1684		case 0x27: sprintf(buffer, "Sync_Pipe"); break;
1685		case 0x28: sprintf(buffer, "Sync_Tile"); break;
1686		case 0x29: sprintf(buffer, "Sync_Full"); break;
1687		case 0x2d: sprintf(buffer, "Set_Scissor %s, %s, %s, %s", sl, tl, sh, th); break;
1688		case 0x2e: sprintf(buffer, "Set_Prim_Depth %04X, %04X", (cmd[1] >> 16) & 0xffff, cmd[1] & 0xffff); break;
1689		case 0x2f: sprintf(buffer, "Set_Other_Modes %08X %08X", cmd[0], cmd[1]); break;
1690		case 0x30: sprintf(buffer, "Load_TLUT %d, %s, %s, %s, %s", tile, sl, tl, sh, th); break;
1691		case 0x32: sprintf(buffer, "Set_Tile_Size %d, %s, %s, %s, %s", tile, sl, tl, sh, th); break;
1692		case 0x33: sprintf(buffer, "Load_Block %d, %03X, %03X, %03X, %03X", tile, (cmd[0] >> 12) & 0xfff, cmd[0] & 0xfff, (cmd[1] >> 12) & 0xfff, cmd[1] & 0xfff); break;
1693		case 0x34: sprintf(buffer, "Load_Tile %d, %s, %s, %s, %s", tile, sl, tl, sh, th); break;
1694		case 0x35: sprintf(buffer, "Set_Tile %d, %s, %s, %d, %04X", tile, format, size, ((cmd[0] >> 9) & 0x1ff) * 8, (cmd[0] & 0x1ff) * 8); break;
1695		case 0x36: sprintf(buffer, "Fill_Rectangle %s, %s, %s, %s", sh, th, sl, tl); break;
1696		case 0x37: sprintf(buffer, "Set_Fill_Color R: %d, G: %d, B: %d, A: %d", r, g, b, a); break;
1697		case 0x38: sprintf(buffer, "Set_Fog_Color R: %d, G: %d, B: %d, A: %d", r, g, b, a); break;
1698		case 0x39: sprintf(buffer, "Set_Blend_Color R: %d, G: %d, B: %d, A: %d", r, g, b, a); break;
1699		case 0x3a: sprintf(buffer, "Set_Prim_Color %d, %d, R: %d, G: %d, B: %d, A: %d", (cmd[0] >> 8) & 0x1f, cmd[0] & 0xff, r, g, b, a); break;
1700		case 0x3b: sprintf(buffer, "Set_Env_Color R: %d, G: %d, B: %d, A: %d", r, g, b, a); break;
1701		case 0x3c: sprintf(buffer, "Set_Combine %08X %08X", cmd[0], cmd[1]); break;
1702		case 0x3d: sprintf(buffer, "Set_Texture_Image %s, %s, %d, %08X", format, size, (cmd[0] & 0x1ff)+1, cmd[1]); break;
1703		case 0x3e: sprintf(buffer, "Set_Mask_Image %08X", cmd[1]); break;
1704		case 0x3f: sprintf(buffer, "Set_Color_Image %s, %s, %d, %08X", format, size, (cmd[0] & 0x1ff)+1, cmd[1]); break;
1705		default: sprintf(buffer, "??? (%08X %08X)", cmd[0], cmd[1]); break;
1706		}
1707		}
1708
1709		/*****************************************************************************/
1710
1711		static UINT32 rightcvghex(UINT32 x, UINT32 fmask)
1712		{
1713		UINT32 stickybit = ((x >> 1) & 0x1fff) > 0;
1714		UINT32 covered = ((x >> 14) & 3) + stickybit;
1715		covered = (0xf0 >> covered) & 0xf;
1716		return (covered & fmask);
1717		}
1718
1719		static UINT32 leftcvghex(UINT32 x, UINT32 fmask)
1720		{
1721		UINT32 stickybit = ((x >> 1) & 0x1fff) > 0;
1722		UINT32 covered = ((x >> 14) & 3) + stickybit;
1723		covered = 0xf >> covered;
1724		return (covered & fmask);
1725		}
1726
1727		static INT32 CLIP(INT32 value,INT32 min,INT32 max)
1728		{
1729		if (value < min)
1730		{
1731		return min;
1732		}
1733		else if (value > max)
1734		{
1735		return max;
1736		}
1737		else
1738		{
1739		return value;
1740		}
1741		}
1742
1743		void n64_rdp::compute_cvg_noflip(extent_t Spans, INT32 majorx, INT32* minorx, INT32* majorxint, INT32* minorxint, INT32 scanline, INT32 yh, INT32 yl, INT32 base)
1744		{
1745		INT32 purgestart = 0xfff;
1746		INT32 purgeend = 0;
1747		bool writablescanline = !(scanline & ~0x3ff);
1748		INT32 scanlinespx = scanline << 2;
1749
1750		if (!writablescanline) return;
1751
1752		for(int i = 0; i < 4; i++)
1753		{
1754		if (minorxint[i] < purgestart)
1755		{
1756		purgestart = minorxint[i];
1757		}
1758		if (majorxint[i] > purgeend)
1759		{
1760		purgeend = majorxint[i];
1761		}
1762		}
1763
1764		purgestart = CLIP(purgestart, 0, 1023);
1765		purgeend = CLIP(purgeend, 0, 1023);
1766		INT32 length = purgeend - purgestart;
1767
1768		if (length < 0) return;
1769
1770		rdp_span_aux userdata = (rdp_span_aux)Spans[scanline - base].userdata;
1771		memset(&userdata->m_cvg[purgestart], 0, (length + 1) << 1);
1772
1773		for(int i = 0; i < 4; i++)
1774		{
1775		INT32 minorcur = minorx[i];
1776		INT32 majorcur = majorx[i];
1777		INT32 minorcurint = minorxint[i];
1778		INT32 majorcurint = majorxint[i];
1779		length = majorcurint - minorcurint;
1780
1781		INT32 fmask = (i & 1) ? 5 : 0xa;
1782		INT32 maskshift = (i ^ 3) << 2;
1783		INT32 fmaskshifted = fmask << maskshift;
1784		INT32 fleft = CLIP(minorcurint + 1, 0, 647);
1785		INT32 fright = CLIP(majorcurint - 1, 0, 647);
1786		bool valid_y = ((scanlinespx + i) >= yh && (scanlinespx + i) < yl);
1787		if (valid_y && length >= 0)
1788		{
1789		if (minorcurint != majorcurint)
1790		{
1791		if (!(minorcurint & ~0x3ff))
1792		{
1793		userdata->m_cvg[minorcurint] \|= (leftcvghex(minorcur, fmask) << maskshift);
1794		}
1795		if (!(majorcurint & ~0x3ff))
1796		{
1797		userdata->m_cvg[majorcurint] \|= (rightcvghex(majorcur, fmask) << maskshift);
1798		}
1799		}
1800		else
1801		{
1802		if (!(majorcurint & ~0x3ff))
1803		{
1804		INT32 samecvg = leftcvghex(minorcur, fmask) & rightcvghex(majorcur, fmask);
1805		userdata->m_cvg[majorcurint] \|= (samecvg << maskshift);
1806		}
1807		}
1808		for (; fleft <= fright; fleft++)
1809		{
1810		userdata->m_cvg[fleft] \|= fmaskshifted;
1811		}
1812		}
1813		}
1814		}
1815
1816		void n64_rdp::compute_cvg_flip(extent_t Spans, INT32 majorx, INT32* minorx, INT32* majorxint, INT32* minorxint, INT32 scanline, INT32 yh, INT32 yl, INT32 base)
1817		{
1818		INT32 purgestart = 0xfff;
1819		INT32 purgeend = 0;
1820		bool writablescanline = !(scanline & ~0x3ff);
1821		INT32 scanlinespx = scanline << 2;
1822
1823		if(!writablescanline) return;
1824
1825		for(int i = 0; i < 4; i++)
1826		{
1827		if (majorxint[i] < purgestart)
1828		{
1829		purgestart = majorxint[i];
1830		}
1831		if (minorxint[i] > purgeend)
1832		{
1833		purgeend = minorxint[i];
1834		}
1835		}
1836
1837		purgestart = CLIP(purgestart, 0, 1023);
1838		purgeend = CLIP(purgeend, 0, 1023);
1839
1840		int length = purgeend - purgestart;
1841
1842		if (length < 0) return;
1843
1844		rdp_span_aux userdata = (rdp_span_aux)Spans[scanline - base].userdata;
1845		memset(&userdata->m_cvg[purgestart], 0, (length + 1) << 1);
1846
1847		for(int i = 0; i < 4; i++)
1848		{
1849		INT32 minorcur = minorx[i];
1850		INT32 majorcur = majorx[i];
1851		INT32 minorcurint = minorxint[i];
1852		INT32 majorcurint = majorxint[i];
1853		length = minorcurint - majorcurint;
1854
1855		INT32 fmask = (i & 1) ? 5 : 0xa;
1856		INT32 maskshift = (i ^ 3) << 2;
1857		INT32 fmaskshifted = fmask << maskshift;
1858		INT32 fleft = CLIP(majorcurint + 1, 0, 647);
1859		INT32 fright = CLIP(minorcurint - 1, 0, 647);
1860		bool valid_y = ((scanlinespx + i) >= yh && (scanlinespx + i) < yl);
1861		if (valid_y && length >= 0)
1862		{
1863		if (minorcurint != majorcurint)
1864		{
1865		if (!(minorcurint & ~0x3ff))
1866		{
1867		userdata->m_cvg[minorcurint] \|= (rightcvghex(minorcur, fmask) << maskshift);
1868		}
1869		if (!(majorcurint & ~0x3ff))
1870		{
1871		userdata->m_cvg[majorcurint] \|= (leftcvghex(majorcur, fmask) << maskshift);
1872		}
1873		}
1874		else
1875		{
1876		if (!(majorcurint & ~0x3ff))
1877		{
1878		INT32 samecvg = rightcvghex(minorcur, fmask) & leftcvghex(majorcur, fmask);
1879		userdata->m_cvg[majorcurint] \|= (samecvg << maskshift);
1880		}
1881		}
1882		for (; fleft <= fright; fleft++)
1883		{
1884		userdata->m_cvg[fleft] \|= fmaskshifted;
1885		}
1886		}
1887		}
1888		}
1889
1890		void n64_rdp::DrawTriangle(bool shade, bool texture, bool zbuffer, bool rect)
1891		{
1892		UINT32 *cmd_data = rect ? m_temp_rect_data : m_cmd_data;
1893		UINT32 fifo_index = rect ? 0 : m_cmd_cur;
1894		UINT32 w1 = cmd_data[fifo_index + 0];
1895		UINT32 w2 = cmd_data[fifo_index + 1];
1896
1897		int flip = (w1 & 0x00800000) ? 1 : 0;
1898		MiscState.MaxLevel = ((w1 >> 19) & 7);
1899		int tilenum = (w1 >> 16) & 0x7;
1900
1901		int dsdiff = 0, dtdiff = 0, dwdiff = 0, drdiff = 0, dgdiff = 0, dbdiff = 0, dadiff = 0, dzdiff = 0;
1902		int dsdeh = 0, dtdeh = 0, dwdeh = 0, drdeh = 0, dgdeh = 0, dbdeh = 0, dadeh = 0, dzdeh = 0;
1903		int dsdxh = 0, dtdxh = 0, dwdxh = 0, drdxh = 0, dgdxh = 0, dbdxh = 0, dadxh = 0, dzdxh = 0;
1904		int dsdyh = 0, dtdyh = 0, dwdyh = 0, drdyh = 0, dgdyh = 0, dbdyh = 0, dadyh = 0, dzdyh = 0;
1905
1906		INT32 maxxmx = 0; // maxxmx / minxhx very opaque names, consider re-naming
1907		INT32 minxmx = 0;
1908		INT32 maxxhx = 0;
1909		INT32 minxhx = 0;
1910
1911		int shade_base = fifo_index + 8;
1912		int texture_base = fifo_index + 8;
1913		int zbuffer_base = fifo_index + 8;
1914		if(shade)
1915		{
1916		texture_base += 16;
1917		zbuffer_base += 16;
1918		}
1919		if(texture)
1920		{
1921		zbuffer_base += 16;
1922		}
1923
1924		UINT32 w3 = cmd_data[fifo_index + 2];
1925		UINT32 w4 = cmd_data[fifo_index + 3];
1926		UINT32 w5 = cmd_data[fifo_index + 4];
1927		UINT32 w6 = cmd_data[fifo_index + 5];
1928		UINT32 w7 = cmd_data[fifo_index + 6];
1929		UINT32 w8 = cmd_data[fifo_index + 7];
1930
1931		INT32 yl = (w1 & 0x3fff);
1932		INT32 ym = ((w2 >> 16) & 0x3fff);
1933		INT32 yh = ((w2 >> 0) & 0x3fff);
1934		INT32 xl = (INT32)(w3 & 0x3fffffff);
1935		INT32 xh = (INT32)(w5 & 0x3fffffff);
1936		INT32 xm = (INT32)(w7 & 0x3fffffff);
1937		// Inverse slopes in 16.16 format
1938		INT32 dxldy = (INT32)(w4);
1939		INT32 dxhdy = (INT32)(w6);
1940		INT32 dxmdy = (INT32)(w8);
1941
1942		if (yl & 0x2000) yl \|= 0xffffc000;
1943		if (ym & 0x2000) ym \|= 0xffffc000;
1944		if (yh & 0x2000) yh \|= 0xffffc000;
1945
1946		if (xl & 0x20000000) xl \|= 0xc0000000;
1947		if (xm & 0x20000000) xm \|= 0xc0000000;
1948		if (xh & 0x20000000) xh \|= 0xc0000000;
1949
1950		int r = (cmd_data[shade_base+0 ] & 0xffff0000) \| ((cmd_data[shade_base+4 ] >> 16) & 0x0000ffff);
1951		int g = ((cmd_data[shade_base+0 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+4 ] & 0x0000ffff);
1952		int b = (cmd_data[shade_base+1 ] & 0xffff0000) \| ((cmd_data[shade_base+5 ] >> 16) & 0x0000ffff);
1953		int a = ((cmd_data[shade_base+1 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+5 ] & 0x0000ffff);
1954		int drdx = (cmd_data[shade_base+2 ] & 0xffff0000) \| ((cmd_data[shade_base+6 ] >> 16) & 0x0000ffff);
1955		int dgdx = ((cmd_data[shade_base+2 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+6 ] & 0x0000ffff);
1956		int dbdx = (cmd_data[shade_base+3 ] & 0xffff0000) \| ((cmd_data[shade_base+7 ] >> 16) & 0x0000ffff);
1957		int dadx = ((cmd_data[shade_base+3 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+7 ] & 0x0000ffff);
1958		int drde = (cmd_data[shade_base+8 ] & 0xffff0000) \| ((cmd_data[shade_base+12] >> 16) & 0x0000ffff);
1959		int dgde = ((cmd_data[shade_base+8 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+12] & 0x0000ffff);
1960		int dbde = (cmd_data[shade_base+9 ] & 0xffff0000) \| ((cmd_data[shade_base+13] >> 16) & 0x0000ffff);
1961		int dade = ((cmd_data[shade_base+9 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+13] & 0x0000ffff);
1962		int drdy = (cmd_data[shade_base+10] & 0xffff0000) \| ((cmd_data[shade_base+14] >> 16) & 0x0000ffff);
1963		int dgdy = ((cmd_data[shade_base+10] << 16) & 0xffff0000) \| (cmd_data[shade_base+14] & 0x0000ffff);
1964		int dbdy = (cmd_data[shade_base+11] & 0xffff0000) \| ((cmd_data[shade_base+15] >> 16) & 0x0000ffff);
1965		int dady = ((cmd_data[shade_base+11] << 16) & 0xffff0000) \| (cmd_data[shade_base+15] & 0x0000ffff);
1966		int s = (cmd_data[texture_base+0 ] & 0xffff0000) \| ((cmd_data[texture_base+4 ] >> 16) & 0x0000ffff);
1967		int t = ((cmd_data[texture_base+0 ] << 16) & 0xffff0000) \| (cmd_data[texture_base+4 ] & 0x0000ffff);
1968		int w = (cmd_data[texture_base+1 ] & 0xffff0000) \| ((cmd_data[texture_base+5 ] >> 16) & 0x0000ffff);
1969		int dsdx = (cmd_data[texture_base+2 ] & 0xffff0000) \| ((cmd_data[texture_base+6 ] >> 16) & 0x0000ffff);
1970		int dtdx = ((cmd_data[texture_base+2 ] << 16) & 0xffff0000) \| (cmd_data[texture_base+6 ] & 0x0000ffff);
1971		int dwdx = (cmd_data[texture_base+3 ] & 0xffff0000) \| ((cmd_data[texture_base+7 ] >> 16) & 0x0000ffff);
1972		int dsde = (cmd_data[texture_base+8 ] & 0xffff0000) \| ((cmd_data[texture_base+12] >> 16) & 0x0000ffff);
1973		int dtde = ((cmd_data[texture_base+8 ] << 16) & 0xffff0000) \| (cmd_data[texture_base+12] & 0x0000ffff);
1974		int dwde = (cmd_data[texture_base+9 ] & 0xffff0000) \| ((cmd_data[texture_base+13] >> 16) & 0x0000ffff);
1975		int dsdy = (cmd_data[texture_base+10] & 0xffff0000) \| ((cmd_data[texture_base+14] >> 16) & 0x0000ffff);
1976		int dtdy = ((cmd_data[texture_base+10] << 16) & 0xffff0000) \| (cmd_data[texture_base+14] & 0x0000ffff);
1977		int dwdy = (cmd_data[texture_base+11] & 0xffff0000) \| ((cmd_data[texture_base+15] >> 16) & 0x0000ffff);
1978		int z = cmd_data[zbuffer_base+0];
1979		int dzdx = cmd_data[zbuffer_base+1];
1980		int dzde = cmd_data[zbuffer_base+2];
1981		int dzdy = cmd_data[zbuffer_base+3];
1982
1983		int dzdy_dz = (dzdy >> 16) & 0xffff;
1984		int dzdx_dz = (dzdx >> 16) & 0xffff;
1985
1986		extent_t Spans[2048];
1987
1988		SpanBase.m_span_drdy = drdy;
1989		SpanBase.m_span_dgdy = dgdy;
1990		SpanBase.m_span_dbdy = dbdy;
1991		SpanBase.m_span_dady = dady;
1992		SpanBase.m_span_dzdy = OtherModes.z_source_sel ? 0 : dzdy;
1993
1994		UINT32 temp_dzpix = ((dzdy_dz & 0x8000) ? ((~dzdy_dz) & 0x7fff) : dzdy_dz) + ((dzdx_dz & 0x8000) ? ((~dzdx_dz) & 0x7fff) : dzdx_dz);
1995		SpanBase.m_span_dr = drdx & ~0x1f;
1996		SpanBase.m_span_dg = dgdx & ~0x1f;
1997		SpanBase.m_span_db = dbdx & ~0x1f;
1998		SpanBase.m_span_da = dadx & ~0x1f;
1999		SpanBase.m_span_ds = dsdx;
2000		SpanBase.m_span_dt = dtdx;
2001		SpanBase.m_span_dw = dwdx;
2002		SpanBase.m_span_dz = OtherModes.z_source_sel ? 0 : dzdx;
2003		SpanBase.m_span_dymax = 0;
2004		SpanBase.m_span_dzpix = NormalizeDZPix(temp_dzpix & 0xffff) & 0xffff;
2005
2006		int xleft_inc = (dxmdy >> 2) & ~1;
2007		int xright_inc = (dxhdy >> 2) & ~1;
2008
2009		int xright = xh & ~1;
2010		int xleft = xm & ~1;
2011
2012		int sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0;
2013		int do_offset = !(sign_dxhdy ^ (flip));
2014
2015		if (do_offset)
2016		{
2017		dsdeh = dsde >> 9; dsdyh = dsdy >> 9;
2018		dtdeh = dtde >> 9; dtdyh = dtdy >> 9;
2019		dwdeh = dwde >> 9; dwdyh = dwdy >> 9;
2020		drdeh = drde >> 9; drdyh = drdy >> 9;
2021		dgdeh = dgde >> 9; dgdyh = dgdy >> 9;
2022		dbdeh = dbde >> 9; dbdyh = dbdy >> 9;
2023		dadeh = dade >> 9; dadyh = dady >> 9;
2024		dzdeh = dzde >> 9; dzdyh = dzdy >> 9;
2025
2026		dsdiff = (dsdeh << 8) + (dsdeh << 7) - (dsdyh << 8) - (dsdyh << 7);
2027		dtdiff = (dtdeh << 8) + (dtdeh << 7) - (dtdyh << 8) - (dtdyh << 7);
2028		dwdiff = (dwdeh << 8) + (dwdeh << 7) - (dwdyh << 8) - (dwdyh << 7);
2029		drdiff = (drdeh << 8) + (drdeh << 7) - (drdyh << 8) - (drdyh << 7);
2030		dgdiff = (dgdeh << 8) + (dgdeh << 7) - (dgdyh << 8) - (dgdyh << 7);
2031		dbdiff = (dbdeh << 8) + (dbdeh << 7) - (dbdyh << 8) - (dbdyh << 7);
2032		dadiff = (dadeh << 8) + (dadeh << 7) - (dadyh << 8) - (dadyh << 7);
2033		dzdiff = (dzdeh << 8) + (dzdeh << 7) - (dzdyh << 8) - (dzdyh << 7);
2034		}
2035		else
2036		{
2037		dsdiff = dtdiff = dwdiff = drdiff = dgdiff = dbdiff = dadiff = dzdiff = 0;
2038		}
2039
2040		dsdxh = dsdx >> 8;
2041		dtdxh = dtdx >> 8;
2042		dwdxh = dwdx >> 8;
2043		drdxh = drdx >> 8;
2044		dgdxh = dgdx >> 8;
2045		dbdxh = dbdx >> 8;
2046		dadxh = dadx >> 8;
2047		dzdxh = dzdx >> 8;
2048
2049		INT32 ycur = yh & ~3;
2050		INT32 ylfar = yl \| 3;
2051		INT32 ldflag = (sign_dxhdy ^ flip) ? 0 : 3;
2052		INT32 majorx[4];
2053		INT32 minorx[4];
2054		INT32 majorxint[4];
2055		INT32 minorxint[4];
2056		bool valid_y = true;
2057
2058		int xfrac = ((xright >> 8) & 0xff);
2059
2060		int clipy1 = Scissor.m_yh;
2061		int clipy2 = Scissor.m_yl;
2062
2063		// Trivial reject
2064		if((ycur >> 2) >= clipy2 && (ylfar >> 2) >= clipy2)
2065		{
2066		return;
2067		}
2068		if((ycur >> 2) < clipy1 && (ylfar >> 2) < clipy1)
2069		{
2070		return;
2071		}
2072
2073		bool new_object = true;
2074		rdp_poly_state *object = NULL;
2075		bool valid = false;
2076
2077		if(flip)
2078		{
2079		for (int k = ycur; k <= ylfar; k++)
2080		{
2081		if (k == ym)
2082		{
2083		xleft = xl & ~1;
2084		xleft_inc = (dxldy >> 2) & ~1;
2085		}
2086
2087		int xstart = xleft >> 16; // 319
2088		int xend = xright >> 16; // 0
2089		int j = k >> 2;
2090		int spanidx = (k - ycur) >> 2;
2091		int spix = k & 3;
2092		valid_y = !(k < yh \|\| k >= yl);
2093
2094		if (spanidx >= 0 && spanidx < 2048)
2095		{
2096		majorxint[spix] = xend; // 0
2097		minorxint[spix] = xstart; // 319
2098		majorx[spix] = xright; // 0x00000000
2099		minorx[spix] = xleft; // 0x013f0000
2100
2101		if (spix == 0)
2102		{
2103		maxxmx = 0;
2104		minxhx = 0xfff;
2105		}
2106
2107		if (valid_y)
2108		{
2109		maxxmx = (xstart > maxxmx) ? xstart : maxxmx;
2110		minxhx = (xend < minxhx) ? xend : minxhx;
2111		}
2112
2113		if (spix == 0)
2114		{
2115		if(new_object)
2116		{
2117		object = &object_data_alloc();
2118		memcpy(object->m_tmem, m_tmem, 0x1000);
2119		new_object = false;
2120		}
2121
2122		Spans[spanidx].userdata = (void)((UINT8)AuxBuf + AuxBufPtr);
2123		AuxBufPtr += sizeof(rdp_span_aux);
2124
2125		if(AuxBufPtr >= EXTENT_AUX_COUNT)
2126		{
2127		fatalerror("n64_rdp::DrawTriangle: span aux buffer overflow\n");
2128		}
2129
2130		rdp_span_aux userdata = (rdp_span_aux)Spans[spanidx].userdata;
2131		valid = true;
2132
2133		userdata->m_tmem = object->m_tmem;
2134
2135		userdata->BlendColor = BlendColor;
2136		userdata->PrimColor = PrimColor;
2137		userdata->EnvColor = EnvColor;
2138		userdata->FogColor = FogColor;
2139		userdata->KeyScale = KeyScale;
2140		userdata->LODFraction = LODFraction;
2141		userdata->PrimLODFraction = PrimLODFraction;
2142
2143		userdata->ColorInputs.combiner_rgbsub_a_r[0] = userdata->ColorInputs.combiner_rgbsub_a_r[1] = &OneColor.i.r;
2144		userdata->ColorInputs.combiner_rgbsub_a_g[0] = userdata->ColorInputs.combiner_rgbsub_a_g[1] = &OneColor.i.g;
2145		userdata->ColorInputs.combiner_rgbsub_a_b[0] = userdata->ColorInputs.combiner_rgbsub_a_b[1] = &OneColor.i.b;
2146		userdata->ColorInputs.combiner_rgbsub_b_r[0] = userdata->ColorInputs.combiner_rgbsub_b_r[1] = &OneColor.i.r;
2147		userdata->ColorInputs.combiner_rgbsub_b_g[0] = userdata->ColorInputs.combiner_rgbsub_b_g[1] = &OneColor.i.g;
2148		userdata->ColorInputs.combiner_rgbsub_b_b[0] = userdata->ColorInputs.combiner_rgbsub_b_b[1] = &OneColor.i.b;
2149		userdata->ColorInputs.combiner_rgbmul_r[0] = userdata->ColorInputs.combiner_rgbmul_r[1] = &OneColor.i.r;
2150		userdata->ColorInputs.combiner_rgbmul_g[0] = userdata->ColorInputs.combiner_rgbmul_g[1] = &OneColor.i.g;
2151		userdata->ColorInputs.combiner_rgbmul_b[0] = userdata->ColorInputs.combiner_rgbmul_b[1] = &OneColor.i.b;
2152		userdata->ColorInputs.combiner_rgbadd_r[0] = userdata->ColorInputs.combiner_rgbadd_r[1] = &OneColor.i.r;
2153		userdata->ColorInputs.combiner_rgbadd_g[0] = userdata->ColorInputs.combiner_rgbadd_g[1] = &OneColor.i.g;
2154		userdata->ColorInputs.combiner_rgbadd_b[0] = userdata->ColorInputs.combiner_rgbadd_b[1] = &OneColor.i.b;
2155		userdata->ColorInputs.combiner_alphasub_a[0] = userdata->ColorInputs.combiner_alphasub_a[1] = &OneColor.i.a;
2156		userdata->ColorInputs.combiner_alphasub_b[0] = userdata->ColorInputs.combiner_alphasub_b[1] = &OneColor.i.a;
2157		userdata->ColorInputs.combiner_alphamul[0] = userdata->ColorInputs.combiner_alphamul[1] = &OneColor.i.a;
2158		userdata->ColorInputs.combiner_alphaadd[0] = userdata->ColorInputs.combiner_alphaadd[1] = &OneColor.i.a;
2159
2160		userdata->ColorInputs.blender1a_r[0] = userdata->ColorInputs.blender1a_r[1] = &userdata->PixelColor.i.r;
2161		userdata->ColorInputs.blender1a_g[0] = userdata->ColorInputs.blender1a_g[1] = &userdata->PixelColor.i.g;
2162		userdata->ColorInputs.blender1a_b[0] = userdata->ColorInputs.blender1a_b[1] = &userdata->PixelColor.i.b;
2163		userdata->ColorInputs.blender1b_a[0] = userdata->ColorInputs.blender1b_a[1] = &userdata->PixelColor.i.a;
2164		userdata->ColorInputs.blender2a_r[0] = userdata->ColorInputs.blender2a_r[1] = &userdata->PixelColor.i.r;
2165		userdata->ColorInputs.blender2a_g[0] = userdata->ColorInputs.blender2a_g[1] = &userdata->PixelColor.i.g;
2166		userdata->ColorInputs.blender2a_b[0] = userdata->ColorInputs.blender2a_b[1] = &userdata->PixelColor.i.b;
2167		userdata->ColorInputs.blender2b_a[0] = userdata->ColorInputs.blender2b_a[1] = &userdata->PixelColor.i.a;
2168
2169		// Setup blender data for this scanline
2170		SetBlenderInput(0, 0, &userdata->ColorInputs.blender1a_r[0],
2171		&userdata->ColorInputs.blender1a_g[0],
2172		&userdata->ColorInputs.blender1a_b[0],
2173		&userdata->ColorInputs.blender1b_a[0], OtherModes.blend_m1a_0, OtherModes.blend_m1b_0, userdata);
2174		SetBlenderInput(0, 1, &userdata->ColorInputs.blender2a_r[0],
2175		&userdata->ColorInputs.blender2a_g[0],
2176		&userdata->ColorInputs.blender2a_b[0],
2177		&userdata->ColorInputs.blender2b_a[0], OtherModes.blend_m2a_0, OtherModes.blend_m2b_0, userdata);
2178		SetBlenderInput(1, 0, &userdata->ColorInputs.blender1a_r[1],
2179		&userdata->ColorInputs.blender1a_g[1],
2180		&userdata->ColorInputs.blender1a_b[1],
2181		&userdata->ColorInputs.blender1b_a[1], OtherModes.blend_m1a_1, OtherModes.blend_m1b_1, userdata);
2182		SetBlenderInput(1, 1, &userdata->ColorInputs.blender2a_r[1],
2183		&userdata->ColorInputs.blender2a_g[1],
2184		&userdata->ColorInputs.blender2a_b[1],
2185		&userdata->ColorInputs.blender2b_a[1], OtherModes.blend_m2a_1, OtherModes.blend_m2b_1, userdata);
2186
2187		// Setup color combiner data for this scanline
2188		SetSubAInputRGB(&userdata->ColorInputs.combiner_rgbsub_a_r[0], &userdata->ColorInputs.combiner_rgbsub_a_g[0], &userdata->ColorInputs.combiner_rgbsub_a_b[0], m_combine.sub_a_rgb0, userdata);
2189		SetSubBInputRGB(&userdata->ColorInputs.combiner_rgbsub_b_r[0], &userdata->ColorInputs.combiner_rgbsub_b_g[0], &userdata->ColorInputs.combiner_rgbsub_b_b[0], m_combine.sub_b_rgb0, userdata);
2190		SetMulInputRGB(&userdata->ColorInputs.combiner_rgbmul_r[0], &userdata->ColorInputs.combiner_rgbmul_g[0], &userdata->ColorInputs.combiner_rgbmul_b[0], m_combine.mul_rgb0, userdata);
2191		SetAddInputRGB(&userdata->ColorInputs.combiner_rgbadd_r[0], &userdata->ColorInputs.combiner_rgbadd_g[0], &userdata->ColorInputs.combiner_rgbadd_b[0], m_combine.add_rgb0, userdata);
2192		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_a[0], m_combine.sub_a_a0, userdata);
2193		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_b[0], m_combine.sub_b_a0, userdata);
2194		SetMulInputAlpha(&userdata->ColorInputs.combiner_alphamul[0], m_combine.mul_a0, userdata);
2195		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphaadd[0], m_combine.add_a0, userdata);
2196
2197		SetSubAInputRGB(&userdata->ColorInputs.combiner_rgbsub_a_r[1], &userdata->ColorInputs.combiner_rgbsub_a_g[1], &userdata->ColorInputs.combiner_rgbsub_a_b[1], m_combine.sub_a_rgb1, userdata);
2198		SetSubBInputRGB(&userdata->ColorInputs.combiner_rgbsub_b_r[1], &userdata->ColorInputs.combiner_rgbsub_b_g[1], &userdata->ColorInputs.combiner_rgbsub_b_b[1], m_combine.sub_b_rgb1, userdata);
2199		SetMulInputRGB(&userdata->ColorInputs.combiner_rgbmul_r[1], &userdata->ColorInputs.combiner_rgbmul_g[1], &userdata->ColorInputs.combiner_rgbmul_b[1], m_combine.mul_rgb1, userdata);
2200		SetAddInputRGB(&userdata->ColorInputs.combiner_rgbadd_r[1], &userdata->ColorInputs.combiner_rgbadd_g[1], &userdata->ColorInputs.combiner_rgbadd_b[1], m_combine.add_rgb1, userdata);
2201		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_a[1], m_combine.sub_a_a1, userdata);
2202		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_b[1], m_combine.sub_b_a1, userdata);
2203		SetMulInputAlpha(&userdata->ColorInputs.combiner_alphamul[1], m_combine.mul_a1, userdata);
2204		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphaadd[1], m_combine.add_a1, userdata);
2205		}
2206
2207		if (spix == 3)
2208		{
2209		Spans[spanidx].startx = maxxmx;
2210		Spans[spanidx].stopx = minxhx;
2211		compute_cvg_flip(Spans, majorx, minorx, majorxint, minorxint, j, yh, yl, ycur >> 2);
2212		}
2213
2214		if (spix == ldflag)
2215		{
2216		((rdp_span_aux*)Spans[spanidx].userdata)->m_unscissored_rx = xend;
2217		xfrac = ((xright >> 8) & 0xff);
2218		Spans[spanidx].param[SPAN_R].start = ((r >> 9) << 9) + drdiff - (xfrac * drdxh);
2219		Spans[spanidx].param[SPAN_G].start = ((g >> 9) << 9) + dgdiff - (xfrac * dgdxh);
2220		Spans[spanidx].param[SPAN_B].start = ((b >> 9) << 9) + dbdiff - (xfrac * dbdxh);
2221		Spans[spanidx].param[SPAN_A].start = ((a >> 9) << 9) + dadiff - (xfrac * dadxh);
2222		Spans[spanidx].param[SPAN_S].start = (((s >> 9) << 9) + dsdiff - (xfrac * dsdxh)) & ~0x1f;
2223		Spans[spanidx].param[SPAN_T].start = (((t >> 9) << 9) + dtdiff - (xfrac * dtdxh)) & ~0x1f;
2224		Spans[spanidx].param[SPAN_W].start = (((w >> 9) << 9) + dwdiff - (xfrac * dwdxh)) & ~0x1f;
2225		Spans[spanidx].param[SPAN_Z].start = ((z >> 9) << 9) + dzdiff - (xfrac * dzdxh);
2226		}
2227		}
2228
2229		if (spix == 3)
2230		{
2231		r += drde;
2232		g += dgde;
2233		b += dbde;
2234		a += dade;
2235		s += dsde;
2236		t += dtde;
2237		w += dwde;
2238		z += dzde;
2239		}
2240
2241		xleft += xleft_inc;
2242		xright += xright_inc;
2243		}
2244		}
2245		else
2246		{
2247		for (int k = ycur; k <= ylfar; k++)
2248		{
2249		if (k == ym)
2250		{
2251		xleft = xl & ~1;
2252		xleft_inc = (dxldy >> 2) & ~1;
2253		}
2254
2255		int xstart = xleft >> 16;
2256		int xend = xright >> 16;
2257		int j = k >> 2;
2258		int spanidx = j - (ycur >> 2);
2259		int spix = k & 3;
2260		valid_y = !(k < yh \|\| k >= yl);
2261
2262		if (k >= 0 && k < 0x1000)
2263		{
2264		majorxint[spix] = xend;
2265		minorxint[spix] = xstart;
2266		majorx[spix] = xright;
2267		minorx[spix] = xleft;
2268
2269		if (spix == 0)
2270		{
2271		maxxhx = 0;
2272		minxmx = 0xfff;
2273		}
2274
2275		if (valid_y)
2276		{
2277		minxmx = (xstart < minxmx) ? xstart : minxmx;
2278		maxxhx = (xend > maxxhx) ? xend : maxxhx;
2279		}
2280
2281		if (spix == 0)
2282		{
2283		if(new_object)
2284		{
2285		object = &object_data_alloc();
2286		memcpy(object->m_tmem, m_tmem, 0x1000);
2287		new_object = false;
2288		}
2289
2290		Spans[spanidx].userdata = (void)((UINT8)AuxBuf + AuxBufPtr);
2291		valid = true;
2292		AuxBufPtr += sizeof(rdp_span_aux);
2293
2294		if(AuxBufPtr >= EXTENT_AUX_COUNT)
2295		{
2296		fatalerror("n64_rdp::DrawTriangle: span aux buffer overflow\n");
2297		}
2298
2299		rdp_span_aux userdata = (rdp_span_aux)Spans[spanidx].userdata;
2300		userdata->m_tmem = object->m_tmem;
2301
2302		userdata->BlendColor = BlendColor;
2303		userdata->PrimColor = PrimColor;
2304		userdata->EnvColor = EnvColor;
2305		userdata->FogColor = FogColor;
2306		userdata->KeyScale = KeyScale;
2307		userdata->LODFraction = LODFraction;
2308		userdata->PrimLODFraction = PrimLODFraction;
2309
2310		userdata->ColorInputs.combiner_rgbsub_a_r[0] = userdata->ColorInputs.combiner_rgbsub_a_r[1] = &OneColor.i.r;
2311		userdata->ColorInputs.combiner_rgbsub_a_g[0] = userdata->ColorInputs.combiner_rgbsub_a_g[1] = &OneColor.i.g;
2312		userdata->ColorInputs.combiner_rgbsub_a_b[0] = userdata->ColorInputs.combiner_rgbsub_a_b[1] = &OneColor.i.b;
2313		userdata->ColorInputs.combiner_rgbsub_b_r[0] = userdata->ColorInputs.combiner_rgbsub_b_r[1] = &OneColor.i.r;
2314		userdata->ColorInputs.combiner_rgbsub_b_g[0] = userdata->ColorInputs.combiner_rgbsub_b_g[1] = &OneColor.i.g;
2315		userdata->ColorInputs.combiner_rgbsub_b_b[0] = userdata->ColorInputs.combiner_rgbsub_b_b[1] = &OneColor.i.b;
2316		userdata->ColorInputs.combiner_rgbmul_r[0] = userdata->ColorInputs.combiner_rgbmul_r[1] = &OneColor.i.r;
2317		userdata->ColorInputs.combiner_rgbmul_g[0] = userdata->ColorInputs.combiner_rgbmul_g[1] = &OneColor.i.g;
2318		userdata->ColorInputs.combiner_rgbmul_b[0] = userdata->ColorInputs.combiner_rgbmul_b[1] = &OneColor.i.b;
2319		userdata->ColorInputs.combiner_rgbadd_r[0] = userdata->ColorInputs.combiner_rgbadd_r[1] = &OneColor.i.r;
2320		userdata->ColorInputs.combiner_rgbadd_g[0] = userdata->ColorInputs.combiner_rgbadd_g[1] = &OneColor.i.g;
2321		userdata->ColorInputs.combiner_rgbadd_b[0] = userdata->ColorInputs.combiner_rgbadd_b[1] = &OneColor.i.b;
2322		userdata->ColorInputs.combiner_alphasub_a[0] = userdata->ColorInputs.combiner_alphasub_a[1] = &OneColor.i.a;
2323		userdata->ColorInputs.combiner_alphasub_b[0] = userdata->ColorInputs.combiner_alphasub_b[1] = &OneColor.i.a;
2324		userdata->ColorInputs.combiner_alphamul[0] = userdata->ColorInputs.combiner_alphamul[1] = &OneColor.i.a;
2325		userdata->ColorInputs.combiner_alphaadd[0] = userdata->ColorInputs.combiner_alphaadd[1] = &OneColor.i.a;
2326
2327		userdata->ColorInputs.blender1a_r[0] = userdata->ColorInputs.blender1a_r[1] = &userdata->PixelColor.i.r;
2328		userdata->ColorInputs.blender1a_g[0] = userdata->ColorInputs.blender1a_g[1] = &userdata->PixelColor.i.g;
2329		userdata->ColorInputs.blender1a_b[0] = userdata->ColorInputs.blender1a_b[1] = &userdata->PixelColor.i.b;
2330		userdata->ColorInputs.blender1b_a[0] = userdata->ColorInputs.blender1b_a[1] = &userdata->PixelColor.i.a;
2331		userdata->ColorInputs.blender2a_r[0] = userdata->ColorInputs.blender2a_r[1] = &userdata->PixelColor.i.r;
2332		userdata->ColorInputs.blender2a_g[0] = userdata->ColorInputs.blender2a_g[1] = &userdata->PixelColor.i.g;
2333		userdata->ColorInputs.blender2a_b[0] = userdata->ColorInputs.blender2a_b[1] = &userdata->PixelColor.i.b;
2334		userdata->ColorInputs.blender2b_a[0] = userdata->ColorInputs.blender2b_a[1] = &userdata->PixelColor.i.a;
2335
2336		// Setup blender data for this scanline
2337		SetBlenderInput(0, 0, &userdata->ColorInputs.blender1a_r[0],
2338		&userdata->ColorInputs.blender1a_g[0],
2339		&userdata->ColorInputs.blender1a_b[0],
2340		&userdata->ColorInputs.blender1b_a[0], OtherModes.blend_m1a_0, OtherModes.blend_m1b_0, userdata);
2341		SetBlenderInput(0, 1, &userdata->ColorInputs.blender2a_r[0],
2342		&userdata->ColorInputs.blender2a_g[0],
2343		&userdata->ColorInputs.blender2a_b[0],
2344		&userdata->ColorInputs.blender2b_a[0], OtherModes.blend_m2a_0, OtherModes.blend_m2b_0, userdata);
2345		SetBlenderInput(1, 0, &userdata->ColorInputs.blender1a_r[1],
2346		&userdata->ColorInputs.blender1a_g[1],
2347		&userdata->ColorInputs.blender1a_b[1],
2348		&userdata->ColorInputs.blender1b_a[1], OtherModes.blend_m1a_1, OtherModes.blend_m1b_1, userdata);
2349		SetBlenderInput(1, 1, &userdata->ColorInputs.blender2a_r[1],
2350		&userdata->ColorInputs.blender2a_g[1],
2351		&userdata->ColorInputs.blender2a_b[1],
2352		&userdata->ColorInputs.blender2b_a[1], OtherModes.blend_m2a_1, OtherModes.blend_m2b_1, userdata);
2353
2354		// Setup color combiner data for this scanline
2355		SetSubAInputRGB(&userdata->ColorInputs.combiner_rgbsub_a_r[0], &userdata->ColorInputs.combiner_rgbsub_a_g[0], &userdata->ColorInputs.combiner_rgbsub_a_b[0], m_combine.sub_a_rgb0, userdata);
2356		SetSubBInputRGB(&userdata->ColorInputs.combiner_rgbsub_b_r[0], &userdata->ColorInputs.combiner_rgbsub_b_g[0], &userdata->ColorInputs.combiner_rgbsub_b_b[0], m_combine.sub_b_rgb0, userdata);
2357		SetMulInputRGB(&userdata->ColorInputs.combiner_rgbmul_r[0], &userdata->ColorInputs.combiner_rgbmul_g[0], &userdata->ColorInputs.combiner_rgbmul_b[0], m_combine.mul_rgb0, userdata);
2358		SetAddInputRGB(&userdata->ColorInputs.combiner_rgbadd_r[0], &userdata->ColorInputs.combiner_rgbadd_g[0], &userdata->ColorInputs.combiner_rgbadd_b[0], m_combine.add_rgb0, userdata);
2359		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_a[0], m_combine.sub_a_a0, userdata);
2360		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_b[0], m_combine.sub_b_a0, userdata);
2361		SetMulInputAlpha(&userdata->ColorInputs.combiner_alphamul[0], m_combine.mul_a0, userdata);
2362		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphaadd[0], m_combine.add_a0, userdata);
2363
2364		SetSubAInputRGB(&userdata->ColorInputs.combiner_rgbsub_a_r[1], &userdata->ColorInputs.combiner_rgbsub_a_g[1], &userdata->ColorInputs.combiner_rgbsub_a_b[1], m_combine.sub_a_rgb1, userdata);
2365		SetSubBInputRGB(&userdata->ColorInputs.combiner_rgbsub_b_r[1], &userdata->ColorInputs.combiner_rgbsub_b_g[1], &userdata->ColorInputs.combiner_rgbsub_b_b[1], m_combine.sub_b_rgb1, userdata);
2366		SetMulInputRGB(&userdata->ColorInputs.combiner_rgbmul_r[1], &userdata->ColorInputs.combiner_rgbmul_g[1], &userdata->ColorInputs.combiner_rgbmul_b[1], m_combine.mul_rgb1, userdata);
2367		SetAddInputRGB(&userdata->ColorInputs.combiner_rgbadd_r[1], &userdata->ColorInputs.combiner_rgbadd_g[1], &userdata->ColorInputs.combiner_rgbadd_b[1], m_combine.add_rgb1, userdata);
2368		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_a[1], m_combine.sub_a_a1, userdata);
2369		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_b[1], m_combine.sub_b_a1, userdata);
2370		SetMulInputAlpha(&userdata->ColorInputs.combiner_alphamul[1], m_combine.mul_a1, userdata);
2371		SetSubInputAlpha(&userdata->ColorInputs.combiner_alphaadd[1], m_combine.add_a1, userdata);
2372		}
2373
2374		if (spix == 3)
2375		{
2376		Spans[spanidx].startx = minxmx;
2377		Spans[spanidx].stopx = maxxhx;
2378		compute_cvg_noflip(Spans, majorx, minorx, majorxint, minorxint, j, yh, yl, ycur >> 2);
2379		}
2380
2381		if (spix == ldflag)
2382		{
2383		((rdp_span_aux*)Spans[spanidx].userdata)->m_unscissored_rx = xend;
2384		xfrac = ((xright >> 8) & 0xff);
2385		Spans[spanidx].param[SPAN_R].start = ((r >> 9) << 9) + drdiff - (xfrac * drdxh);
2386		Spans[spanidx].param[SPAN_G].start = ((g >> 9) << 9) + dgdiff - (xfrac * dgdxh);
2387		Spans[spanidx].param[SPAN_B].start = ((b >> 9) << 9) + dbdiff - (xfrac * dbdxh);
2388		Spans[spanidx].param[SPAN_A].start = ((a >> 9) << 9) + dadiff - (xfrac * dadxh);
2389		Spans[spanidx].param[SPAN_S].start = (((s >> 9) << 9) + dsdiff - (xfrac * dsdxh)) & ~0x1f;
2390		Spans[spanidx].param[SPAN_T].start = (((t >> 9) << 9) + dtdiff - (xfrac * dtdxh)) & ~0x1f;
2391		Spans[spanidx].param[SPAN_W].start = (((w >> 9) << 9) + dwdiff - (xfrac * dwdxh)) & ~0x1f;
2392		Spans[spanidx].param[SPAN_Z].start = ((z >> 9) << 9) + dzdiff - (xfrac * dzdxh);
2393		}
2394		}
2395
2396		if (spix == 3)
2397		{
2398		r += drde;
2399		g += dgde;
2400		b += dbde;
2401		a += dade;
2402		s += dsde;
2403		t += dtde;
2404		w += dwde;
2405		z += dzde;
2406		}
2407		xleft += xleft_inc;
2408		xright += xright_inc;
2409		}
2410		}
2411
2412		if(!new_object && valid)
2413		{
2414		RenderSpans(yh >> 2, yl >> 2, tilenum, flip ? true : false, Spans, rect, object);
2415		}
2416
2417		//wait("DrawTriangle");
2418		}
2419
2420		/*****************************************************************************/
2421
2422		////////////////////////
2423		// FB ACCESSORS
2424		////////////////////////
2425		void n64_rdp::_Write16Bit_Cvg0_Blend(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2426		{
2427		UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
2428		UINT32 hb = fb;
2429
2430		UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
2431		UINT32 finalcvg = 0;
2432
2433		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2434		{
2435		finalcolor = RREADIDX16(fb) & 0xfffe;
2436		CHECK16(fb);
2437		}
2438
2439		finalcvg = userdata->CurrentPixCvg + userdata->CurrentMemCvg;
2440		if (finalcvg & 8)
2441		{
2442		finalcvg = 7;
2443		}
2444		RWRITEIDX16(fb, finalcolor \| ((finalcvg >> 2) & 1));
2445		HWRITEADDR8(hb, finalcvg & 3);
2446		}
2447
2448		void n64_rdp::_Write16Bit_Cvg0_NoBlend(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2449		{
2450		UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
2451		UINT32 hb = fb;
2452
2453		UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
2454		UINT32 finalcvg = 0;
2455
2456		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2457		{
2458		finalcolor = RREADIDX16(fb) & 0xfffe;
2459		CHECK16(fb);
2460		}
2461
2462		finalcvg = (userdata->CurrentPixCvg - 1) & 7;
2463		RWRITEIDX16(fb, finalcolor \| ((finalcvg >> 2) & 1));
2464		HWRITEADDR8(hb, finalcvg & 3);
2465		}
2466
2467		void n64_rdp::_Write16Bit_Cvg1(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2468		{
2469		UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
2470		UINT32 hb = fb;
2471
2472		UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
2473		UINT32 finalcvg = 0;
2474
2475		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2476		{
2477		finalcolor = RREADIDX16(fb) & 0xfffe;
2478		CHECK16(fb);
2479		}
2480
2481		finalcvg = (userdata->CurrentPixCvg + userdata->CurrentMemCvg) & 7;
2482		RWRITEIDX16(fb, finalcolor \| ((finalcvg >> 2) & 1));
2483		HWRITEADDR8(hb, finalcvg & 3);
2484		}
2485
2486		void n64_rdp::_Write16Bit_Cvg2(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2487		{
2488		UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
2489		UINT32 hb = fb;
2490
2491		UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
2492
2493		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2494		{
2495		finalcolor = RREADIDX16(fb) & 0xfffe;
2496		CHECK16(fb);
2497		}
2498
2499		RWRITEIDX16(fb, finalcolor \| 1);
2500		HWRITEADDR8(hb, 3);
2501		}
2502
2503		void n64_rdp::_Write16Bit_Cvg3(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2504		{
2505		UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
2506		UINT32 hb = fb;
2507
2508		UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
2509
2510		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2511		{
2512		finalcolor = RREADIDX16(fb) & 0xfffe;
2513		CHECK16(fb);
2514		}
2515
2516		RWRITEIDX16(fb, finalcolor \| ((userdata->CurrentMemCvg >> 2) & 1));
2517		HWRITEADDR8(hb, userdata->CurrentMemCvg & 3);
2518		}
2519
2520		void n64_rdp::_Write32Bit_Cvg0_Blend(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2521		{
2522		UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
2523		UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
2524		UINT32 finalcvg = 0;
2525
2526		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2527		{
2528		finalcolor = RREADIDX32(fb) & 0xffffff00;
2529		CHECK32(fb);
2530		}
2531
2532		finalcvg = userdata->CurrentPixCvg + userdata->CurrentMemCvg;
2533		if (finalcvg & 8)
2534		{
2535		finalcvg = 7;
2536		}
2537		RWRITEIDX32(fb, finalcolor \| (finalcvg << 5));
2538		}
2539
2540		void n64_rdp::_Write32Bit_Cvg0_NoBlend(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2541		{
2542		UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
2543		UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
2544		UINT32 finalcvg = 0;
2545
2546		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2547		{
2548		finalcolor = RREADIDX32(fb) & 0xffffff00;
2549		CHECK32(fb);
2550		}
2551
2552		finalcvg = (userdata->CurrentPixCvg - 1) & 7;
2553		RWRITEIDX32(fb, finalcolor \| (finalcvg << 5));
2554		}
2555
2556		void n64_rdp::_Write32Bit_Cvg1(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2557		{
2558		UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
2559		UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
2560		UINT32 finalcvg = 0;
2561
2562		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2563		{
2564		finalcolor = RREADIDX32(fb) & 0xffffff00;
2565		CHECK32(fb);
2566		}
2567
2568		finalcvg = (userdata->CurrentPixCvg + userdata->CurrentMemCvg) & 7;
2569		finalcolor \|= (finalcvg << 5);
2570		RWRITEIDX32(fb, finalcolor);
2571		}
2572
2573		void n64_rdp::_Write32Bit_Cvg2(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2574		{
2575		UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
2576		UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
2577
2578		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2579		{
2580		finalcolor = RREADIDX32(fb) & 0xffffff00;
2581		CHECK32(fb);
2582		}
2583
2584		RWRITEIDX32(fb, finalcolor \| 0xE0);
2585		}
2586
2587		void n64_rdp::_Write32Bit_Cvg3(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
2588		{
2589		UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
2590		UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
2591
2592		if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
2593		{
2594		finalcolor = RREADIDX32(fb) & 0xffffff00;
2595		CHECK32(fb);
2596		}
2597
2598		RWRITEIDX32(fb, finalcolor \| (userdata->CurrentMemCvg << 5));
2599		}
2600
2601
2602		void n64_rdp::_Read16Bit_ImgRead0(UINT32 curpixel, rdp_span_aux *userdata, const rdp_poly_state &object)
2603		{
2604		UINT16 fword = RREADIDX16((object.MiscState.FBAddress >> 1) + curpixel);
2605		CHECK16((object.MiscState.FBAddress >> 1) + curpixel);
2606		userdata->MemoryColor.i.r = GETHICOL(fword);
2607		userdata->MemoryColor.i.g = GETMEDCOL(fword);
2608		userdata->MemoryColor.i.b = GETLOWCOL(fword);
2609		userdata->MemoryColor.i.a = 0xff;
2610		userdata->CurrentMemCvg = 7;
2611		}
2612
2613		void n64_rdp::_Read16Bit_ImgRead1(UINT32 curpixel, rdp_span_aux *userdata, const rdp_poly_state &object)
2614		{
2615		UINT16 fword = RREADIDX16((object.MiscState.FBAddress >> 1) + curpixel);
2616		CHECK16((object.MiscState.FBAddress >> 1) + curpixel);
2617		UINT8 hbyte = HREADADDR8((object.MiscState.FBAddress >> 1) + curpixel);
2618		userdata->MemoryColor.i.r = GETHICOL(fword);
2619		userdata->MemoryColor.i.g = GETMEDCOL(fword);
2620		userdata->MemoryColor.i.b = GETLOWCOL(fword);
2621		userdata->MemoryColor.i.a = userdata->CurrentMemCvg << 5;
2622		userdata->CurrentMemCvg = ((fword & 1) << 2) \| (hbyte & 3);
2623		}
2624
2625		void n64_rdp::_Read32Bit_ImgRead0(UINT32 curpixel, rdp_span_aux *userdata, const rdp_poly_state &object)
2626		{
2627		UINT32 mem = RREADIDX32((object.MiscState.FBAddress >> 2) + curpixel);
2628		CHECK32((object.MiscState.FBAddress >> 2) + curpixel);
2629		userdata->MemoryColor.i.r = (mem >> 24) & 0xff;
2630		userdata->MemoryColor.i.g = (mem >> 16) & 0xff;
2631		userdata->MemoryColor.i.b = (mem >> 8) & 0xff;
2632		userdata->MemoryColor.i.a = 0xff;
2633		userdata->CurrentMemCvg = 7;
2634		}
2635
2636		void n64_rdp::_Read32Bit_ImgRead1(UINT32 curpixel, rdp_span_aux *userdata, const rdp_poly_state &object)
2637		{
2638		UINT32 mem = RREADIDX32((object.MiscState.FBAddress >> 2) + curpixel);
2639		CHECK32((object.MiscState.FBAddress >> 2) + curpixel);
2640		userdata->MemoryColor.i.r = (mem >> 24) & 0xff;
2641		userdata->MemoryColor.i.g = (mem >> 16) & 0xff;
2642		userdata->MemoryColor.i.b = (mem >> 8) & 0xff;
2643		userdata->MemoryColor.i.a = (mem) & 0xff;
2644		userdata->CurrentMemCvg = (mem >> 5) & 7;
2645		}
2646
2647		void n64_rdp::_Copy16Bit(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, int CurrentPixCvg, const rdp_poly_state &object)
2648		{
2649		UINT16 val = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1) \| ((CurrentPixCvg >> 2) & 1);
2650		RWRITEIDX16((object.MiscState.FBAddress >> 1) + curpixel, val);
2651		HWRITEADDR8((object.MiscState.FBAddress >> 1) + curpixel, CurrentPixCvg & 3);
2652		CHECK16((object.MiscState.FBAddress >> 1) + curpixel);
2653		}
2654
2655		void n64_rdp::_Copy32Bit(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, int CurrentPixCvg, const rdp_poly_state &object)
2656		{
2657		UINT32 val = (r << 24) \| (g << 16) \| (b << 8) \| (CurrentPixCvg << 5);
2658		RWRITEIDX32((object.MiscState.FBAddress >> 2) + curpixel, val);
2659		CHECK32((object.MiscState.FBAddress >> 2) + curpixel);
2660		}
2661
2662
2663		void n64_rdp::_Fill16Bit(UINT32 curpixel, const rdp_poly_state &object)
2664		{
2665		UINT16 val;
2666		if (curpixel & 1)
2667		{
2668		val = object.FillColor & 0xffff;
2669		}
2670		else
2671		{
2672		val = (object.FillColor >> 16) & 0xffff;
2673		}
2674		RWRITEIDX16((object.MiscState.FBAddress >> 1) + curpixel, val);
2675		CHECK16((object.MiscState.FBAddress >> 1) + curpixel);
2676		HWRITEADDR8((object.MiscState.FBAddress >> 1) + curpixel, ((val & 1) << 1) \| (val & 1));
2677		}
2678
2679		void n64_rdp::_Fill32Bit(UINT32 curpixel, const rdp_poly_state &object)
2680		{
2681		UINT32 FillColor = object.FillColor;
2682		RWRITEIDX32((object.MiscState.FBAddress >> 2) + curpixel, FillColor);
2683		CHECK32((object.MiscState.FBAddress >> 2) + curpixel);
2684		HWRITEADDR8((object.MiscState.FBAddress >> 1) + (curpixel << 1), (FillColor & 0x10000) ? 3 : 0);
2685		HWRITEADDR8((object.MiscState.FBAddress >> 1) + (curpixel << 1) + 1, (FillColor & 0x1) ? 3 : 0);
2686		}
2687
2688		////////////////////////
2689		// RDP COMMANDS
2690		////////////////////////
2691		void n64_rdp::Triangle(bool shade, bool texture, bool zbuffer)
2692		{
2693		DrawTriangle(shade, texture, zbuffer, false);
2694		m_pipe_clean = false;
2695		//wait();
2696		}
2697
2698		void n64_rdp::CmdTriangle(UINT32 w1, UINT32 w2)
2699		{
2700		Triangle(false, false, false);
2701		}
2702
2703		void n64_rdp::CmdTriangleZ(UINT32 w1, UINT32 w2)
2704		{
2705		Triangle(false, false, true);
2706		}
2707
2708		void n64_rdp::CmdTriangleT(UINT32 w1, UINT32 w2)
2709		{
2710		Triangle(false, true, false);
2711		}
2712
2713		void n64_rdp::CmdTriangleTZ(UINT32 w1, UINT32 w2)
2714		{
2715		Triangle(false, true, true);
2716		}
2717
2718		void n64_rdp::CmdTriangleS(UINT32 w1, UINT32 w2)
2719		{
2720		Triangle(true, false, false);
2721		}
2722
2723		void n64_rdp::CmdTriangleSZ(UINT32 w1, UINT32 w2)
2724		{
2725		Triangle(true, false, true);
2726		}
2727
2728		void n64_rdp::CmdTriangleST(UINT32 w1, UINT32 w2)
2729		{
2730		Triangle(true, true, false);
2731		}
2732
2733		void n64_rdp::CmdTriangleSTZ(UINT32 w1, UINT32 w2)
2734		{
2735		Triangle(true, true, true);
2736		}
2737
2738		void n64_rdp::CmdTexRect(UINT32 w1, UINT32 w2)
2739		{
2740		//if(m_pending_mode_block) { wait("Block on pending mode-change"); m_pending_mode_block = false; }
2741
2742		UINT32 *data = m_cmd_data + m_cmd_cur;
2743
2744		UINT32 w3 = data[2];
2745		UINT32 w4 = data[3];
2746
2747		UINT32 tilenum = (w2 >> 24) & 0x7;
2748		UINT32 xl = (w1 >> 12) & 0xfff;
2749		UINT32 yl = (w1 >> 0) & 0xfff;
2750		UINT32 xh = (w2 >> 12) & 0xfff;
2751		UINT32 yh = (w2 >> 0) & 0xfff;
2752		INT32 s = (w3 >> 16) & 0xffff;
2753		INT32 t = (w3 >> 0) & 0xffff;
2754		INT32 dsdx = (w4 >> 16) & 0xffff;
2755		INT32 dtdy = (w4 >> 0) & 0xffff;
2756
2757		dsdx = SIGN16(dsdx);
2758		dtdy = SIGN16(dtdy);
2759
2760		if (OtherModes.cycle_type == CYCLE_TYPE_FILL \|\| OtherModes.cycle_type == CYCLE_TYPE_COPY)
2761		{
2762		yl \|= 3;
2763		}
2764
2765		UINT32 xlint = (xl >> 2) & 0x3ff;
2766		UINT32 xhint = (xh >> 2) & 0x3ff;
2767
2768		UINT32* ewdata = m_temp_rect_data;
2769		ewdata[0] = (0x24 << 24) \| ((0x80 \| tilenum) << 16) \| yl; // command, flipped, tile, yl
2770		ewdata[1] = (yl << 16) \| yh; // ym, yh
2771		ewdata[2] = (xlint << 16) \| ((xl & 3) << 14); // xl, xl frac
2772		ewdata[3] = 0; // dxldy, dxldy frac
2773		ewdata[4] = (xhint << 16) \| ((xh & 3) << 14); // xh, xh frac
2774		ewdata[5] = 0; // dxhdy, dxhdy frac
2775		ewdata[6] = (xlint << 16) \| ((xl & 3) << 14); // xm, xm frac
2776		ewdata[7] = 0; // dxmdy, dxmdy frac
2777		memset(&ewdata[8], 0, 16 * sizeof(UINT32)); // shade
2778		ewdata[24] = (s << 16) \| t; // s, t
2779		ewdata[25] = 0; // w
2780		ewdata[26] = ((dsdx >> 5) << 16); // dsdx, dtdx
2781		ewdata[27] = 0; // dwdx
2782		ewdata[28] = 0; // s frac, t frac
2783		ewdata[29] = 0; // w frac
2784		ewdata[30] = ((dsdx & 0x1f) << 11) << 16; // dsdx frac, dtdx frac
2785		ewdata[31] = 0; // dwdx frac
2786		ewdata[32] = (dtdy >> 5) & 0xffff;//dsde, dtde
2787		ewdata[33] = 0;//dwde
2788		ewdata[34] = (dtdy >> 5) & 0xffff;//dsdy, dtdy
2789		ewdata[35] = 0;//dwdy
2790		ewdata[36] = (dtdy & 0x1f) << 11;//dsde frac, dtde frac
2791		ewdata[37] = 0;//dwde frac
2792		ewdata[38] = (dtdy & 0x1f) << 11;//dsdy frac, dtdy frac
2793		ewdata[39] = 0;//dwdy frac
2794		memset(&ewdata[40], 0, 4 * sizeof(UINT32));//depth
2795
2796		DrawTriangle(true, true, false, true);
2797		}
2798
2799		void n64_rdp::CmdTexRectFlip(UINT32 w1, UINT32 w2)
2800		{
2801		//if(m_pending_mode_block) { wait("Block on pending mode-change"); m_pending_mode_block = false; }
2802
2803		UINT32 *data = m_cmd_data + m_cmd_cur;
2804
2805		UINT32 w3 = data[2];
2806		UINT32 w4 = data[3];
2807
2808		UINT32 tilenum = (w2 >> 24) & 0x7;
2809		UINT32 xl = (w1 >> 12) & 0xfff;
2810		UINT32 yl = (w1 >> 0) & 0xfff;
2811		UINT32 xh = (w2 >> 12) & 0xfff;
2812		UINT32 yh = (w2 >> 0) & 0xfff;
2813		INT32 s = (w3 >> 16) & 0xffff;
2814		INT32 t = (w3 >> 0) & 0xffff;
2815		INT32 dsdx = (w4 >> 16) & 0xffff;
2816		INT32 dtdy = (w4 >> 0) & 0xffff;
2817
2818		dsdx = SIGN16(dsdx);
2819		dtdy = SIGN16(dtdy);
2820
2821		if (OtherModes.cycle_type == CYCLE_TYPE_FILL \|\| OtherModes.cycle_type == CYCLE_TYPE_COPY)
2822		{
2823		yl \|= 3;
2824		}
2825
2826		UINT32 xlint = (xl >> 2) & 0x3ff;
2827		UINT32 xhint = (xh >> 2) & 0x3ff;
2828
2829		UINT32* ewdata = m_temp_rect_data;
2830		ewdata[0] = (0x25 << 24) \| ((0x80 \| tilenum) << 16) \| yl;//command, flipped, tile, yl
2831		ewdata[1] = (yl << 16) \| yh;//ym, yh
2832		ewdata[2] = (xlint << 16) \| ((xl & 3) << 14);//xl, xl frac
2833		ewdata[3] = 0;//dxldy, dxldy frac
2834		ewdata[4] = (xhint << 16) \| ((xh & 3) << 14);//xh, xh frac
2835		ewdata[5] = 0;//dxhdy, dxhdy frac
2836		ewdata[6] = (xlint << 16) \| ((xl & 3) << 14);//xm, xm frac
2837		ewdata[7] = 0;//dxmdy, dxmdy frac
2838		memset(&ewdata[8], 0, 16 * sizeof(UINT32));//shade
2839		ewdata[24] = (s << 16) \| t;//s, t
2840		ewdata[25] = 0;//w
2841		ewdata[26] = (dtdy >> 5) & 0xffff;//dsdx, dtdx
2842		ewdata[27] = 0;//dwdx
2843		ewdata[28] = 0;//s frac, t frac
2844		ewdata[29] = 0;//w frac
2845		ewdata[30] = ((dtdy & 0x1f) << 11);//dsdx frac, dtdx frac
2846		ewdata[31] = 0;//dwdx frac
2847		ewdata[32] = (dsdx >> 5) << 16;//dsde, dtde
2848		ewdata[33] = 0;//dwde
2849		ewdata[34] = (dsdx >> 5) << 16;//dsdy, dtdy
2850		ewdata[35] = 0;//dwdy
2851		ewdata[36] = (dsdx & 0x1f) << 27;//dsde frac, dtde frac
2852		ewdata[37] = 0;//dwde frac
2853		ewdata[38] = (dsdx & 0x1f) << 27;//dsdy frac, dtdy frac
2854		ewdata[39] = 0;//dwdy frac
2855		memset(&ewdata[40], 0, 4 * sizeof(UINT32));//depth
2856
2857		DrawTriangle(true, true, false, true);
2858		}
2859
2860		void n64_rdp::CmdSyncLoad(UINT32 w1, UINT32 w2)
2861		{
2862		//wait("SyncLoad");
2863		}
2864
2865		void n64_rdp::CmdSyncPipe(UINT32 w1, UINT32 w2)
2866		{
2867		//wait("SyncPipe");
2868		}
2869
2870		void n64_rdp::CmdSyncTile(UINT32 w1, UINT32 w2)
2871		{
2872		//wait("SyncTile");
2873		}
2874
2875		void n64_rdp::CmdSyncFull(UINT32 w1, UINT32 w2)
2876		{
2877		//wait("SyncFull");
2878		dp_full_sync(*m_machine);
2879		}
2880
2881		void n64_rdp::CmdSetKeyGB(UINT32 w1, UINT32 w2)
2882		{
2883		KeyScale.i.b = w2 & 0xff;
2884		KeyScale.i.g = (w2 >> 16) & 0xff;
2885		}
2886
2887		void n64_rdp::CmdSetKeyR(UINT32 w1, UINT32 w2)
2888		{
2889		KeyScale.i.r = w2 & 0xff;
2890		}
2891
2892		void n64_rdp::CmdSetFillColor32(UINT32 w1, UINT32 w2)
2893		{
2894		//wait("SetFillColor");
2895		FillColor = w2;
2896		}
2897
2898		void n64_rdp::CmdSetConvert(UINT32 w1, UINT32 w2)
2899		{
2900		if(!m_pipe_clean) { m_pipe_clean = true; wait("SetConvert"); }
2901		INT32 k0 = (w1 >> 13) & 0xff;
2902		INT32 k1 = (w1 >> 4) & 0xff;
2903		INT32 k2 = ((w1 & 7) << 5) \| ((w2 >> 27) & 0x1f);
2904		INT32 k3 = (w2 >> 18) & 0xff;
2905		INT32 k4 = (w2 >> 9) & 0xff;
2906		INT32 k5 = w2 & 0xff;
2907		k0 = ((w1 >> 21) & 1) ? (-(0x100 - k0)) : k0;
2908		k1 = ((w1 >> 12) & 1) ? (-(0x100 - k1)) : k1;
2909		k2 = (w1 & 0xf) ? (-(0x100 - k2)) : k2;
2910		k3 = ((w2 >> 26) & 1) ? (-(0x100 - k3)) : k3;
2911		k4 = ((w2 >> 17) & 1) ? (-(0x100 - k4)) : k4;
2912		k5 = ((w2 >> 8) & 1) ? (-(0x100 - k5)) : k5;
2913		SetYUVFactors(k0, k1, k2, k3, k4, k5);
2914		}
2915
2916		void n64_rdp::CmdSetScissor(UINT32 w1, UINT32 w2)
2917		{
2918		Scissor.m_xh = ((w1 >> 12) & 0xfff) >> 2;
2919		Scissor.m_yh = ((w1 >> 0) & 0xfff) >> 2;
2920		Scissor.m_xl = ((w2 >> 12) & 0xfff) >> 2;
2921		Scissor.m_yl = ((w2 >> 0) & 0xfff) >> 2;
2922
2923		// TODO: handle f & o?
2924		}
2925
2926		void n64_rdp::CmdSetPrimDepth(UINT32 w1, UINT32 w2)
2927		{
2928		MiscState.PrimitiveZ = (UINT16)(w2 >> 16) & 0x7fff;
2929		MiscState.PrimitiveDZ = (UINT16)(w1);
2930		}
2931
2932		void n64_rdp::CmdSetOtherModes(UINT32 w1, UINT32 w2)
2933		{
2934		//wait("SetOtherModes");
2935		OtherModes.cycle_type = (w1 >> 20) & 0x3; // 01
2936		OtherModes.persp_tex_en = (w1 & 0x80000) ? 1 : 0; // 1
2937		OtherModes.detail_tex_en = (w1 & 0x40000) ? 1 : 0; // 0
2938		OtherModes.sharpen_tex_en = (w1 & 0x20000) ? 1 : 0; // 0
2939		OtherModes.tex_lod_en = (w1 & 0x10000) ? 1 : 0; // 0
2940		OtherModes.en_tlut = (w1 & 0x08000) ? 1 : 0; // 0
2941		OtherModes.tlut_type = (w1 & 0x04000) ? 1 : 0; // 0
2942		OtherModes.sample_type = (w1 & 0x02000) ? 1 : 0; // 1
2943		OtherModes.mid_texel = (w1 & 0x01000) ? 1 : 0; // 0
2944		OtherModes.bi_lerp0 = (w1 & 0x00800) ? 1 : 0; // 1
2945		OtherModes.bi_lerp1 = (w1 & 0x00400) ? 1 : 0; // 1
2946		OtherModes.convert_one = (w1 & 0x00200) ? 1 : 0; // 0
2947		OtherModes.key_en = (w1 & 0x00100) ? 1 : 0; // 0
2948		OtherModes.rgb_dither_sel = (w1 >> 6) & 0x3; // 00
2949		OtherModes.alpha_dither_sel = (w1 >> 4) & 0x3; // 01
2950		OtherModes.blend_m1a_0 = (w2 >> 30) & 0x3; // 11
2951		OtherModes.blend_m1a_1 = (w2 >> 28) & 0x3; // 00
2952		OtherModes.blend_m1b_0 = (w2 >> 26) & 0x3; // 10
2953		OtherModes.blend_m1b_1 = (w2 >> 24) & 0x3; // 00
2954		OtherModes.blend_m2a_0 = (w2 >> 22) & 0x3; // 00
2955		OtherModes.blend_m2a_1 = (w2 >> 20) & 0x3; // 01
2956		OtherModes.blend_m2b_0 = (w2 >> 18) & 0x3; // 00
2957		OtherModes.blend_m2b_1 = (w2 >> 16) & 0x3; // 01
2958		OtherModes.force_blend = (w2 >> 14) & 1; // 0
2959		OtherModes.alpha_cvg_select = (w2 >> 13) & 1; // 1
2960		OtherModes.cvg_times_alpha = (w2 >> 12) & 1; // 0
2961		OtherModes.z_mode = (w2 >> 10) & 0x3; // 00
2962		OtherModes.cvg_dest = (w2 >> 8) & 0x3; // 00
2963		OtherModes.color_on_cvg = (w2 >> 7) & 1; // 0
2964		OtherModes.image_read_en = (w2 >> 6) & 1; // 1
2965		OtherModes.z_update_en = (w2 >> 5) & 1; // 1
2966		OtherModes.z_compare_en = (w2 >> 4) & 1; // 1
2967		OtherModes.antialias_en = (w2 >> 3) & 1; // 1
2968		OtherModes.z_source_sel = (w2 >> 2) & 1; // 0
2969		OtherModes.dither_alpha_en = (w2 >> 1) & 1; // 0
2970		OtherModes.alpha_compare_en = (w2) & 1; // 0
2971		}
2972
2973		void n64_rdp::CmdLoadTLUT(UINT32 w1, UINT32 w2)
2974		{
2975		//wait("LoadTLUT");
2976		N64Tile* tile = m_tiles;
2977
2978		int tilenum = (w2 >> 24) & 0x7;
2979		int sl = tile[tilenum].sl = ((w1 >> 12) & 0xfff);
2980		int tl = tile[tilenum].tl = w1 & 0xfff;
2981		int sh = tile[tilenum].sh = ((w2 >> 12) & 0xfff);
2982		int th = tile[tilenum].th = w2 & 0xfff;
2983
2984		if (tl != th)
2985		{
2986		fatalerror("Load tlut: tl=%d, th=%d\n",tl,th);
2987		}
2988
2989		int count = (sh >> 2) - (sl >> 2) + 1;
2990		count <<= 2;
2991
2992		switch (MiscState.TISize)
2993		{
2994		case PIXEL_SIZE_16BIT:
2995		{
2996		if (tile[tilenum].tmem < 256)
2997		{
2998		fatalerror("rdp_load_tlut: loading tlut into low half at %d qwords\n",tile[tilenum].tmem);
2999		}
3000		UINT32 srcstart = (MiscState.TIAddress + (tl >> 2) * (MiscState.TIWidth << 1) + (sl >> 1)) >> 1;
3001		UINT16 *dst = GetTMEM16();
3002		UINT32 dststart = tile[tilenum].tmem << 2;
3003
3004		for (int i = 0; i < count; i += 4)
3005		{
3006		if (dststart < 2048)
3007		{
3008		dst[dststart] = U_RREADIDX16(srcstart);
3009		dst[dststart + 1] = dst[dststart];
3010		dst[dststart + 2] = dst[dststart];
3011		dst[dststart + 3] = dst[dststart];
3012		dststart += 4;
3013		srcstart += 1;
3014		}
3015		}
3016		break;
3017		}
3018		default: fatalerror("RDP: load_tlut: size = %d\n", MiscState.TISize);
3019		}
3020		}
3021
3022		void n64_rdp::CmdSetTileSize(UINT32 w1, UINT32 w2)
3023		{
3024		//wait("SetTileSize");
3025
3026		const int tilenum = (w2 >> 24) & 0x7;
3027
3028		m_tiles[tilenum].sl = (w1 >> 12) & 0xfff;
3029		m_tiles[tilenum].tl = (w1 >> 0) & 0xfff;
3030		m_tiles[tilenum].sh = (w2 >> 12) & 0xfff;
3031		m_tiles[tilenum].th = (w2 >> 0) & 0xfff;
3032		}
3033
3034		void n64_rdp::CmdLoadBlock(UINT32 w1, UINT32 w2)
3035		{
3036		//wait("LoadBlock");
3037		N64Tile* tile = m_tiles;
3038
3039		int tilenum = (w2 >> 24) & 0x7;
3040		UINT16* tc = GetTMEM16();
3041
3042		UINT16 sl, sh, tl;
3043		tile[tilenum].sl = sl = ((w1 >> 12) & 0xfff);
3044		tile[tilenum].tl = tl = ((w1 >> 0) & 0xfff);
3045		tile[tilenum].sh = sh = ((w2 >> 12) & 0xfff);
3046		UINT16 dxt = ((w2 >> 0) & 0xfff);
3047
3048		if (sh < sl)
3049		{
3050		fatalerror("load_block: sh < sl\n");
3051		}
3052
3053		INT32 width = (sh - sl) + 1;
3054
3055		width = (width << MiscState.TISize) >> 1;
3056		if (width & 7)
3057		{
3058		width = (width & ~7) + 8;
3059		}
3060		width >>= 3;
3061
3062		UINT32 tb = tile[tilenum].tmem << 2;
3063
3064		int tiwinwords = MiscState.TIWidth;
3065		UINT32 slinwords = sl;
3066
3067		tiwinwords = (tiwinwords << MiscState.TISize) >> 2;
3068		slinwords = (slinwords << MiscState.TISize) >> 2;
3069
3070		int ptr = 0, srcptr = 0;
3071		UINT16 first, sec;
3072		UINT32 src = (MiscState.TIAddress >> 1) + (tl * tiwinwords) + slinwords;
3073
3074		if (dxt != 0)
3075		{
3076		int j = 0;
3077		int t = 0;
3078		int oldt = 0;
3079
3080		if (tile[tilenum].size != PIXEL_SIZE_32BIT && tile[tilenum].format != FORMAT_YUV)
3081		{
3082		for (int i = 0; i < width; i ++)
3083		{
3084		oldt = t;
3085		t = ((j >> 11) & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
3086		if (t != oldt)
3087		{
3088		i += tile[tilenum].line;
3089		}
3090
3091		ptr = tb + (i << 2);
3092		srcptr = src + (i << 2);
3093
3094		tc[(ptr ^ t) & 0x7ff] = U_RREADIDX16(srcptr);
3095		tc[((ptr + 1) ^ t) & 0x7ff] = U_RREADIDX16(srcptr + 1);
3096		tc[((ptr + 2) ^ t) & 0x7ff] = U_RREADIDX16(srcptr + 2);
3097		tc[((ptr + 3) ^ t) & 0x7ff] = U_RREADIDX16(srcptr + 3);
3098		j += dxt;
3099		}
3100		}
3101		else if (tile[tilenum].format == FORMAT_YUV)
3102		{
3103		for (int i = 0; i < width; i ++)
3104		{
3105		oldt = t;
3106		t = ((j >> 11) & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
3107		if (t != oldt)
3108		{
3109		i += tile[tilenum].line;
3110		}
3111
3112		ptr = ((tb + (i << 1)) ^ t) & 0x3ff;
3113		srcptr = src + (i << 2);
3114
3115		first = U_RREADIDX16(srcptr);
3116		sec = U_RREADIDX16(srcptr + 1);
3117		tc[ptr] = ((first >> 8) << 8) \| (sec >> 8);
3118		tc[ptr \| 0x400] = ((first & 0xff) << 8) \| (sec & 0xff);
3119
3120		ptr = ((tb + (i << 1) + 1) ^ t) & 0x3ff;
3121		first = U_RREADIDX16(srcptr + 2);
3122		sec = U_RREADIDX16(srcptr + 3);
3123		tc[ptr] = ((first >> 8) << 8) \| (sec >> 8);
3124		tc[ptr \| 0x400] = ((first & 0xff) << 8) \| (sec & 0xff);
3125
3126		j += dxt;
3127		}
3128		}
3129		else
3130		{
3131		for (int i = 0; i < width; i ++)
3132		{
3133		oldt = t;
3134		t = ((j >> 11) & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
3135		if (t != oldt)
3136		i += tile[tilenum].line;
3137
3138		ptr = ((tb + (i << 1)) ^ t) & 0x3ff;
3139		srcptr = src + (i << 2);
3140		tc[ptr] = U_RREADIDX16(srcptr);
3141		tc[ptr \| 0x400] = U_RREADIDX16(srcptr + 1);
3142
3143		ptr = ((tb + (i << 1) + 1) ^ t) & 0x3ff;
3144		tc[ptr] = U_RREADIDX16(srcptr + 2);
3145		tc[ptr \| 0x400] = U_RREADIDX16(srcptr + 3);
3146
3147		j += dxt;
3148		}
3149		}
3150		tile[tilenum].th = tl + (j >> 11);
3151		}
3152		else
3153		{
3154		if (tile[tilenum].size != PIXEL_SIZE_32BIT && tile[tilenum].format != FORMAT_YUV)
3155		{
3156		for (int i = 0; i < width; i ++)
3157		{
3158		ptr = tb + (i << 2);
3159		srcptr = src + (i << 2);
3160		tc[(ptr ^ WORD_ADDR_XOR) & 0x7ff] = U_RREADIDX16(srcptr);
3161		tc[((ptr + 1) ^ WORD_ADDR_XOR) & 0x7ff] = U_RREADIDX16(srcptr + 1);
3162		tc[((ptr + 2) ^ WORD_ADDR_XOR) & 0x7ff] = U_RREADIDX16(srcptr + 2);
3163		tc[((ptr + 3) ^ WORD_ADDR_XOR) & 0x7ff] = U_RREADIDX16(srcptr + 3);
3164		}
3165		}
3166		else if (tile[tilenum].format == FORMAT_YUV)
3167		{
3168		for (int i = 0; i < width; i ++)
3169		{
3170		ptr = ((tb + (i << 1)) ^ WORD_ADDR_XOR) & 0x3ff;
3171		srcptr = src + (i << 2);
3172		first = U_RREADIDX16(srcptr);
3173		sec = U_RREADIDX16(srcptr + 1);
3174		tc[ptr] = ((first >> 8) << 8) \| (sec >> 8);//UV pair
3175		tc[ptr \| 0x400] = ((first & 0xff) << 8) \| (sec & 0xff);
3176
3177		ptr = ((tb + (i << 1) + 1) ^ WORD_ADDR_XOR) & 0x3ff;
3178		first = U_RREADIDX16(srcptr + 2);
3179		sec = U_RREADIDX16(srcptr + 3);
3180		tc[ptr] = ((first >> 8) << 8) \| (sec >> 8);
3181		tc[ptr \| 0x400] = ((first & 0xff) << 8) \| (sec & 0xff);
3182		}
3183		}
3184		else
3185		{
3186		for (int i = 0; i < width; i ++)
3187		{
3188		ptr = ((tb + (i << 1)) ^ WORD_ADDR_XOR) & 0x3ff;
3189		srcptr = src + (i << 2);
3190		tc[ptr] = U_RREADIDX16(srcptr);
3191		tc[ptr \| 0x400] = U_RREADIDX16(srcptr + 1);
3192
3193		ptr = ((tb + (i << 1) + 1) ^ WORD_ADDR_XOR) & 0x3ff;
3194		tc[ptr] = U_RREADIDX16(srcptr + 2);
3195		tc[ptr \| 0x400] = U_RREADIDX16(srcptr + 3);
3196		}
3197		}
3198		tile[tilenum].th = tl;
3199		}
3200		}
3201
3202		void n64_rdp::CmdLoadTile(UINT32 w1, UINT32 w2)
3203		{
3204		//wait("LoadTile");
3205		N64Tile* tile = m_tiles;
3206		int tilenum = (w2 >> 24) & 0x7;
3207
3208		tile[tilenum].sl = ((w1 >> 12) & 0xfff);
3209		tile[tilenum].tl = ((w1 >> 0) & 0xfff);
3210		tile[tilenum].sh = ((w2 >> 12) & 0xfff);
3211		tile[tilenum].th = ((w2 >> 0) & 0xfff);
3212
3213		UINT16 sl = tile[tilenum].sl >> 2;
3214		UINT16 tl = tile[tilenum].tl >> 2;
3215		UINT16 sh = tile[tilenum].sh >> 2;
3216		UINT16 th = tile[tilenum].th >> 2;
3217
3218		INT32 width = (sh - sl) + 1;
3219		INT32 height = (th - tl) + 1;
3220		/*
3221		int topad;
3222		if (MiscState.TISize < 3)
3223		{
3224		topad = (width * MiscState.TISize) & 0x7;
3225		}
3226		else
3227		{
3228		topad = (width << 2) & 0x7;
3229		}
3230		topad = 0; // ????
3231		*/
3232		switch (MiscState.TISize)
3233		{
3234		case PIXEL_SIZE_8BIT:
3235		{
3236		UINT32 src = MiscState.TIAddress;
3237		UINT8 *tc = GetTMEM();
3238		int tb = tile[tilenum].tmem << 3;
3239
3240		for (int j = 0; j < height; j++)
3241		{
3242		int tline = tb + ((tile[tilenum].line << 3) * j);
3243		int s = ((j + tl) * MiscState.TIWidth) + sl;
3244
3245		int xorval8 = ((j & 1) ? BYTE_XOR_DWORD_SWAP : BYTE_ADDR_XOR);
3246		for (int i = 0; i < width; i++)
3247		{
3248		tc[((tline + i) ^ xorval8) & 0xfff] = U_RREADADDR8(src + s + i);
3249		}
3250		}
3251		break;
3252		}
3253		case PIXEL_SIZE_16BIT:
3254		{
3255		UINT32 src = MiscState.TIAddress >> 1;
3256		UINT16 *tc = GetTMEM16();
3257		UINT16 yuvword;
3258
3259		if (tile[tilenum].format != FORMAT_YUV)
3260		{
3261		for (int j = 0; j < height; j++)
3262		{
3263		int tb = tile[tilenum].tmem << 2;
3264		int tline = tb + ((tile[tilenum].line << 2) * j);
3265		int s = ((j + tl) * MiscState.TIWidth) + sl;
3266		int xorval16 = (j & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
3267
3268		for (int i = 0; i < width; i++)
3269		{
3270		UINT32 taddr = (tline + i) ^ xorval16;
3271		tc[taddr & 0x7ff] = U_RREADIDX16(src + s + i);
3272		}
3273		}
3274		}
3275		else
3276		{
3277		for (int j = 0; j < height; j++)
3278		{
3279		int tb = tile[tilenum].tmem << 3;
3280		int tline = tb + ((tile[tilenum].line << 3) * j);
3281		int s = ((j + tl) * MiscState.TIWidth) + sl;
3282		int xorval8 = (j & 1) ? BYTE_XOR_DWORD_SWAP : BYTE_ADDR_XOR;
3283
3284		for (int i = 0; i < width; i++)
3285		{
3286		UINT32 taddr = ((tline + i) ^ xorval8) & 0x7ff;
3287		yuvword = U_RREADIDX16(src + s + i);
3288		GetTMEM()[taddr] = yuvword >> 8;
3289		GetTMEM()[taddr \| 0x800] = yuvword & 0xff;
3290		}
3291		}
3292		}
3293		break;
3294		}
3295		case PIXEL_SIZE_32BIT:
3296		{
3297		UINT32 src = MiscState.TIAddress >> 2;
3298		UINT16 *tc16 = GetTMEM16();
3299		int tb = (tile[tilenum].tmem << 2);
3300
3301		for (int j = 0; j < height; j++)
3302		{
3303		int tline = tb + ((tile[tilenum].line << 2) * j);
3304
3305		int s = ((j + tl) * MiscState.TIWidth) + sl;
3306		int xorval32cur = (j & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
3307		for (int i = 0; i < width; i++)
3308		{
3309		UINT32 c = U_RREADIDX32(src + s + i);
3310		UINT32 ptr = ((tline + i) ^ xorval32cur) & 0x3ff;
3311		tc16[ptr] = c >> 16;
3312		tc16[ptr \| 0x400] = c & 0xffff;
3313		}
3314		}
3315		break;
3316		}
3317
3318		default: fatalerror("RDP: load_tile: size = %d\n", MiscState.TISize);
3319		}
3320		}
3321
3322		void n64_rdp::CmdSetTile(UINT32 w1, UINT32 w2)
3323		{
3324		//wait("SetTile");
3325
3326		int tilenum = (w2 >> 24) & 0x7;
3327		N64Tile* tex_tile = &m_tiles[tilenum];
3328
3329		tex_tile->format = (w1 >> 21) & 0x7;
3330		tex_tile->size = (w1 >> 19) & 0x3;
3331		tex_tile->line = (w1 >> 9) & 0x1ff;
3332		tex_tile->tmem = (w1 >> 0) & 0x1ff;
3333		tex_tile->palette = (w2 >> 20) & 0xf;
3334		tex_tile->ct = (w2 >> 19) & 0x1;
3335		tex_tile->mt = (w2 >> 18) & 0x1;
3336		tex_tile->mask_t = (w2 >> 14) & 0xf;
3337		tex_tile->shift_t = (w2 >> 10) & 0xf;
3338		tex_tile->cs = (w2 >> 9) & 0x1;
3339		tex_tile->ms = (w2 >> 8) & 0x1;
3340		tex_tile->mask_s = (w2 >> 4) & 0xf;
3341		tex_tile->shift_s = (w2 >> 0) & 0xf;
3342
3343		if (tex_tile->format == FORMAT_I && tex_tile->size > PIXEL_SIZE_8BIT)
3344		{
3345		tex_tile->format = FORMAT_RGBA; // Used by Supercross 2000 (in-game)
3346		}
3347		if (tex_tile->format == FORMAT_CI && tex_tile->size > PIXEL_SIZE_8BIT)
3348		{
3349		tex_tile->format = FORMAT_RGBA; // Used by Clay Fighter - Sculptor's Cut
3350		}
3351
3352		if (tex_tile->format == FORMAT_RGBA && tex_tile->size < PIXEL_SIZE_16BIT)
3353		{
3354		tex_tile->format = FORMAT_CI; // Used by Exterem-G2, Madden Football 64, and Rat Attack
3355		}
3356
3357		//m_pending_mode_block = true;
3358		}
3359
3360		void n64_rdp::CmdFillRect(UINT32 w1, UINT32 w2)
3361		{
3362		//if(m_pending_mode_block) { wait("Block on pending mode-change"); m_pending_mode_block = false; }
3363
3364		UINT32 xl = (w1 >> 12) & 0xfff;
3365		UINT32 yl = (w1 >> 0) & 0xfff;
3366		UINT32 xh = (w2 >> 12) & 0xfff;
3367		UINT32 yh = (w2 >> 0) & 0xfff;
3368
3369		if (OtherModes.cycle_type == CYCLE_TYPE_FILL \|\| OtherModes.cycle_type == CYCLE_TYPE_COPY)
3370		{
3371		yl \|= 3;
3372		}
3373
3374		UINT32 xlint = (xl >> 2) & 0x3ff;
3375		UINT32 xhint = (xh >> 2) & 0x3ff;
3376
3377		UINT32* ewdata = m_temp_rect_data;
3378		ewdata[0] = (0x3680 << 16) \| yl;//command, flipped, tile, yl
3379		ewdata[1] = (yl << 16) \| yh;//ym, yh
3380		ewdata[2] = (xlint << 16) \| ((xl & 3) << 14);//xl, xl frac
3381		ewdata[3] = 0;//dxldy, dxldy frac
3382		ewdata[4] = (xhint << 16) \| ((xh & 3) << 14);//xh, xh frac
3383		ewdata[5] = 0;//dxhdy, dxhdy frac
3384		ewdata[6] = (xlint << 16) \| ((xl & 3) << 14);//xm, xm frac
3385		ewdata[7] = 0;//dxmdy, dxmdy frac
3386		memset(&ewdata[8], 0, 36 * sizeof(UINT32));//shade, texture, depth
3387
3388		DrawTriangle(false, false, false, true);
3389		}
3390
3391		void n64_rdp::CmdSetFogColor(UINT32 w1, UINT32 w2)
3392		{
3393		FogColor.c = w2;
3394		}
3395
3396		void n64_rdp::CmdSetBlendColor(UINT32 w1, UINT32 w2)
3397		{
3398		BlendColor.c = w2;
3399		}
3400
3401		void n64_rdp::CmdSetPrimColor(UINT32 w1, UINT32 w2)
3402		{
3403		MiscState.MinLevel = (w1 >> 8) & 0x1f;
3404		PrimLODFraction = w1 & 0xff;
3405		PrimColor.c = w2;
3406		}
3407
3408		void n64_rdp::CmdSetEnvColor(UINT32 w1, UINT32 w2)
3409		{
3410		EnvColor.c = w2;
3411		}
3412
3413		void n64_rdp::CmdSetCombine(UINT32 w1, UINT32 w2)
3414		{
3415		m_combine.sub_a_rgb0 = (w1 >> 20) & 0xf;
3416		m_combine.mul_rgb0 = (w1 >> 15) & 0x1f;
3417		m_combine.sub_a_a0 = (w1 >> 12) & 0x7;
3418		m_combine.mul_a0 = (w1 >> 9) & 0x7;
3419		m_combine.sub_a_rgb1 = (w1 >> 5) & 0xf;
3420		m_combine.mul_rgb1 = (w1 >> 0) & 0x1f;
3421
3422		m_combine.sub_b_rgb0 = (w2 >> 28) & 0xf;
3423		m_combine.sub_b_rgb1 = (w2 >> 24) & 0xf;
3424		m_combine.sub_a_a1 = (w2 >> 21) & 0x7;
3425		m_combine.mul_a1 = (w2 >> 18) & 0x7;
3426		m_combine.add_rgb0 = (w2 >> 15) & 0x7;
3427		m_combine.sub_b_a0 = (w2 >> 12) & 0x7;
3428		m_combine.add_a0 = (w2 >> 9) & 0x7;
3429		m_combine.add_rgb1 = (w2 >> 6) & 0x7;
3430		m_combine.sub_b_a1 = (w2 >> 3) & 0x7;
3431		m_combine.add_a1 = (w2 >> 0) & 0x7;
3432		}
3433
3434		void n64_rdp::CmdSetTextureImage(UINT32 w1, UINT32 w2)
3435		{
3436		MiscState.TIFormat = (w1 >> 21) & 0x7;
3437		MiscState.TISize = (w1 >> 19) & 0x3;
3438		MiscState.TIWidth = (w1 & 0x3ff) + 1;
3439		MiscState.TIAddress = w2 & 0x01ffffff;
3440		}
3441
3442		void n64_rdp::CmdSetMaskImage(UINT32 w1, UINT32 w2)
3443		{
3444		//wait("SetMaskImage");
3445
3446		MiscState.ZBAddress = w2 & 0x01ffffff;
3447		}
3448
3449		void n64_rdp::CmdSetColorImage(UINT32 w1, UINT32 w2)
3450		{
3451		//wait("SetColorImage");
3452
3453		MiscState.FBFormat = (w1 >> 21) & 0x7;
3454		MiscState.FBSize = (w1 >> 19) & 0x3;
3455		MiscState.FBWidth = (w1 & 0x3ff) + 1;
3456		MiscState.FBAddress = w2 & 0x01ffffff;
3457
3458		if (MiscState.FBFormat < 2 \|\| MiscState.FBFormat > 32) // Jet Force Gemini sets the format to 4, Intensity. Protection?
3459		{
3460		MiscState.FBFormat = 2;
3461		}
3462		}
3463
3464		UINT32 n64_rdp::AddRightCvg(UINT32 x, UINT32 k)
3465		{
3466		//#undef FULL_SUBPIXELS
3467		#define FULL_SUBPIXELS
3468		UINT32 coveredsubpixels=((x >> 14) & 3);
3469		if (!(x & 0xffff))
3470		{
3471		return 0;
3472		}
3473		#ifdef FULL_SUBPIXELS
3474		if (!coveredsubpixels)
3475		{
3476		return 0;
3477		}
3478		if (!(k & 1))
3479		{
3480		return (coveredsubpixels<3) ? 1 : 2;
3481		}
3482		else
3483		{
3484		return (coveredsubpixels<2) ? 0 : 1;
3485		}
3486		#endif
3487		if (!(k & 1))
3488		{
3489		return (coveredsubpixels<2) ? 1 : 2;
3490		}
3491		else
3492		{
3493		if (coveredsubpixels<1)
3494		{
3495		return 0;
3496		}
3497		else if (coveredsubpixels<3)
3498		{
3499		return 1;
3500		}
3501		else
3502		{
3503		return 2;
3504		}
3505		}
3506		}
3507
3508		UINT32 n64_rdp::AddLeftCvg(UINT32 x, UINT32 k)
3509		{
3510		UINT32 coveredsubpixels = 3 - ((x >> 14) & 3);
3511		if (!(x & 0xffff))
3512		{
3513		return 2;
3514		}
3515		#ifdef FULL_SUBPIXELS
3516		if (!coveredsubpixels)
3517		{
3518		return 0;
3519		}
3520		if (!(k & 1))
3521		{
3522		return (coveredsubpixels<2) ? 0 : 1;
3523		}
3524		else
3525		{
3526		return (coveredsubpixels<3) ? 1 : 2;
3527		}
3528		#endif
3529		if (k & 1)
3530		{
3531		return (coveredsubpixels<2) ? 1 : 2;
3532		}
3533		else
3534		{
3535		if (coveredsubpixels < 1)
3536		{
3537		return 0;
3538		}
3539		else if (coveredsubpixels < 3)
3540		{
3541		return 1;
3542		}
3543		else
3544		{
3545		return 2;
3546		}
3547		}
3548		}
3549
3550		/*****************************************************************************/
3551
3552		void n64_rdp::CmdInvalid(UINT32 w1, UINT32 w2)
3553		{
3554		fatalerror("n64_rdp::Invalid: %d, %08x %08x\n", (w1 >> 24) & 0x3f, w1, w2);
3555		}
3556
3557		void n64_rdp::CmdNoOp(UINT32 w1, UINT32 w2)
3558		{
3559		// Do nothing
3560		}
3561
3562
3563		void n64_rdp::ProcessList()
3564		{
3565		INT32 length = m_end - m_current;
3566
3567		if(length < 0)
3568		{
3569		m_current = m_end;
3570		return;
3571		}
3572
3573		// load command data
3574		for(int i = 0; i < length; i += 4)
3575		{
3576		m_cmd_data[m_cmd_ptr++] = ReadData((m_current & 0x1fffffff) + i);
3577		}
3578
3579		m_current = m_end;
3580
3581		UINT32 cmd = (m_cmd_data[0] >> 24) & 0x3f;
3582		UINT32 cmd_length = (m_cmd_ptr + 1) * 4;
3583
3584		SetStatusReg(GetStatusReg() &~ DP_STATUS_FREEZE);
3585
3586		// check if more data is needed
3587		if (cmd_length < rdp_command_length[cmd])
3588		{
3589		return;
3590		}
3591
3592		while (m_cmd_cur < m_cmd_ptr)
3593		{
3594		cmd = (m_cmd_data[m_cmd_cur] >> 24) & 0x3f;
3595
3596		if (((m_cmd_ptr - m_cmd_cur) * 4) < rdp_command_length[cmd])
3597		{
3598		return;
3599		//fatalerror("rdp_process_list: not enough rdp command data: cur = %d, ptr = %d, expected = %d\n", m_cmd_cur, m_cmd_ptr, rdp_command_length[cmd]);
3600		}
3601
3602		if (LOG_RDP_EXECUTION)
3603		{
3604		char string[4000];
3605		Dasm(string);
3606
3607		fprintf(rdp_exec, "%08X: %08X %08X %s\n", m_start+(m_cmd_cur * 4), m_cmd_data[m_cmd_cur+0], m_cmd_data[m_cmd_cur+1], string);
3608		fflush(rdp_exec);
3609		}
3610
3611		// execute the command
3612		UINT32 w1 = m_cmd_data[m_cmd_cur+0];
3613		UINT32 w2 = m_cmd_data[m_cmd_cur+1];
3614
3615		switch(cmd)
3616		{
3617		case 0x00: CmdNoOp(w1, w2); break;
3618
3619		case 0x08: CmdTriangle(w1, w2); break;
3620		case 0x09: CmdTriangleZ(w1, w2); break;
3621		case 0x0a: CmdTriangleT(w1, w2); break;
3622		case 0x0b: CmdTriangleTZ(w1, w2); break;
3623		case 0x0c: CmdTriangleS(w1, w2); break;
3624		case 0x0d: CmdTriangleSZ(w1, w2); break;
3625		case 0x0e: CmdTriangleST(w1, w2); break;
3626		case 0x0f: CmdTriangleSTZ(w1, w2); break;
3627
3628		case 0x24: CmdTexRect(w1, w2); break;
3629		case 0x25: CmdTexRectFlip(w1, w2); break;
3630
3631		case 0x26: CmdSyncLoad(w1, w2); break;
3632		case 0x27: CmdSyncPipe(w1, w2); break;
3633		case 0x28: CmdSyncTile(w1, w2); break;
3634		case 0x29: CmdSyncFull(w1, w2); break;
3635
3636		case 0x2a: CmdSetKeyGB(w1, w2); break;
3637		case 0x2b: CmdSetKeyR(w1, w2); break;
3638
3639		case 0x2c: CmdSetConvert(w1, w2); break;
3640		case 0x3c: CmdSetCombine(w1, w2); break;
3641		case 0x2d: CmdSetScissor(w1, w2); break;
3642		case 0x2e: CmdSetPrimDepth(w1, w2); break;
3643		case 0x2f: CmdSetOtherModes(w1, w2); break;
3644
3645		case 0x30: CmdLoadTLUT(w1, w2); break;
3646		case 0x33: CmdLoadBlock(w1, w2); break;
3647		case 0x34: CmdLoadTile(w1, w2); break;
3648
3649		case 0x32: CmdSetTileSize(w1, w2); break;
3650		case 0x35: CmdSetTile(w1, w2); break;
3651
3652		case 0x36: CmdFillRect(w1, w2); break;
3653
3654		case 0x37: CmdSetFillColor32(w1, w2); break;
3655		case 0x38: CmdSetFogColor(w1, w2); break;
3656		case 0x39: CmdSetBlendColor(w1, w2); break;
3657		case 0x3a: CmdSetPrimColor(w1, w2); break;
3658		case 0x3b: CmdSetEnvColor(w1, w2); break;
3659
3660		case 0x3d: CmdSetTextureImage(w1, w2); break;
3661		case 0x3e: CmdSetMaskImage(w1, w2); break;
3662		case 0x3f: CmdSetColorImage(w1, w2); break;
3663		}
3664
3665		m_cmd_cur += rdp_command_length[cmd] / 4;
3666		};
3667		m_cmd_ptr = 0;
3668		m_cmd_cur = 0;
3669
3670		m_start = m_current = m_end;
3671		}
3672
3673		/*****************************************************************************/
3674
3675		n64_rdp::n64_rdp(n64_state &state) : poly_manager<UINT32, rdp_poly_state, 8, 32000>(state.machine())
3676		{
3677		AuxBufPtr = 0;
3678		AuxBuf = NULL;
3679		m_pipe_clean = true;
3680
3681		m_pending_mode_block = false;
3682
3683		m_cmd_ptr = 0;
3684		m_cmd_cur = 0;
3685
3686		m_start = 0;
3687		m_end = 0;
3688		m_current = 0;
3689		m_status = 0x88;
3690
3691		for (int i = 0; i < 8; i++)
3692		{
3693		m_tiles[i].num = i;
3694		}
3695
3696		OneColor.c = 0xffffffff;
3697		ZeroColor.c = 0x00000000;
3698
3699		m_tmem = NULL;
3700
3701		m_machine = NULL;
3702
3703		//memset(m_hidden_bits, 3, 8388608);
3704
3705		PrimLODFraction = 0;
3706
3707		for (int i = 0; i < 256; i++)
3708		{
3709		m_gamma_table[i] = sqrt((float)(i << 6));
3710		m_gamma_table[i] <<= 1;
3711		}
3712
3713		for (int i = 0; i < 0x4000; i++)
3714		{
3715		m_gamma_dither_table[i] = sqrt((float)i);
3716		m_gamma_dither_table[i] <<= 1;
3717		}
3718
3719		z_build_com_table();
3720
3721		for (int i = 0; i < 0x4000; i++)
3722		{
3723		UINT32 exponent = (i >> 11) & 7;
3724		UINT32 mantissa = i & 0x7ff;
3725		z_complete_dec_table[i] = ((mantissa << z_dec_table[exponent].shift) + z_dec_table[exponent].add) & 0x3fffff;
3726		}
3727
3728		precalc_cvmask_derivatives();
3729
3730		for(int i = 0; i < 0x200; i++)
3731		{
3732		switch((i >> 7) & 3)
3733		{
3734		case 0:
3735		case 1:
3736		s_special_9bit_clamptable[i] = i & 0xff;
3737		break;
3738		case 2:
3739		s_special_9bit_clamptable[i] = 0xff;
3740		break;
3741		case 3:
3742		s_special_9bit_clamptable[i] = 0;
3743		break;
3744		}
3745		}
3746
3747		for(int i = 0; i < 32; i++)
3748		{
3749		ReplicatedRGBA[i] = (i << 3) \| ((i >> 2) & 7);
3750		}
3751
3752		_Write[0] = &n64_rdp::_Write16Bit_Cvg0_NoBlend;
3753		_Write[1] = &n64_rdp::_Write16Bit_Cvg0_Blend;
3754		_Write[2] = &n64_rdp::_Write16Bit_Cvg1;
3755		_Write[3] = &n64_rdp::_Write16Bit_Cvg1;
3756		_Write[4] = &n64_rdp::_Write16Bit_Cvg2;
3757		_Write[5] = &n64_rdp::_Write16Bit_Cvg2;
3758		_Write[6] = &n64_rdp::_Write16Bit_Cvg3;
3759		_Write[7] = &n64_rdp::_Write16Bit_Cvg3;
3760		_Write[8] = &n64_rdp::_Write32Bit_Cvg0_NoBlend;
3761		_Write[9] = &n64_rdp::_Write32Bit_Cvg0_Blend;
3762		_Write[10] = &n64_rdp::_Write32Bit_Cvg1;
3763		_Write[11] = &n64_rdp::_Write32Bit_Cvg1;
3764		_Write[12] = &n64_rdp::_Write32Bit_Cvg2;
3765		_Write[13] = &n64_rdp::_Write32Bit_Cvg2;
3766		_Write[14] = &n64_rdp::_Write32Bit_Cvg3;
3767		_Write[15] = &n64_rdp::_Write32Bit_Cvg3;
3768
3769		_Read[0] = &n64_rdp::_Read16Bit_ImgRead0;
3770		_Read[1] = &n64_rdp::_Read16Bit_ImgRead1;
3771		_Read[2] = &n64_rdp::_Read32Bit_ImgRead0;
3772		_Read[3] = &n64_rdp::_Read32Bit_ImgRead1;
3773
3774		_Copy[0] = &n64_rdp::_Copy16Bit;
3775		_Copy[1] = &n64_rdp::_Copy32Bit;
3776
3777		_Fill[0] = &n64_rdp::_Fill16Bit;
3778		_Fill[1] = &n64_rdp::_Fill32Bit;
3779		}
3780
3781		void n64_state::video_start()
3782		{
3783		m_rdp = auto_alloc(machine(), n64_rdp(*this));
3784
3785		m_rdp->SetMachine(machine());
3786		m_rdp->InitInternalState();
3787
3788		m_rdp->Blender.SetMachine(machine());
3789		m_rdp->Blender.SetProcessor(m_rdp);
3790
3791		m_rdp->TexPipe.SetMachine(machine());
3792
3793		m_rdp->AuxBuf = auto_alloc_array_clear(machine(), UINT8, EXTENT_AUX_COUNT);
3794
3795		if (LOG_RDP_EXECUTION)
3796		{
3797		rdp_exec = fopen("rdp_execute.txt", "wt");
3798		}
3799		}
3800
3801		UINT32 n64_state::screen_update_n64(screen_device &screen, bitmap_rgb32 &bitmap, const rectangle &cliprect)
3802		{
3803		n64_periphs *n64 = machine().device<n64_periphs>("rcp");
3804		m_rdp->visarea = screen.visible_area();
3805
3806		//UINT16 frame_buffer = (UINT16)&rdram[(n64->vi_origin & 0xffffff) >> 2];
3807		//UINT8 *cvg_buffer = &m_rdp.HiddenBits[((n64->vi_origin & 0xffffff) >> 2) >> 1];
3808		//int vibuffering = ((n64->vi_control & 2) && fsaa && divot);
3809
3810		//vibuffering = 0; // Disabled for now
3811
3812		/*
3813		if (vibuffering && ((n64->vi_control & 3) == 2))
3814		{
3815		if (frame_buffer)
3816		{
3817		for (j=0; j < vres; j++)
3818		{
3819		for (i=0; i < hres; i++)
3820		{
3821		UINT16 pix;
3822		pix = frame_buffer[pixels ^ WORD_ADDR_XOR];
3823		curpixel_cvg = ((pix & 1) << 2) \| (cvg_buffer[pixels ^ BYTE_ADDR_XOR] & 3); // Reuse of this variable
3824		if (curpixel_cvg < 7 && i > 1 && j > 1 && i < (hres - 2) && j < (vres - 2) && fsaa)
3825		{
3826		newc = video_filter16(&frame_buffer[pixels ^ WORD_ADDR_XOR], &cvg_buffer[pixels ^ BYTE_ADDR_XOR], n64->vi_width);
3827		ViBuffer[i][j] = newc;
3828		}
3829		else
3830		{
3831		newc.i.r = ((pix >> 8) & 0xf8) \| (pix >> 13);
3832		newc.i.g = ((pix >> 3) & 0xf8) \| ((pix >> 8) & 0x07);
3833		newc.i.b = ((pix << 2) & 0xf8) \| ((pix >> 3) & 0x07);
3834		ViBuffer[i][j] = newc;
3835		}
3836		pixels++;
3837		}
3838		pixels += invisiblewidth;
3839		}
3840		}
3841		}
3842		*/
3843
3844		m_rdp->wait();
3845		m_rdp->AuxBufPtr = 0;
3846
3847		if (n64->vi_blank)
3848		{
3849		bitmap.fill(0, m_rdp->visarea);
3850		return 0;
3851		}
3852
3853		m_rdp->VideoUpdate(n64, bitmap);
3854
3855		return 0;
3856		}
	1	/******************************************************************************
	2
	3
	4	SGI/Nintendo Reality Display Processor
	5	-------------------
	6
	7	by MooglyGuy
	8	based on initial C code by Ville Linde
	9	contains additional improvements from angrylion, Ziggy, Gonetz and Orkin
	10
	11
	12	*******************************************************************************
	13
	14	STATUS:
	15
	16	Much behavior needs verification against real hardware. Many edge cases must
	17	be verified on real hardware as well.
	18
	19	TODO:
	20
	21	- Further re-work class structure to avoid dependencies
	22
	23	*******************************************************************************/
	24
	25	#include "emu.h"
	26	#include "video/n64.h"
	27
	28	#define LOG_RDP_EXECUTION 0
	29
	30	static FILE *rdp_exec;
	31
	32	UINT32 n64_rdp::s_special_9bit_clamptable[512];
	33
	34	bool n64_rdp::rdp_range_check(UINT32 addr)
	35	{
	36	if(MiscState.FBSize == 0) return false;
	37
	38	int fbcount = ((MiscState.FBWidth * Scissor.m_yl) << (MiscState.FBSize - 1)) * 3;
	39	int fbaddr = MiscState.FBAddress & 0x007fffff;
	40	if ((addr >= fbaddr) && (addr < (fbaddr + fbcount)))
	41	{
	42	return false;
	43	}
	44
	45	int zbcount = MiscState.FBWidth * Scissor.m_yl * 2;
	46	int zbaddr = MiscState.ZBAddress & 0x007fffff;
	47	if ((addr >= zbaddr) && (addr < (zbaddr + zbcount)))
	48	{
	49	return false;
	50	}
	51
	52	printf("Check failed: %08x vs. %08x-%08x, %08x-%08x (%d, %d)\n", addr, fbaddr, fbaddr + fbcount, zbaddr, zbaddr + zbcount, MiscState.FBWidth, Scissor.m_yl);
	53	fflush(stdout);
	54	return true;
	55	}
	56
	57	/*****************************************************************************/
	58
	59	// The functions in this file should be moved into the parent Processor class.
	60	#include "rdpfiltr.inc"
	61
	62	void n64_rdp::GetAlphaCvg(UINT8 comb_alpha, rdp_span_aux userdata, const rdp_poly_state &object)
	63	{
	64	INT32 temp = *comb_alpha;
	65	INT32 temp2 = userdata->CurrentPixCvg;
	66	INT32 temp3 = 0;
	67
	68	if (object.OtherModes.cvg_times_alpha)
	69	{
	70	temp3 = (temp * temp2) + 4;
	71	userdata->CurrentPixCvg = (temp3 >> 8) & 0xf;
	72	}
	73	if (object.OtherModes.alpha_cvg_select)
	74	{
	75	temp = (OtherModes.cvg_times_alpha) ? (temp3 >> 3) : (temp2 << 5);
	76	}
	77	if (temp > 0xff)
	78	{
	79	temp = 0xff;
	80	}
	81	*comb_alpha = temp;
	82	}
	83
	84	/*****************************************************************************/
	85
	86	void n64_rdp::VideoUpdate(n64_periphs *n64, bitmap_rgb32 &bitmap)
	87	{
	88	switch(n64->vi_control & 0x3)
	89	{
	90	case PIXEL_SIZE_16BIT:
	91	VideoUpdate16(n64, bitmap);
	92	break;
	93
	94	case PIXEL_SIZE_32BIT:
	95	VideoUpdate32(n64, bitmap);
	96	break;
	97
	98	default:
	99	//fatalerror("Unsupported framebuffer depth: m_fb_size=%d\n", MiscState.FBSize);
	100	break;
	101	}
	102	}
	103
	104	void n64_rdp::VideoUpdate16(n64_periphs *n64, bitmap_rgb32 &bitmap)
	105	{
	106	//int fsaa = (((n64->vi_control >> 8) & 3) < 2);
	107	//int divot = (n64->vi_control >> 4) & 1;
	108
	109	//UINT32 prev_cvg = 0;
	110	//UINT32 next_cvg = 0;
	111	//int dither_filter = (n64->vi_control >> 16) & 1;
	112	//int vibuffering = ((n64->vi_control & 2) && fsaa && divot);
	113
	114	UINT16 frame_buffer = (UINT16)&rdram[(n64->vi_origin & 0xffffff) >> 2];
	115	//UINT32 hb = ((n64->vi_origin & 0xffffff) >> 2) >> 1;
	116	//UINT8* hidden_buffer = &HiddenBits[hb];
	117
	118	INT32 hdiff = (n64->vi_hstart & 0x3ff) - ((n64->vi_hstart >> 16) & 0x3ff);
	119	float hcoeff = ((float)(n64->vi_xscale & 0xfff) / (1 << 10));
	120	UINT32 hres = ((float)hdiff * hcoeff);
	121	INT32 invisiblewidth = n64->vi_width - hres;
	122
	123	INT32 vdiff = ((n64->vi_vstart & 0x3ff) - ((n64->vi_vstart >> 16) & 0x3ff)) >> 1;
	124	float vcoeff = ((float)(n64->vi_yscale & 0xfff) / (1 << 10));
	125	UINT32 vres = ((float)vdiff * vcoeff);
	126
	127	if (vdiff <= 0 \|\| hdiff <= 0)
	128	{
	129	return;
	130	}
	131
	132	if (hres > 640) // Needed by Top Gear Overdrive (E)
	133	{
	134	invisiblewidth += (hres - 640);
	135	hres = 640;
	136	}
	137
	138	if (vres > bitmap.height()) // makes Perfect Dark boot w/o crashing
	139	{
	140	vres = bitmap.height();
	141	}
	142
	143	UINT32 pixels = 0;
	144
	145	if (frame_buffer)
	146	{
	147	for(int j = 0; j < vres; j++)
	148	{
	149	UINT32 *d = &bitmap.pix32(j);
	150
	151	for(int i = 0; i < hres; i++)
	152	{
	153	Color c;
	154	//int r, g, b;
	155
	156	UINT16 pix = frame_buffer[pixels ^ WORD_ADDR_XOR];
	157	//MiscState.CurrentPixCvg = ((pix & 1) << 2) \| (hidden_buffer[pixels ^ BYTE_ADDR_XOR] & 3);
	158
	159	//if(divot)
	160	//{
	161	// if(i > 0 && i < (hres - 1))
	162	// {
	163	// prev_cvg = ((frame_buffer[(pixels - 1)^WORD_ADDR_XOR] & 1) << 2) \| (hidden_buffer[(pixels - 1)^BYTE_ADDR_XOR] & 3);
	164	// next_cvg = ((frame_buffer[(pixels + 1)^WORD_ADDR_XOR] & 1) << 2) \| (hidden_buffer[(pixels + 1)^BYTE_ADDR_XOR] & 3);
	165	// }
	166	//}
	167	c.i.r = ((pix >> 8) & 0xf8) \| (pix >> 13);
	168	c.i.g = ((pix >> 3) & 0xf8) \| ((pix >> 8) & 0x07);
	169	c.i.b = ((pix << 2) & 0xf8) \| ((pix >> 3) & 0x07);
	170
	171	//if(fsaa)
	172	//{
	173	//if (/!vibuffering &&/ state->m_rdp.MiscState.CurrentPixCvg < 7 && i > 1 && j > 1 && i < (hres - 2) && j < (vres - 2))
	174	//{
	175	//video_filter16(&c.i.r, &c.i.g, &c.i.b, &frame_buffer[pixels ^ WORD_ADDR_XOR],&hidden_buffer[pixels ^ BYTE_ADDR_XOR], n64->vi_width);
	176	//}
	177	//}
	178	//else if (dither_filter && state->m_rdp.MiscState.CurrentPixCvg == 7 && i > 0 && j > 0 && i < (hres - 1) && j < (vres - 1))
	179	//{
	180	//if (vibuffering)
	181	//{
	182	// restore_filter16_buffer(&r, &g, &b, &ViBuffer[i][j], n64->vi_width);
	183	//}
	184	//else
	185	//{
	186	//restore_filter16(&c.i.r, &c.i.g, &c.i.b, &frame_buffer[pixels ^ WORD_ADDR_XOR], pixels ^ WORD_ADDR_XOR, n64->vi_width);
	187	//}
	188	//}
	189	//if(divot)
	190	//{
	191	//if (i > 0 && i < (hres - 1) && (MiscState.CurrentPixCvg != 7 \|\| prev_cvg != 7 \|\| next_cvg != 7))
	192	//{
	193	//if (vibuffering)
	194	//{
	195	// divot_filter16_buffer(&r, &g, &b, &ViBuffer[i][j]);
	196	//}
	197	//else
	198	//{
	199	//divot_filter16(&c.i.r, &c.i.g, &c.i.b, &frame_buffer[pixels ^ WORD_ADDR_XOR], pixels ^ WORD_ADDR_XOR);
	200	//}
	201	//}
	202	//}
	203
	204	/*
	205	if (gamma_dither)
	206	{
	207	dith = screen.machine().rand() & 0x3f;
	208	}
	209	if (gamma)
	210	{
	211	if (gamma_dither)
	212	{
	213	r = m_gamma_dither_table[(r << 6)\|dith];
	214	g = m_gamma_dither_table[(g << 6)\|dith];
	215	b = m_gamma_dither_table[(b << 6)\|dith];
	216	}
	217	else
	218	{
	219	r = m_gamma_table[r];
	220	g = m_gamma_table[g];
	221	b = m_gamma_table[b];
	222	}
	223	}
	224	else if (gamma_dither)
	225	{
	226	if (r < 255)
	227	r += (dith & 1);
	228	if (g < 255)
	229	g += (dith & 1);
	230	if (b < 255)
	231	b += (dith & 1);
	232	}
	233	*/
	234	pixels++;
	235
	236	d[i] = c.c >> 8;//(r << 16) \| (g << 8) \| b; // Fix me for endianness
	237	}
	238	pixels +=invisiblewidth;
	239	}
	240	}
	241	}
	242
	243	void n64_rdp::VideoUpdate32(n64_periphs *n64, bitmap_rgb32 &bitmap)
	244	{
	245	int gamma = (n64->vi_control >> 3) & 1;
	246	int gamma_dither = (n64->vi_control >> 2) & 1;
	247	//int vibuffering = ((n64->vi_control & 2) && fsaa && divot);
	248
	249	UINT32 frame_buffer32 = (UINT32)&rdram[(n64->vi_origin & 0xffffff) >> 2];
	250
	251	const INT32 hdiff = (n64->vi_hstart & 0x3ff) - ((n64->vi_hstart >> 16) & 0x3ff);
	252	const float hcoeff = ((float)(n64->vi_xscale & 0xfff) / (1 << 10));
	253	UINT32 hres = ((float)hdiff * hcoeff);
	254	INT32 invisiblewidth = n64->vi_width - hres;
	255
	256	const INT32 vdiff = ((n64->vi_vstart & 0x3ff) - ((n64->vi_vstart >> 16) & 0x3ff)) >> 1;
	257	const float vcoeff = ((float)(n64->vi_yscale & 0xfff) / (1 << 10));
	258	const UINT32 vres = ((float)vdiff * vcoeff);
	259
	260	if (vdiff <= 0 \|\| hdiff <= 0)
	261	{
	262	return;
	263	}
	264
	265	if (hres > 640) // Needed by Top Gear Overdrive (E)
	266	{
	267	invisiblewidth += (hres - 640);
	268	hres = 640;
	269	}
	270
	271	if (frame_buffer32)
	272	{
	273	for (int j = 0; j < vres; j++)
	274	{
	275	UINT32 *d = &bitmap.pix32(j);
	276	for (int i = 0; i < hres; i++)
	277	{
	278	UINT32 pix = *frame_buffer32++;
	279	if (gamma \|\| gamma_dither)
	280	{
	281	int r = (pix >> 24) & 0xff;
	282	int g = (pix >> 16) & 0xff;
	283	int b = (pix >> 8) & 0xff;
	284	int dith = 0;
	285	if (gamma_dither)
	286	{
	287	dith = GetRandom() & 0x3f;
	288	}
	289	if (gamma)
	290	{
	291	if (gamma_dither)
	292	{
	293	r = m_gamma_dither_table[(r << 6)\| dith];
	294	g = m_gamma_dither_table[(g << 6)\| dith];
	295	b = m_gamma_dither_table[(b << 6)\| dith];
	296	}
	297	else
	298	{
	299	r = m_gamma_table[r];
	300	g = m_gamma_table[g];
	301	b = m_gamma_table[b];
	302	}
	303	}
	304	else if (gamma_dither)
	305	{
	306	if (r < 255)
	307	r += (dith & 1);
	308	if (g < 255)
	309	g += (dith & 1);
	310	if (b < 255)
	311	b += (dith & 1);
	312	}
	313	pix = (r << 24) \| (g << 16) \| (b << 8);
	314	}
	315
	316
	317	d[i] = (pix >> 8);
	318	}
	319	frame_buffer32 += invisiblewidth;
	320	}
	321	}
	322	}
	323
	324	/*****************************************************************************/
	325
	326	void n64_rdp::TCDivNoPersp(INT32 ss, INT32 st, INT32 sw, INT32* sss, INT32* sst)
	327	{
	328	*sss = (SIGN16(ss)) & 0x1ffff;
	329	*sst = (SIGN16(st)) & 0x1ffff;
	330	}
	331
	332	void n64_rdp::TCDiv(INT32 ss, INT32 st, INT32 sw, INT32* sss, INT32* sst)
	333	{
	334	int w_carry = 0;
	335	if ((sw & 0x8000) \|\| !(sw & 0x7fff))
	336	{
	337	w_carry = 1;
	338	}
	339
	340	sw &= 0x7fff;
	341
	342	int shift;
	343	for (shift = 1; shift <= 14 && !((sw << shift) & 0x8000); shift++);
	344	shift -= 1;
	345
	346	int normout = (sw << shift) & 0x3fff;
	347	int wnorm = (normout & 0xff) << 2;
	348	normout >>= 8;
	349
	350	int temppoint = m_norm_point_rom[normout];
	351	int tempslope = m_norm_slope_rom[normout];
	352
	353	int tlu_rcp = ((-(tempslope * wnorm)) >> 10) + temppoint;
	354
	355	int sprod = SIGN16(ss) * tlu_rcp;
	356	int tprod = SIGN16(st) * tlu_rcp;
	357	int tempmask = ((1 << (shift + 1)) - 1) << (29 - shift);
	358	int shift_value = 13 - shift;
	359
	360	int outofbounds_s = sprod & tempmask;
	361	int outofbounds_t = tprod & tempmask;
	362	if (shift == 0xe)
	363	{
	364	*sss = sprod << 1;
	365	*sst = tprod << 1;
	366	}
	367	else
	368	{
	369	*sss = sprod = (sprod >> shift_value);
	370	*sst = tprod = (tprod >> shift_value);
	371	}
	372	//compute clamp flags
	373	int under_s = 0;
	374	int under_t = 0;
	375	int over_s = 0;
	376	int over_t = 0;
	377
	378	if (outofbounds_s != tempmask && outofbounds_s != 0)
	379	{
	380	if (sprod & (1 << 29))
	381	{
	382	under_s = 1;
	383	}
	384	else
	385	{
	386	over_s = 1;
	387	}
	388	}
	389
	390	if (outofbounds_t != tempmask && outofbounds_t != 0)
	391	{
	392	if (tprod & (1 << 29))
	393	{
	394	under_t = 1;
	395	}
	396	else
	397	{
	398	over_t = 1;
	399	}
	400	}
	401
	402	over_s \|= w_carry;
	403	over_t \|= w_carry;
	404
	405	sss = (sss & 0x1ffff) \| (over_s << 18) \| (under_s << 17);
	406	sst = (sst & 0x1ffff) \| (over_t << 18) \| (under_t << 17);
	407	}
	408
	409	INT32 n64_rdp::ColorCombinerEquation(INT32 a, INT32 b, INT32 c, INT32 d)
	410	{
	411	a = KURT_AKELEY_SIGN9(a);
	412	b = KURT_AKELEY_SIGN9(b);
	413	c = SIGN9(c);
	414	d = KURT_AKELEY_SIGN9(d);
	415	a = (((a - b) * c) + (d << 8) + 0x80);
	416	a = SIGN17(a) >> 8;
	417	a = s_special_9bit_clamptable[a & 0x1ff];
	418	return a;
	419	}
	420
	421	INT32 n64_rdp::AlphaCombinerEquation(INT32 a, INT32 b, INT32 c, INT32 d)
	422	{
	423	a = KURT_AKELEY_SIGN9(a);
	424	b = KURT_AKELEY_SIGN9(b);
	425	c = SIGN9(c);
	426	d = KURT_AKELEY_SIGN9(d);
	427	a = (((a - b) * c) + (d << 8) + 0x80) >> 8;
	428	a = SIGN9(a);
	429	a = s_special_9bit_clamptable[a & 0x1ff];
	430	return a;
	431	}
	432
	433	void n64_rdp::SetSubAInputRGB(UINT8 input_r, UINT8 input_g, UINT8 *input_b, int code, rdp_span_aux userdata)
	434	{
	435	switch (code & 0xf)
	436	{
	437	case 0: input_r = &userdata->CombinedColor.i.r; input_g = &userdata->CombinedColor.i.g; *input_b = &userdata->CombinedColor.i.b; break;
	438	case 1: input_r = &userdata->Texel0Color.i.r; input_g = &userdata->Texel0Color.i.g; *input_b = &userdata->Texel0Color.i.b; break;
	439	case 2: input_r = &userdata->Texel1Color.i.r; input_g = &userdata->Texel1Color.i.g; *input_b = &userdata->Texel1Color.i.b; break;
	440	case 3: input_r = &userdata->PrimColor.i.r; input_g = &userdata->PrimColor.i.g; *input_b = &userdata->PrimColor.i.b; break;
	441	case 4: input_r = &userdata->ShadeColor.i.r; input_g = &userdata->ShadeColor.i.g; *input_b = &userdata->ShadeColor.i.b; break;
	442	case 5: input_r = &userdata->EnvColor.i.r; input_g = &userdata->EnvColor.i.g; *input_b = &userdata->EnvColor.i.b; break;
	443	case 6: input_r = &OneColor.i.r; input_g = &OneColor.i.g; *input_b = &OneColor.i.b; break;
	444	case 7: input_r = &userdata->NoiseColor.i.r; input_g = &userdata->NoiseColor.i.g; *input_b = &userdata->NoiseColor.i.b; break;
	445	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
	446	{
	447	input_r = &ZeroColor.i.r; input_g = &ZeroColor.i.g; *input_b = &ZeroColor.i.b; break;
	448	}
	449	}
	450	}
	451
	452	void n64_rdp::SetSubBInputRGB(UINT8 input_r, UINT8 input_g, UINT8 *input_b, int code, rdp_span_aux userdata)
	453	{
	454	switch (code & 0xf)
	455	{
	456	case 0: input_r = &userdata->CombinedColor.i.r; input_g = &userdata->CombinedColor.i.g; *input_b = &userdata->CombinedColor.i.b; break;
	457	case 1: input_r = &userdata->Texel0Color.i.r; input_g = &userdata->Texel0Color.i.g; *input_b = &userdata->Texel0Color.i.b; break;
	458	case 2: input_r = &userdata->Texel1Color.i.r; input_g = &userdata->Texel1Color.i.g; *input_b = &userdata->Texel1Color.i.b; break;
	459	case 3: input_r = &userdata->PrimColor.i.r; input_g = &userdata->PrimColor.i.g; *input_b = &userdata->PrimColor.i.b; break;
	460	case 4: input_r = &userdata->ShadeColor.i.r; input_g = &userdata->ShadeColor.i.g; *input_b = &userdata->ShadeColor.i.b; break;
	461	case 5: input_r = &userdata->EnvColor.i.r; input_g = &userdata->EnvColor.i.g; *input_b = &userdata->EnvColor.i.b; break;
	462	case 6: fatalerror("SET_SUBB_RGB_INPUT: key_center\n");
	463	case 7: input_r = (UINT8)&m_k4; input_g = (UINT8)&m_k4; input_b = (UINT8)&m_k4; break;
	464	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
	465	{
	466	input_r = &ZeroColor.i.r; input_g = &ZeroColor.i.g; *input_b = &ZeroColor.i.b; break;
	467	}
	468	}
	469	}
	470
	471	void n64_rdp::SetMulInputRGB(UINT8 input_r, UINT8 input_g, UINT8 *input_b, int code, rdp_span_aux userdata)
	472	{
	473	switch (code & 0x1f)
	474	{
	475	case 0: input_r = &userdata->CombinedColor.i.r; input_g = &userdata->CombinedColor.i.g; *input_b = &userdata->CombinedColor.i.b; break;
	476	case 1: input_r = &userdata->Texel0Color.i.r; input_g = &userdata->Texel0Color.i.g; *input_b = &userdata->Texel0Color.i.b; break;
	477	case 2: input_r = &userdata->Texel1Color.i.r; input_g = &userdata->Texel1Color.i.g; *input_b = &userdata->Texel1Color.i.b; break;
	478	case 3: input_r = &userdata->PrimColor.i.r; input_g = &userdata->PrimColor.i.g; *input_b = &userdata->PrimColor.i.b; break;
	479	case 4: input_r = &userdata->ShadeColor.i.r; input_g = &userdata->ShadeColor.i.g; *input_b = &userdata->ShadeColor.i.b; break;
	480	case 5: input_r = &userdata->EnvColor.i.r; input_g = &userdata->EnvColor.i.g; *input_b = &userdata->EnvColor.i.b; break;
	481	case 6: input_r = &userdata->KeyScale.i.r; input_g = &userdata->KeyScale.i.g; *input_b = &userdata->KeyScale.i.b; break;
	482	case 7: input_r = &userdata->CombinedColor.i.a; input_g = &userdata->CombinedColor.i.a; *input_b = &userdata->CombinedColor.i.a; break;
	483	case 8: input_r = &userdata->Texel0Color.i.a; input_g = &userdata->Texel0Color.i.a; *input_b = &userdata->Texel0Color.i.a; break;
	484	case 9: input_r = &userdata->Texel1Color.i.a; input_g = &userdata->Texel1Color.i.a; *input_b = &userdata->Texel1Color.i.a; break;
	485	case 10: input_r = &userdata->PrimColor.i.a; input_g = &userdata->PrimColor.i.a; *input_b = &userdata->PrimColor.i.a; break;
	486	case 11: input_r = &userdata->ShadeColor.i.a; input_g = &userdata->ShadeColor.i.a; *input_b = &userdata->ShadeColor.i.a; break;
	487	case 12: input_r = &userdata->EnvColor.i.a; input_g = &userdata->EnvColor.i.a; *input_b = &userdata->EnvColor.i.a; break;
	488	case 13: input_r = &userdata->LODFraction; input_g = &userdata->LODFraction; *input_b = &userdata->LODFraction; break;
	489	case 14: input_r = &userdata->PrimLODFraction; input_g = &userdata->PrimLODFraction; *input_b = &userdata->PrimLODFraction; break;
	490	case 15: input_r = (UINT8)&m_k5; input_g = (UINT8)&m_k5; input_b = (UINT8)&m_k5; break;
	491	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
	492	case 24: case 25: case 26: case 27: case 28: case 29: case 30: case 31:
	493	{
	494	input_r = &ZeroColor.i.r; input_g = &ZeroColor.i.g; *input_b = &ZeroColor.i.b; break;
	495	}
	496	}
	497	}
	498
	499	void n64_rdp::SetAddInputRGB(UINT8 input_r, UINT8 input_g, UINT8 *input_b, int code, rdp_span_aux userdata)
	500	{
	501	switch (code & 0x7)
	502	{
	503	case 0: input_r = &userdata->CombinedColor.i.r; input_g = &userdata->CombinedColor.i.g; *input_b = &userdata->CombinedColor.i.b; break;
	504	case 1: input_r = &userdata->Texel0Color.i.r; input_g = &userdata->Texel0Color.i.g; *input_b = &userdata->Texel0Color.i.b; break;
	505	case 2: input_r = &userdata->Texel1Color.i.r; input_g = &userdata->Texel1Color.i.g; *input_b = &userdata->Texel1Color.i.b; break;
	506	case 3: input_r = &userdata->PrimColor.i.r; input_g = &userdata->PrimColor.i.g; *input_b = &userdata->PrimColor.i.b; break;
	507	case 4: input_r = &userdata->ShadeColor.i.r; input_g = &userdata->ShadeColor.i.g; *input_b = &userdata->ShadeColor.i.b; break;
	508	case 5: input_r = &userdata->EnvColor.i.r; input_g = &userdata->EnvColor.i.g; *input_b = &userdata->EnvColor.i.b; break;
	509	case 6: input_r = &OneColor.i.r; input_g = &OneColor.i.g; *input_b = &OneColor.i.b; break;
	510	case 7: input_r = &ZeroColor.i.r; input_g = &ZeroColor.i.g; *input_b = &ZeroColor.i.b; break;
	511	}
	512	}
	513
	514	void n64_rdp::SetSubInputAlpha(UINT8 *input, int code, rdp_span_aux userdata)
	515	{
	516	switch (code & 0x7)
	517	{
	518	case 0: *input = &userdata->CombinedColor.i.a; break;
	519	case 1: *input = &userdata->Texel0Color.i.a; break;
	520	case 2: *input = &userdata->Texel1Color.i.a; break;
	521	case 3: *input = &userdata->PrimColor.i.a; break;
	522	case 4: *input = &userdata->ShadeColor.i.a; break;
	523	case 5: *input = &userdata->EnvColor.i.a; break;
	524	case 6: *input = &OneColor.i.a; break;
	525	case 7: *input = &ZeroColor.i.a; break;
	526	}
	527	}
	528
	529	void n64_rdp::SetMulInputAlpha(UINT8 *input, int code, rdp_span_aux userdata)
	530	{
	531	switch (code & 0x7)
	532	{
	533	case 0: *input = &userdata->LODFraction; break;
	534	case 1: *input = &userdata->Texel0Color.i.a; break;
	535	case 2: *input = &userdata->Texel1Color.i.a; break;
	536	case 3: *input = &userdata->PrimColor.i.a; break;
	537	case 4: *input = &userdata->ShadeColor.i.a; break;
	538	case 5: *input = &userdata->EnvColor.i.a; break;
	539	case 6: *input = &userdata->PrimLODFraction; break;
	540	case 7: *input = &ZeroColor.i.a; break;
	541	}
	542	}
	543
	544	void n64_rdp::SetBlenderInput(int cycle, int which, UINT8 input_r, UINT8 input_g, UINT8 input_b, UINT8 input_a, int a, int b, rdp_span_aux *userdata)
	545	{
	546	switch (a & 0x3)
	547	{
	548	case 0:
	549	{
	550	if (cycle == 0)
	551	{
	552	*input_r = &userdata->PixelColor.i.r;
	553	*input_g = &userdata->PixelColor.i.g;
	554	*input_b = &userdata->PixelColor.i.b;
	555	}
	556	else
	557	{
	558	*input_r = &userdata->BlendedPixelColor.i.r;
	559	*input_g = &userdata->BlendedPixelColor.i.g;
	560	*input_b = &userdata->BlendedPixelColor.i.b;
	561	}
	562	break;
	563	}
	564
	565	case 1:
	566	{
	567	*input_r = &userdata->MemoryColor.i.r;
	568	*input_g = &userdata->MemoryColor.i.g;
	569	*input_b = &userdata->MemoryColor.i.b;
	570	break;
	571	}
	572
	573	case 2:
	574	{
	575	*input_r = &userdata->BlendColor.i.r;
	576	*input_g = &userdata->BlendColor.i.g;
	577	*input_b = &userdata->BlendColor.i.b;
	578	break;
	579	}
	580
	581	case 3:
	582	{
	583	*input_r = &userdata->FogColor.i.r;
	584	*input_g = &userdata->FogColor.i.g;
	585	*input_b = &userdata->FogColor.i.b;
	586	break;
	587	}
	588	}
	589
	590	if (which == 0)
	591	{
	592	switch (b & 0x3)
	593	{
	594	case 0: *input_a = &userdata->PixelColor.i.a; break;
	595	case 1: *input_a = &userdata->FogColor.i.a; break;
	596	case 2: *input_a = &userdata->ShadeColor.i.a; break;
	597	case 3: *input_a = &ZeroColor.i.a; break;
	598	}
	599	}
	600	else
	601	{
	602	switch (b & 0x3)
	603	{
	604	case 0: *input_a = &userdata->InvPixelColor.i.a; break;
	605	case 1: *input_a = &userdata->MemoryColor.i.a; break;
	606	case 2: *input_a = &OneColor.i.a; break;
	607	case 3: *input_a = &ZeroColor.i.a; break;
	608	}
	609	}
	610	}
	611
	612	const UINT8 n64_rdp::s_bayer_matrix[16] =
	613	{ /* Bayer matrix */
	614	0, 4, 1, 5,
	615	6, 2, 7, 3,
	616	1, 5, 0, 4,
	617	7, 3, 6, 2
	618	};
	619
	620	const UINT8 n64_rdp::s_magic_matrix[16] =
	621	{ /* Magic square matrix */
	622	0, 6, 1, 7,
	623	4, 2, 5, 3,
	624	3, 5, 2, 4,
	625	7, 1, 6, 0
	626	};
	627
	628	const n64_rdp::ZDecompressEntry n64_rdp::z_dec_table[8] =
	629	{
	630	{ 6, 0x00000 },
	631	{ 5, 0x20000 },
	632	{ 4, 0x30000 },
	633	{ 3, 0x38000 },
	634	{ 2, 0x3c000 },
	635	{ 1, 0x3e000 },
	636	{ 0, 0x3f000 },
	637	{ 0, 0x3f800 },
	638	};
	639
	640	/*****************************************************************************/
	641
	642	void n64_rdp::z_build_com_table(void)
	643	{
	644	UINT16 altmem = 0;
	645	for(int z = 0; z < 0x40000; z++)
	646	{
	647	switch((z >> 11) & 0x7f)
	648	{
	649	case 0x00:
	650	case 0x01:
	651	case 0x02:
	652	case 0x03:
	653	case 0x04:
	654	case 0x05:
	655	case 0x06:
	656	case 0x07:
	657	case 0x08:
	658	case 0x09:
	659	case 0x0a:
	660	case 0x0b:
	661	case 0x0c:
	662	case 0x0d:
	663	case 0x0e:
	664	case 0x0f:
	665	case 0x10:
	666	case 0x11:
	667	case 0x12:
	668	case 0x13:
	669	case 0x14:
	670	case 0x15:
	671	case 0x16:
	672	case 0x17:
	673	case 0x18:
	674	case 0x19:
	675	case 0x1a:
	676	case 0x1b:
	677	case 0x1c:
	678	case 0x1d:
	679	case 0x1e:
	680	case 0x1f:
	681	case 0x20:
	682	case 0x21:
	683	case 0x22:
	684	case 0x23:
	685	case 0x24:
	686	case 0x25:
	687	case 0x26:
	688	case 0x27:
	689	case 0x28:
	690	case 0x29:
	691	case 0x2a:
	692	case 0x2b:
	693	case 0x2c:
	694	case 0x2d:
	695	case 0x2e:
	696	case 0x2f:
	697	case 0x30:
	698	case 0x31:
	699	case 0x32:
	700	case 0x33:
	701	case 0x34:
	702	case 0x35:
	703	case 0x36:
	704	case 0x37:
	705	case 0x38:
	706	case 0x39:
	707	case 0x3a:
	708	case 0x3b:
	709	case 0x3c:
	710	case 0x3d:
	711	case 0x3e:
	712	case 0x3f:
	713	altmem = (z >> 4) & 0x1ffc;
	714	break;
	715	case 0x40:
	716	case 0x41:
	717	case 0x42:
	718	case 0x43:
	719	case 0x44:
	720	case 0x45:
	721	case 0x46:
	722	case 0x47:
	723	case 0x48:
	724	case 0x49:
	725	case 0x4a:
	726	case 0x4b:
	727	case 0x4c:
	728	case 0x4d:
	729	case 0x4e:
	730	case 0x4f:
	731	case 0x50:
	732	case 0x51:
	733	case 0x52:
	734	case 0x53:
	735	case 0x54:
	736	case 0x55:
	737	case 0x56:
	738	case 0x57:
	739	case 0x58:
	740	case 0x59:
	741	case 0x5a:
	742	case 0x5b:
	743	case 0x5c:
	744	case 0x5d:
	745	case 0x5e:
	746	case 0x5f:
	747	altmem = ((z >> 3) & 0x1ffc) \| 0x2000;
	748	break;
	749	case 0x60:
	750	case 0x61:
	751	case 0x62:
	752	case 0x63:
	753	case 0x64:
	754	case 0x65:
	755	case 0x66:
	756	case 0x67:
	757	case 0x68:
	758	case 0x69:
	759	case 0x6a:
	760	case 0x6b:
	761	case 0x6c:
	762	case 0x6d:
	763	case 0x6e:
	764	case 0x6f:
	765	altmem = ((z >> 2) & 0x1ffc) \| 0x4000;
	766	break;
	767	case 0x70:
	768	case 0x71:
	769	case 0x72:
	770	case 0x73:
	771	case 0x74:
	772	case 0x75:
	773	case 0x76:
	774	case 0x77:
	775	altmem = ((z >> 1) & 0x1ffc) \| 0x6000;
	776	break;
	777	case 0x78://uncompressed z = 0x3c000
	778	case 0x79:
	779	case 0x7a:
	780	case 0x7b:
	781	altmem = (z & 0x1ffc) \| 0x8000;
	782	break;
	783	case 0x7c://uncompressed z = 0x3e000
	784	case 0x7d:
	785	altmem = ((z << 1) & 0x1ffc) \| 0xa000;
	786	break;
	787	case 0x7e://uncompressed z = 0x3f000
	788	altmem = ((z << 2) & 0x1ffc) \| 0xc000;
	789	break;
	790	case 0x7f://uncompressed z = 0x3f000
	791	altmem = ((z << 2) & 0x1ffc) \| 0xe000;
	792	break;
	793	}
	794
	795	z_com_table[z] = altmem;
	796
	797	}
	798	}
	799
	800	void n64_rdp::precalc_cvmask_derivatives(void)
	801	{
	802	const UINT8 yarray[16] = {0, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0};
	803	const UINT8 xarray[16] = {0, 3, 2, 2, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0};
	804
	805	for (int i = 0; i < 0x10000; i++)
	806	{
	807	compressed_cvmasks[i] = (i & 1) \| ((i & 4) >> 1) \| ((i & 0x20) >> 3) \| ((i & 0x80) >> 4) \|
	808	((i & 0x100) >> 4) \| ((i & 0x400) >> 5) \| ((i & 0x2000) >> 7) \| ((i & 0x8000) >> 8);
	809	}
	810
	811	for (int i = 0; i < 0x100; i++)
	812	{
	813	UINT16 mask = decompress_cvmask_frombyte(i);
	814	cvarray[i].cvg = cvarray[i].cvbit = 0;
	815	cvarray[i].cvbit = (i >> 7) & 1;
	816	for (int k = 0; k < 8; k++)
	817	{
	818	cvarray[i].cvg += ((i >> k) & 1);
	819	}
	820
	821	UINT16 masky = 0;
	822	for (int k = 0; k < 4; k++)
	823	{
	824	masky \|= ((mask & (0xf000 >> (k << 2))) > 0) << k;
	825	}
	826	UINT8 offy = yarray[masky];
	827
	828	UINT16 maskx = (mask & (0xf000 >> (offy << 2))) >> ((offy ^ 3) << 2);
	829	UINT8 offx = xarray[maskx];
	830
	831	cvarray[i].xoff = offx;
	832	cvarray[i].yoff = offy;
	833	}
	834	}
	835
	836	UINT16 n64_rdp::decompress_cvmask_frombyte(UINT8 x)
	837	{
	838	UINT16 y = (x & 1) \| ((x & 2) << 1) \| ((x & 4) << 3) \| ((x & 8) << 4) \|
	839	((x & 0x10) << 4) \| ((x & 0x20) << 5) \| ((x & 0x40) << 7) \| ((x & 0x80) << 8);
	840	return y;
	841	}
	842
	843	void n64_rdp::lookup_cvmask_derivatives(UINT32 mask, UINT8* offx, UINT8* offy, rdp_span_aux *userdata)
	844	{
	845	UINT32 index = compressed_cvmasks[mask];
	846	userdata->CurrentPixCvg = cvarray[index].cvg;
	847	userdata->CurrentCvgBit = cvarray[index].cvbit;
	848	*offx = cvarray[index].xoff;
	849	*offy = cvarray[index].yoff;
	850	}
	851
	852	void n64_rdp::ZStore(const rdp_poly_state &object, UINT32 zcurpixel, UINT32 dzcurpixel, UINT32 z, UINT32 enc)
	853	{
	854	UINT16 zval = z_com_table[z & 0x3ffff]\|(enc >> 2);
	855	if(zcurpixel <= MEM16_LIMIT)
	856	{
	857	((UINT16*)rdram)[zcurpixel ^ WORD_ADDR_XOR] = zval;
	858	}
	859	if(dzcurpixel <= MEM8_LIMIT)
	860	{
	861	HiddenBits[dzcurpixel ^ BYTE_ADDR_XOR] = enc & 3;
	862	}
	863	}
	864
	865	INT32 n64_rdp::NormalizeDZPix(INT32 sum)
	866	{
	867	if (sum & 0xc000)
	868	{
	869	return 0x8000;
	870	}
	871	if (!(sum & 0xffff))
	872	{
	873	return 1;
	874	}
	875	for(int count = 0x2000; count > 0; count >>= 1)
	876	{
	877	if (sum & count)
	878	{
	879	return(count << 1);
	880	}
	881	}
	882	return 0;
	883	}
	884
	885	UINT32 n64_rdp::ZDecompress(UINT32 zcurpixel)
	886	{
	887	UINT32 zb = RREADIDX16(zcurpixel);
	888	CHECK16(zcurpixel);
	889	return z_complete_dec_table[(zb >> 2) & 0x3fff];
	890	}
	891
	892	UINT32 n64_rdp::DZDecompress(UINT32 zcurpixel, UINT32 dzcurpixel)
	893	{
	894	UINT16 zval = RREADIDX16(zcurpixel);
	895	CHECK16(zcurpixel);
	896	UINT8 dzval = (((dzcurpixel) <= 0x7fffff) ? (HiddenBits[(dzcurpixel) ^ BYTE_ADDR_XOR]) : 0);
	897	UINT32 dz_compressed = ((zval & 3) << 2) \| (dzval & 3);
	898	return (1 << dz_compressed);
	899	}
	900
	901	UINT32 n64_rdp::DZCompress(UINT32 value)
	902	{
	903	INT32 j = 0;
	904	for (; value > 1; j++, value >>= 1);
	905	return j;
	906	}
	907
	908	void n64_rdp::GetDitherValues(int x, int y, int* cdith, int* adith, const rdp_poly_state& object)
	909	{
	910	int dithindex = ((y & 3) << 2) \| (x & 3);
	911	switch((object.OtherModes.rgb_dither_sel << 2) \| object.OtherModes.alpha_dither_sel)
	912	{
	913	case 0:
	914	adith = cdith = s_magic_matrix[dithindex];
	915	break;
	916	case 1:
	917	*cdith = s_magic_matrix[dithindex];
	918	adith = (~(cdith)) & 7;
	919	break;
	920	case 2:
	921	*cdith = s_magic_matrix[dithindex];
	922	*adith = machine().rand() & 7;
	923	break;
	924	case 3:
	925	*cdith = s_magic_matrix[dithindex];
	926	*adith = 0;
	927	break;
	928	case 4:
	929	adith = cdith = s_bayer_matrix[dithindex];
	930	break;
	931	case 5:
	932	*cdith = s_bayer_matrix[dithindex];
	933	adith = (~(cdith)) & 7;
	934	break;
	935	case 6:
	936	*cdith = s_bayer_matrix[dithindex];
	937	*adith = machine().rand() & 7;
	938	break;
	939	case 7:
	940	*cdith = s_bayer_matrix[dithindex];
	941	*adith = 0;
	942	break;
	943	case 8:
	944	*cdith = machine().rand() & 7;
	945	*adith = s_magic_matrix[dithindex];
	946	break;
	947	case 9:
	948	*cdith = machine().rand() & 7;
	949	*adith = (~s_magic_matrix[dithindex]) & 7;
	950	break;
	951	case 10:
	952	*cdith = machine().rand() & 7;
	953	adith = (cdith + 17) & 7;
	954	break;
	955	case 11:
	956	*cdith = machine().rand() & 7;
	957	*adith = 0;
	958	break;
	959	case 12:
	960	*cdith = 0;
	961	*adith = s_bayer_matrix[dithindex];
	962	break;
	963	case 13:
	964	*cdith = 0;
	965	*adith = (~s_bayer_matrix[dithindex]) & 7;
	966	break;
	967	case 14:
	968	*cdith = 0;
	969	*adith = m_machine->rand() & 7;
	970	break;
	971	case 15:
	972	adith = cdith = 0;
	973	break;
	974	}
	975	}
	976
	977	INT32 CLAMP(INT32 in, INT32 min, INT32 max)
	978	{
	979	if(in < min) return min;
	980	if(in > max) return max;
	981	return in;
	982	}
	983
	984	bool n64_rdp::ZCompare(UINT32 zcurpixel, UINT32 dzcurpixel, UINT32 sz, UINT16 dzpix, rdp_span_aux *userdata, const rdp_poly_state &object)
	985	{
	986	bool force_coplanar = false;
	987	sz &= 0x3ffff;
	988
	989	UINT32 oz;
	990	UINT32 dzmem;
	991	UINT32 zval;
	992	INT32 rawdzmem;
	993
	994	if (object.OtherModes.z_compare_en)
	995	{
	996	oz = ZDecompress(zcurpixel);
	997	dzmem = DZDecompress(zcurpixel, dzcurpixel);
	998	zval = RREADIDX16(zcurpixel);
	999	CHECK16(zcurpixel);
	1000	rawdzmem = ((zval & 3) << 2) \| ((((dzcurpixel) <= 0x3fffff) ? (HiddenBits[(dzcurpixel) ^ BYTE_ADDR_XOR]) : 0) & 3);
	1001	}
	1002	else
	1003	{
	1004	oz = 0;
	1005	dzmem = 1 << 0xf;
	1006	zval = 0x3;
	1007	rawdzmem = 0xf;
	1008	}
	1009
	1010	userdata->m_dzpix_enc = DZCompress(dzpix & 0xffff);
	1011	userdata->ShiftA = CLAMP(userdata->m_dzpix_enc - rawdzmem, 0, 4);
	1012	userdata->ShiftB = CLAMP(rawdzmem - userdata->m_dzpix_enc, 0, 4);
	1013
	1014	int precision_factor = (zval >> 13) & 0xf;
	1015	if (precision_factor < 3)
	1016	{
	1017	int dzmemmodifier = 16 >> precision_factor;
	1018	if (dzmem == 0x8000)
	1019	{
	1020	force_coplanar = true;
	1021	}
	1022	dzmem <<= 1;
	1023	if (dzmem <= dzmemmodifier)
	1024	{
	1025	dzmem = dzmemmodifier;
	1026	}
	1027	if (!dzmem)
	1028	{
	1029	dzmem = 0xffff;
	1030	}
	1031	}
	1032	if (dzmem > 0x8000)
	1033	{
	1034	dzmem = 0xffff;
	1035	}
	1036
	1037	UINT32 dznew = (dzmem > dzpix) ? dzmem : (UINT32)dzpix;
	1038	UINT32 dznotshift = dznew;
	1039	dznew <<= 3;
	1040
	1041	bool farther = (sz + dznew) >= oz;
	1042	bool infront = sz < oz;
	1043
	1044	if (force_coplanar)
	1045	{
	1046	farther = true;
	1047	}
	1048
	1049	bool overflow = ((userdata->CurrentMemCvg + userdata->CurrentPixCvg) & 8) > 0;
	1050	userdata->BlendEnable = (object.OtherModes.force_blend \|\| (!overflow && object.OtherModes.antialias_en && farther)) ? 1 : 0;
	1051	userdata->PreWrap = overflow;
	1052
	1053	int cvgcoeff = 0;
	1054	UINT32 dzenc = 0;
	1055
	1056	if (object.OtherModes.z_mode == 1 && infront && farther && overflow)
	1057	{
	1058	dzenc = DZCompress(dznotshift & 0xffff);
	1059	cvgcoeff = ((oz >> dzenc) - (sz >> dzenc)) & 0xf;
	1060	userdata->CurrentPixCvg = ((cvgcoeff * userdata->CurrentPixCvg) >> 3) & 0xf;
	1061	}
	1062
	1063	if (!object.OtherModes.z_compare_en)
	1064	{
	1065	return true;
	1066	}
	1067
	1068	INT32 diff = (INT32)sz - (INT32)dznew;
	1069	bool nearer = diff <= (INT32)oz;
	1070	bool max = (oz == 0x3ffff);
	1071	if (force_coplanar)
	1072	{
	1073	nearer = true;
	1074	}
	1075
	1076	switch(object.OtherModes.z_mode)
	1077	{
	1078	case 0:
	1079	return (max \|\| (overflow ? infront : nearer));
	1080	case 1:
	1081	return (max \|\| (overflow ? infront : nearer));
	1082	case 2:
	1083	return (infront \|\| max);
	1084	case 3:
	1085	return (farther && nearer && !max);
	1086	}
	1087
	1088	return false;
	1089	}
	1090
	1091	UINT32 n64_rdp::GetLog2(UINT32 lod_clamp)
	1092	{
	1093	if (lod_clamp < 2)
	1094	{
	1095	return 0;
	1096	}
	1097	else
	1098	{
	1099	for (int i = 7; i > 0; i--)
	1100	{
	1101	if ((lod_clamp >> i) & 1)
	1102	{
	1103	return i;
	1104	}
	1105	}
	1106	}
	1107
	1108	return 0;
	1109	}
	1110
	1111	/*****************************************************************************/
	1112
	1113	UINT32 n64_rdp::ReadData(UINT32 address)
	1114	{
	1115	if (m_status & 0x1) // XBUS_DMEM_DMA enabled
	1116	{
	1117	return rsp_dmem[(address & 0xfff) / 4];
	1118	}
	1119	else
	1120	{
	1121	return rdram[((address & 0xffffff) / 4)];
	1122	}
	1123	}
	1124
	1125	static const char *const image_format[] = { "RGBA", "YUV", "CI", "IA", "I", "???", "???", "???" };
	1126	static const char *const image_size[] = { "4-bit", "8-bit", "16-bit", "32-bit" };
	1127
	1128	static const int rdp_command_length[64] =
	1129	{
	1130	8, // 0x00, No Op
	1131	8, // 0x01, ???
	1132	8, // 0x02, ???
	1133	8, // 0x03, ???
	1134	8, // 0x04, ???
	1135	8, // 0x05, ???
	1136	8, // 0x06, ???
	1137	8, // 0x07, ???
	1138	32, // 0x08, Non-Shaded Triangle
	1139	32+16, // 0x09, Non-Shaded, Z-Buffered Triangle
	1140	32+64, // 0x0a, Textured Triangle
	1141	32+64+16, // 0x0b, Textured, Z-Buffered Triangle
	1142	32+64, // 0x0c, Shaded Triangle
	1143	32+64+16, // 0x0d, Shaded, Z-Buffered Triangle
	1144	32+64+64, // 0x0e, Shaded+Textured Triangle
	1145	32+64+64+16,// 0x0f, Shaded+Textured, Z-Buffered Triangle
	1146	8, // 0x10, ???
	1147	8, // 0x11, ???
	1148	8, // 0x12, ???
	1149	8, // 0x13, ???
	1150	8, // 0x14, ???
	1151	8, // 0x15, ???
	1152	8, // 0x16, ???
	1153	8, // 0x17, ???
	1154	8, // 0x18, ???
	1155	8, // 0x19, ???
	1156	8, // 0x1a, ???
	1157	8, // 0x1b, ???
	1158	8, // 0x1c, ???
	1159	8, // 0x1d, ???
	1160	8, // 0x1e, ???
	1161	8, // 0x1f, ???
	1162	8, // 0x20, ???
	1163	8, // 0x21, ???
	1164	8, // 0x22, ???
	1165	8, // 0x23, ???
	1166	16, // 0x24, Texture_Rectangle
	1167	16, // 0x25, Texture_Rectangle_Flip
	1168	8, // 0x26, Sync_Load
	1169	8, // 0x27, Sync_Pipe
	1170	8, // 0x28, Sync_Tile
	1171	8, // 0x29, Sync_Full
	1172	8, // 0x2a, Set_Key_GB
	1173	8, // 0x2b, Set_Key_R
	1174	8, // 0x2c, Set_Convert
	1175	8, // 0x2d, Set_Scissor
	1176	8, // 0x2e, Set_Prim_Depth
	1177	8, // 0x2f, Set_Other_Modes
	1178	8, // 0x30, Load_TLUT
	1179	8, // 0x31, ???
	1180	8, // 0x32, Set_Tile_Size
	1181	8, // 0x33, Load_Block
	1182	8, // 0x34, Load_Tile
	1183	8, // 0x35, Set_Tile
	1184	8, // 0x36, Fill_Rectangle
	1185	8, // 0x37, Set_Fill_Color
	1186	8, // 0x38, Set_Fog_Color
	1187	8, // 0x39, Set_Blend_Color
	1188	8, // 0x3a, Set_Prim_Color
	1189	8, // 0x3b, Set_Env_Color
	1190	8, // 0x3c, Set_Combine
	1191	8, // 0x3d, Set_Texture_Image
	1192	8, // 0x3e, Set_Mask_Image
	1193	8 // 0x3f, Set_Color_Image
	1194	};
	1195
	1196	void n64_rdp::Dasm(char *buffer)
	1197	{
	1198	int i;
	1199	int tile;
	1200	const char format, size;
	1201	char sl[32], tl[32], sh[32], th[32];
	1202	char s[32], t[32], w[32];
	1203	char dsdx[32], dtdx[32], dwdx[32];
	1204	char dsdy[32], dtdy[32], dwdy[32];
	1205	char dsde[32], dtde[32], dwde[32];
	1206	char yl[32], yh[32], ym[32], xl[32], xh[32], xm[32];
	1207	char dxldy[32], dxhdy[32], dxmdy[32];
	1208	char rt[32], gt[32], bt[32], at[32];
	1209	char drdx[32], dgdx[32], dbdx[32], dadx[32];
	1210	char drdy[32], dgdy[32], dbdy[32], dady[32];
	1211	char drde[32], dgde[32], dbde[32], dade[32];
	1212	UINT32 r,g,b,a;
	1213
	1214	UINT32 cmd[64];
	1215	UINT32 length;
	1216	UINT32 command;
	1217
	1218	length = m_cmd_ptr * 4;
	1219	if (length < 8)
	1220	{
	1221	sprintf(buffer, "ERROR: length = %d\n", length);
	1222	return;
	1223	}
	1224
	1225	cmd[0] = m_cmd_data[m_cmd_cur+0];
	1226	cmd[1] = m_cmd_data[m_cmd_cur+1];
	1227
	1228	tile = (cmd[1] >> 24) & 0x7;
	1229	sprintf(sl, "%4.2f", (float)((cmd[0] >> 12) & 0xfff) / 4.0f);
	1230	sprintf(tl, "%4.2f", (float)((cmd[0] >> 0) & 0xfff) / 4.0f);
	1231	sprintf(sh, "%4.2f", (float)((cmd[1] >> 12) & 0xfff) / 4.0f);
	1232	sprintf(th, "%4.2f", (float)((cmd[1] >> 0) & 0xfff) / 4.0f);
	1233
	1234	format = image_format[(cmd[0] >> 21) & 0x7];
	1235	size = image_size[(cmd[0] >> 19) & 0x3];
	1236
	1237	r = (cmd[1] >> 24) & 0xff;
	1238	g = (cmd[1] >> 16) & 0xff;
	1239	b = (cmd[1] >> 8) & 0xff;
	1240	a = (cmd[1] >> 0) & 0xff;
	1241
	1242	command = (cmd[0] >> 24) & 0x3f;
	1243	switch (command)
	1244	{
	1245	case 0x00: sprintf(buffer, "No Op"); break;
	1246	case 0x08: // Tri_NoShade
	1247	{
	1248	int lft = (command >> 23) & 0x1;
	1249
	1250	if (length != rdp_command_length[command])
	1251	{
	1252	sprintf(buffer, "ERROR: Tri_NoShade length = %d\n", length);
	1253	return;
	1254	}
	1255
	1256	cmd[2] = m_cmd_data[m_cmd_cur+2];
	1257	cmd[3] = m_cmd_data[m_cmd_cur+3];
	1258	cmd[4] = m_cmd_data[m_cmd_cur+4];
	1259	cmd[5] = m_cmd_data[m_cmd_cur+5];
	1260	cmd[6] = m_cmd_data[m_cmd_cur+6];
	1261	cmd[7] = m_cmd_data[m_cmd_cur+7];
	1262
	1263	sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
	1264	sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
	1265	sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
	1266	sprintf(xl, "%4.4f", (float)(cmd[2] / 65536.0f));
	1267	sprintf(dxldy, "%4.4f", (float)(cmd[3] / 65536.0f));
	1268	sprintf(xh, "%4.4f", (float)(cmd[4] / 65536.0f));
	1269	sprintf(dxhdy, "%4.4f", (float)(cmd[5] / 65536.0f));
	1270	sprintf(xm, "%4.4f", (float)(cmd[6] / 65536.0f));
	1271	sprintf(dxmdy, "%4.4f", (float)(cmd[7] / 65536.0f));
	1272
	1273	sprintf(buffer, "Tri_NoShade %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
	1274	break;
	1275	}
	1276	case 0x09: // Tri_NoShadeZ
	1277	{
	1278	int lft = (command >> 23) & 0x1;
	1279
	1280	if (length != rdp_command_length[command])
	1281	{
	1282	sprintf(buffer, "ERROR: Tri_NoShadeZ length = %d\n", length);
	1283	return;
	1284	}
	1285
	1286	cmd[2] = m_cmd_data[m_cmd_cur+2];
	1287	cmd[3] = m_cmd_data[m_cmd_cur+3];
	1288	cmd[4] = m_cmd_data[m_cmd_cur+4];
	1289	cmd[5] = m_cmd_data[m_cmd_cur+5];
	1290	cmd[6] = m_cmd_data[m_cmd_cur+6];
	1291	cmd[7] = m_cmd_data[m_cmd_cur+7];
	1292
	1293	sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
	1294	sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
	1295	sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
	1296	sprintf(xl, "%4.4f", (float)(cmd[2] / 65536.0f));
	1297	sprintf(dxldy, "%4.4f", (float)(cmd[3] / 65536.0f));
	1298	sprintf(xh, "%4.4f", (float)(cmd[4] / 65536.0f));
	1299	sprintf(dxhdy, "%4.4f", (float)(cmd[5] / 65536.0f));
	1300	sprintf(xm, "%4.4f", (float)(cmd[6] / 65536.0f));
	1301	sprintf(dxmdy, "%4.4f", (float)(cmd[7] / 65536.0f));
	1302
	1303	sprintf(buffer, "Tri_NoShadeZ %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
	1304	break;
	1305	}
	1306	case 0x0a: // Tri_Tex
	1307	{
	1308	int lft = (command >> 23) & 0x1;
	1309
	1310	if (length < rdp_command_length[command])
	1311	{
	1312	sprintf(buffer, "ERROR: Tri_Tex length = %d\n", length);
	1313	return;
	1314	}
	1315
	1316	for (i=2; i < 24; i++)
	1317	{
	1318	cmd[i] = m_cmd_data[m_cmd_cur+i];
	1319	}
	1320
	1321	sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
	1322	sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
	1323	sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
	1324	sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
	1325	sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
	1326	sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
	1327	sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
	1328	sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
	1329	sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
	1330
	1331	sprintf(s, "%4.4f", (float)(INT32)((cmd[ 8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
	1332	sprintf(t, "%4.4f", (float)(INT32)(((cmd[ 8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
	1333	sprintf(w, "%4.4f", (float)(INT32)((cmd[ 9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
	1334	sprintf(dsdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
	1335	sprintf(dtdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
	1336	sprintf(dwdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
	1337	sprintf(dsde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
	1338	sprintf(dtde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
	1339	sprintf(dwde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
	1340	sprintf(dsdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
	1341	sprintf(dtdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
	1342	sprintf(dwdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
	1343
	1344
	1345	buffer+=sprintf(buffer, "Tri_Tex %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
	1346	buffer+=sprintf(buffer, " ");
	1347	buffer+=sprintf(buffer, " S: %s, T: %s, W: %s\n", s, t, w);
	1348	buffer+=sprintf(buffer, " ");
	1349	buffer+=sprintf(buffer, " DSDX: %s, DTDX: %s, DWDX: %s\n", dsdx, dtdx, dwdx);
	1350	buffer+=sprintf(buffer, " ");
	1351	buffer+=sprintf(buffer, " DSDE: %s, DTDE: %s, DWDE: %s\n", dsde, dtde, dwde);
	1352	buffer+=sprintf(buffer, " ");
	1353	buffer+=sprintf(buffer, " DSDY: %s, DTDY: %s, DWDY: %s\n", dsdy, dtdy, dwdy);
	1354	break;
	1355	}
	1356	case 0x0b: // Tri_TexZ
	1357	{
	1358	int lft = (command >> 23) & 0x1;
	1359
	1360	if (length < rdp_command_length[command])
	1361	{
	1362	sprintf(buffer, "ERROR: Tri_TexZ length = %d\n", length);
	1363	return;
	1364	}
	1365
	1366	for (i=2; i < 24; i++)
	1367	{
	1368	cmd[i] = m_cmd_data[m_cmd_cur+i];
	1369	}
	1370
	1371	sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
	1372	sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
	1373	sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
	1374	sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
	1375	sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
	1376	sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
	1377	sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
	1378	sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
	1379	sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
	1380
	1381	sprintf(s, "%4.4f", (float)(INT32)((cmd[ 8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
	1382	sprintf(t, "%4.4f", (float)(INT32)(((cmd[ 8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
	1383	sprintf(w, "%4.4f", (float)(INT32)((cmd[ 9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
	1384	sprintf(dsdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
	1385	sprintf(dtdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
	1386	sprintf(dwdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
	1387	sprintf(dsde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
	1388	sprintf(dtde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
	1389	sprintf(dwde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
	1390	sprintf(dsdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
	1391	sprintf(dtdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
	1392	sprintf(dwdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
	1393
	1394
	1395	buffer+=sprintf(buffer, "Tri_TexZ %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
	1396	buffer+=sprintf(buffer, " ");
	1397	buffer+=sprintf(buffer, " S: %s, T: %s, W: %s\n", s, t, w);
	1398	buffer+=sprintf(buffer, " ");
	1399	buffer+=sprintf(buffer, " DSDX: %s, DTDX: %s, DWDX: %s\n", dsdx, dtdx, dwdx);
	1400	buffer+=sprintf(buffer, " ");
	1401	buffer+=sprintf(buffer, " DSDE: %s, DTDE: %s, DWDE: %s\n", dsde, dtde, dwde);
	1402	buffer+=sprintf(buffer, " ");
	1403	buffer+=sprintf(buffer, " DSDY: %s, DTDY: %s, DWDY: %s\n", dsdy, dtdy, dwdy);
	1404	break;
	1405	}
	1406	case 0x0c: // Tri_Shade
	1407	{
	1408	int lft = (command >> 23) & 0x1;
	1409
	1410	if (length != rdp_command_length[command])
	1411	{
	1412	sprintf(buffer, "ERROR: Tri_Shade length = %d\n", length);
	1413	return;
	1414	}
	1415
	1416	for (i=2; i < 24; i++)
	1417	{
	1418	cmd[i] = m_cmd_data[i];
	1419	}
	1420
	1421	sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
	1422	sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
	1423	sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
	1424	sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
	1425	sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
	1426	sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
	1427	sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
	1428	sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
	1429	sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
	1430	sprintf(rt, "%4.4f", (float)(INT32)((cmd[8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
	1431	sprintf(gt, "%4.4f", (float)(INT32)(((cmd[8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
	1432	sprintf(bt, "%4.4f", (float)(INT32)((cmd[9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
	1433	sprintf(at, "%4.4f", (float)(INT32)(((cmd[9] & 0xffff) << 16) \| (cmd[13] & 0xffff)) / 65536.0f);
	1434	sprintf(drdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
	1435	sprintf(dgdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
	1436	sprintf(dbdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
	1437	sprintf(dadx, "%4.4f", (float)(INT32)(((cmd[11] & 0xffff) << 16) \| (cmd[15] & 0xffff)) / 65536.0f);
	1438	sprintf(drde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
	1439	sprintf(dgde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
	1440	sprintf(dbde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
	1441	sprintf(dade, "%4.4f", (float)(INT32)(((cmd[17] & 0xffff) << 16) \| (cmd[21] & 0xffff)) / 65536.0f);
	1442	sprintf(drdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
	1443	sprintf(dgdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
	1444	sprintf(dbdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
	1445	sprintf(dady, "%4.4f", (float)(INT32)(((cmd[19] & 0xffff) << 16) \| (cmd[23] & 0xffff)) / 65536.0f);
	1446
	1447	buffer+=sprintf(buffer, "Tri_Shade %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
	1448	buffer+=sprintf(buffer, " ");
	1449	buffer+=sprintf(buffer, " R: %s, G: %s, B: %s, A: %s\n", rt, gt, bt, at);
	1450	buffer+=sprintf(buffer, " ");
	1451	buffer+=sprintf(buffer, " DRDX: %s, DGDX: %s, DBDX: %s, DADX: %s\n", drdx, dgdx, dbdx, dadx);
	1452	buffer+=sprintf(buffer, " ");
	1453	buffer+=sprintf(buffer, " DRDE: %s, DGDE: %s, DBDE: %s, DADE: %s\n", drde, dgde, dbde, dade);
	1454	buffer+=sprintf(buffer, " ");
	1455	buffer+=sprintf(buffer, " DRDY: %s, DGDY: %s, DBDY: %s, DADY: %s\n", drdy, dgdy, dbdy, dady);
	1456	break;
	1457	}
	1458	case 0x0d: // Tri_ShadeZ
	1459	{
	1460	int lft = (command >> 23) & 0x1;
	1461
	1462	if (length != rdp_command_length[command])
	1463	{
	1464	sprintf(buffer, "ERROR: Tri_ShadeZ length = %d\n", length);
	1465	return;
	1466	}
	1467
	1468	for (i=2; i < 24; i++)
	1469	{
	1470	cmd[i] = m_cmd_data[i];
	1471	}
	1472
	1473	sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
	1474	sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
	1475	sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
	1476	sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
	1477	sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
	1478	sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
	1479	sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
	1480	sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
	1481	sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
	1482	sprintf(rt, "%4.4f", (float)(INT32)((cmd[8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
	1483	sprintf(gt, "%4.4f", (float)(INT32)(((cmd[8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
	1484	sprintf(bt, "%4.4f", (float)(INT32)((cmd[9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
	1485	sprintf(at, "%4.4f", (float)(INT32)(((cmd[9] & 0xffff) << 16) \| (cmd[13] & 0xffff)) / 65536.0f);
	1486	sprintf(drdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
	1487	sprintf(dgdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
	1488	sprintf(dbdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
	1489	sprintf(dadx, "%4.4f", (float)(INT32)(((cmd[11] & 0xffff) << 16) \| (cmd[15] & 0xffff)) / 65536.0f);
	1490	sprintf(drde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
	1491	sprintf(dgde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
	1492	sprintf(dbde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
	1493	sprintf(dade, "%4.4f", (float)(INT32)(((cmd[17] & 0xffff) << 16) \| (cmd[21] & 0xffff)) / 65536.0f);
	1494	sprintf(drdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
	1495	sprintf(dgdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
	1496	sprintf(dbdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
	1497	sprintf(dady, "%4.4f", (float)(INT32)(((cmd[19] & 0xffff) << 16) \| (cmd[23] & 0xffff)) / 65536.0f);
	1498
	1499	buffer+=sprintf(buffer, "Tri_ShadeZ %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
	1500	buffer+=sprintf(buffer, " ");
	1501	buffer+=sprintf(buffer, " R: %s, G: %s, B: %s, A: %s\n", rt, gt, bt, at);
	1502	buffer+=sprintf(buffer, " ");
	1503	buffer+=sprintf(buffer, " DRDX: %s, DGDX: %s, DBDX: %s, DADX: %s\n", drdx, dgdx, dbdx, dadx);
	1504	buffer+=sprintf(buffer, " ");
	1505	buffer+=sprintf(buffer, " DRDE: %s, DGDE: %s, DBDE: %s, DADE: %s\n", drde, dgde, dbde, dade);
	1506	buffer+=sprintf(buffer, " ");
	1507	buffer+=sprintf(buffer, " DRDY: %s, DGDY: %s, DBDY: %s, DADY: %s\n", drdy, dgdy, dbdy, dady);
	1508	break;
	1509	}
	1510	case 0x0e: // Tri_TexShade
	1511	{
	1512	int lft = (command >> 23) & 0x1;
	1513
	1514	if (length < rdp_command_length[command])
	1515	{
	1516	sprintf(buffer, "ERROR: Tri_TexShade length = %d\n", length);
	1517	return;
	1518	}
	1519
	1520	for (i=2; i < 40; i++)
	1521	{
	1522	cmd[i] = m_cmd_data[m_cmd_cur+i];
	1523	}
	1524
	1525	sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
	1526	sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
	1527	sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
	1528	sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
	1529	sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
	1530	sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
	1531	sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
	1532	sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
	1533	sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
	1534	sprintf(rt, "%4.4f", (float)(INT32)((cmd[8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
	1535	sprintf(gt, "%4.4f", (float)(INT32)(((cmd[8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
	1536	sprintf(bt, "%4.4f", (float)(INT32)((cmd[9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
	1537	sprintf(at, "%4.4f", (float)(INT32)(((cmd[9] & 0xffff) << 16) \| (cmd[13] & 0xffff)) / 65536.0f);
	1538	sprintf(drdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
	1539	sprintf(dgdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
	1540	sprintf(dbdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
	1541	sprintf(dadx, "%4.4f", (float)(INT32)(((cmd[11] & 0xffff) << 16) \| (cmd[15] & 0xffff)) / 65536.0f);
	1542	sprintf(drde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
	1543	sprintf(dgde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
	1544	sprintf(dbde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
	1545	sprintf(dade, "%4.4f", (float)(INT32)(((cmd[17] & 0xffff) << 16) \| (cmd[21] & 0xffff)) / 65536.0f);
	1546	sprintf(drdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
	1547	sprintf(dgdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
	1548	sprintf(dbdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
	1549	sprintf(dady, "%4.4f", (float)(INT32)(((cmd[19] & 0xffff) << 16) \| (cmd[23] & 0xffff)) / 65536.0f);
	1550
	1551	sprintf(s, "%4.4f", (float)(INT32)((cmd[24] & 0xffff0000) \| ((cmd[28] >> 16) & 0xffff)) / 65536.0f);
	1552	sprintf(t, "%4.4f", (float)(INT32)(((cmd[24] & 0xffff) << 16) \| (cmd[28] & 0xffff)) / 65536.0f);
	1553	sprintf(w, "%4.4f", (float)(INT32)((cmd[25] & 0xffff0000) \| ((cmd[29] >> 16) & 0xffff)) / 65536.0f);
	1554	sprintf(dsdx, "%4.4f", (float)(INT32)((cmd[26] & 0xffff0000) \| ((cmd[30] >> 16) & 0xffff)) / 65536.0f);
	1555	sprintf(dtdx, "%4.4f", (float)(INT32)(((cmd[26] & 0xffff) << 16) \| (cmd[30] & 0xffff)) / 65536.0f);
	1556	sprintf(dwdx, "%4.4f", (float)(INT32)((cmd[27] & 0xffff0000) \| ((cmd[31] >> 16) & 0xffff)) / 65536.0f);
	1557	sprintf(dsde, "%4.4f", (float)(INT32)((cmd[32] & 0xffff0000) \| ((cmd[36] >> 16) & 0xffff)) / 65536.0f);
	1558	sprintf(dtde, "%4.4f", (float)(INT32)(((cmd[32] & 0xffff) << 16) \| (cmd[36] & 0xffff)) / 65536.0f);
	1559	sprintf(dwde, "%4.4f", (float)(INT32)((cmd[33] & 0xffff0000) \| ((cmd[37] >> 16) & 0xffff)) / 65536.0f);
	1560	sprintf(dsdy, "%4.4f", (float)(INT32)((cmd[34] & 0xffff0000) \| ((cmd[38] >> 16) & 0xffff)) / 65536.0f);
	1561	sprintf(dtdy, "%4.4f", (float)(INT32)(((cmd[34] & 0xffff) << 16) \| (cmd[38] & 0xffff)) / 65536.0f);
	1562	sprintf(dwdy, "%4.4f", (float)(INT32)((cmd[35] & 0xffff0000) \| ((cmd[39] >> 16) & 0xffff)) / 65536.0f);
	1563
	1564
	1565	buffer+=sprintf(buffer, "Tri_TexShade %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
	1566	buffer+=sprintf(buffer, " ");
	1567	buffer+=sprintf(buffer, " R: %s, G: %s, B: %s, A: %s\n", rt, gt, bt, at);
	1568	buffer+=sprintf(buffer, " ");
	1569	buffer+=sprintf(buffer, " DRDX: %s, DGDX: %s, DBDX: %s, DADX: %s\n", drdx, dgdx, dbdx, dadx);
	1570	buffer+=sprintf(buffer, " ");
	1571	buffer+=sprintf(buffer, " DRDE: %s, DGDE: %s, DBDE: %s, DADE: %s\n", drde, dgde, dbde, dade);
	1572	buffer+=sprintf(buffer, " ");
	1573	buffer+=sprintf(buffer, " DRDY: %s, DGDY: %s, DBDY: %s, DADY: %s\n", drdy, dgdy, dbdy, dady);
	1574
	1575	buffer+=sprintf(buffer, " ");
	1576	buffer+=sprintf(buffer, " S: %s, T: %s, W: %s\n", s, t, w);
	1577	buffer+=sprintf(buffer, " ");
	1578	buffer+=sprintf(buffer, " DSDX: %s, DTDX: %s, DWDX: %s\n", dsdx, dtdx, dwdx);
	1579	buffer+=sprintf(buffer, " ");
	1580	buffer+=sprintf(buffer, " DSDE: %s, DTDE: %s, DWDE: %s\n", dsde, dtde, dwde);
	1581	buffer+=sprintf(buffer, " ");
	1582	buffer+=sprintf(buffer, " DSDY: %s, DTDY: %s, DWDY: %s\n", dsdy, dtdy, dwdy);
	1583	break;
	1584	}
	1585	case 0x0f: // Tri_TexShadeZ
	1586	{
	1587	int lft = (command >> 23) & 0x1;
	1588
	1589	if (length < rdp_command_length[command])
	1590	{
	1591	sprintf(buffer, "ERROR: Tri_TexShadeZ length = %d\n", length);
	1592	return;
	1593	}
	1594
	1595	for (i=2; i < 40; i++)
	1596	{
	1597	cmd[i] = m_cmd_data[m_cmd_cur+i];
	1598	}
	1599
	1600	sprintf(yl, "%4.4f", (float)((cmd[0] >> 0) & 0x1fff) / 4.0f);
	1601	sprintf(ym, "%4.4f", (float)((cmd[1] >> 16) & 0x1fff) / 4.0f);
	1602	sprintf(yh, "%4.4f", (float)((cmd[1] >> 0) & 0x1fff) / 4.0f);
	1603	sprintf(xl, "%4.4f", (float)((INT32)cmd[2] / 65536.0f));
	1604	sprintf(dxldy, "%4.4f", (float)((INT32)cmd[3] / 65536.0f));
	1605	sprintf(xh, "%4.4f", (float)((INT32)cmd[4] / 65536.0f));
	1606	sprintf(dxhdy, "%4.4f", (float)((INT32)cmd[5] / 65536.0f));
	1607	sprintf(xm, "%4.4f", (float)((INT32)cmd[6] / 65536.0f));
	1608	sprintf(dxmdy, "%4.4f", (float)((INT32)cmd[7] / 65536.0f));
	1609	sprintf(rt, "%4.4f", (float)(INT32)((cmd[8] & 0xffff0000) \| ((cmd[12] >> 16) & 0xffff)) / 65536.0f);
	1610	sprintf(gt, "%4.4f", (float)(INT32)(((cmd[8] & 0xffff) << 16) \| (cmd[12] & 0xffff)) / 65536.0f);
	1611	sprintf(bt, "%4.4f", (float)(INT32)((cmd[9] & 0xffff0000) \| ((cmd[13] >> 16) & 0xffff)) / 65536.0f);
	1612	sprintf(at, "%4.4f", (float)(INT32)(((cmd[9] & 0xffff) << 16) \| (cmd[13] & 0xffff)) / 65536.0f);
	1613	sprintf(drdx, "%4.4f", (float)(INT32)((cmd[10] & 0xffff0000) \| ((cmd[14] >> 16) & 0xffff)) / 65536.0f);
	1614	sprintf(dgdx, "%4.4f", (float)(INT32)(((cmd[10] & 0xffff) << 16) \| (cmd[14] & 0xffff)) / 65536.0f);
	1615	sprintf(dbdx, "%4.4f", (float)(INT32)((cmd[11] & 0xffff0000) \| ((cmd[15] >> 16) & 0xffff)) / 65536.0f);
	1616	sprintf(dadx, "%4.4f", (float)(INT32)(((cmd[11] & 0xffff) << 16) \| (cmd[15] & 0xffff)) / 65536.0f);
	1617	sprintf(drde, "%4.4f", (float)(INT32)((cmd[16] & 0xffff0000) \| ((cmd[20] >> 16) & 0xffff)) / 65536.0f);
	1618	sprintf(dgde, "%4.4f", (float)(INT32)(((cmd[16] & 0xffff) << 16) \| (cmd[20] & 0xffff)) / 65536.0f);
	1619	sprintf(dbde, "%4.4f", (float)(INT32)((cmd[17] & 0xffff0000) \| ((cmd[21] >> 16) & 0xffff)) / 65536.0f);
	1620	sprintf(dade, "%4.4f", (float)(INT32)(((cmd[17] & 0xffff) << 16) \| (cmd[21] & 0xffff)) / 65536.0f);
	1621	sprintf(drdy, "%4.4f", (float)(INT32)((cmd[18] & 0xffff0000) \| ((cmd[22] >> 16) & 0xffff)) / 65536.0f);
	1622	sprintf(dgdy, "%4.4f", (float)(INT32)(((cmd[18] & 0xffff) << 16) \| (cmd[22] & 0xffff)) / 65536.0f);
	1623	sprintf(dbdy, "%4.4f", (float)(INT32)((cmd[19] & 0xffff0000) \| ((cmd[23] >> 16) & 0xffff)) / 65536.0f);
	1624	sprintf(dady, "%4.4f", (float)(INT32)(((cmd[19] & 0xffff) << 16) \| (cmd[23] & 0xffff)) / 65536.0f);
	1625
	1626	sprintf(s, "%4.4f", (float)(INT32)((cmd[24] & 0xffff0000) \| ((cmd[28] >> 16) & 0xffff)) / 65536.0f);
	1627	sprintf(t, "%4.4f", (float)(INT32)(((cmd[24] & 0xffff) << 16) \| (cmd[28] & 0xffff)) / 65536.0f);
	1628	sprintf(w, "%4.4f", (float)(INT32)((cmd[25] & 0xffff0000) \| ((cmd[29] >> 16) & 0xffff)) / 65536.0f);
	1629	sprintf(dsdx, "%4.4f", (float)(INT32)((cmd[26] & 0xffff0000) \| ((cmd[30] >> 16) & 0xffff)) / 65536.0f);
	1630	sprintf(dtdx, "%4.4f", (float)(INT32)(((cmd[26] & 0xffff) << 16) \| (cmd[30] & 0xffff)) / 65536.0f);
	1631	sprintf(dwdx, "%4.4f", (float)(INT32)((cmd[27] & 0xffff0000) \| ((cmd[31] >> 16) & 0xffff)) / 65536.0f);
	1632	sprintf(dsde, "%4.4f", (float)(INT32)((cmd[32] & 0xffff0000) \| ((cmd[36] >> 16) & 0xffff)) / 65536.0f);
	1633	sprintf(dtde, "%4.4f", (float)(INT32)(((cmd[32] & 0xffff) << 16) \| (cmd[36] & 0xffff)) / 65536.0f);
	1634	sprintf(dwde, "%4.4f", (float)(INT32)((cmd[33] & 0xffff0000) \| ((cmd[37] >> 16) & 0xffff)) / 65536.0f);
	1635	sprintf(dsdy, "%4.4f", (float)(INT32)((cmd[34] & 0xffff0000) \| ((cmd[38] >> 16) & 0xffff)) / 65536.0f);
	1636	sprintf(dtdy, "%4.4f", (float)(INT32)(((cmd[34] & 0xffff) << 16) \| (cmd[38] & 0xffff)) / 65536.0f);
	1637	sprintf(dwdy, "%4.4f", (float)(INT32)((cmd[35] & 0xffff0000) \| ((cmd[39] >> 16) & 0xffff)) / 65536.0f);
	1638
	1639
	1640	buffer+=sprintf(buffer, "Tri_TexShadeZ %d, XL: %s, XM: %s, XH: %s, YL: %s, YM: %s, YH: %s\n", lft, xl,xm,xh,yl,ym,yh);
	1641	buffer+=sprintf(buffer, " ");
	1642	buffer+=sprintf(buffer, " R: %s, G: %s, B: %s, A: %s\n", rt, gt, bt, at);
	1643	buffer+=sprintf(buffer, " ");
	1644	buffer+=sprintf(buffer, " DRDX: %s, DGDX: %s, DBDX: %s, DADX: %s\n", drdx, dgdx, dbdx, dadx);
	1645	buffer+=sprintf(buffer, " ");
	1646	buffer+=sprintf(buffer, " DRDE: %s, DGDE: %s, DBDE: %s, DADE: %s\n", drde, dgde, dbde, dade);
	1647	buffer+=sprintf(buffer, " ");
	1648	buffer+=sprintf(buffer, " DRDY: %s, DGDY: %s, DBDY: %s, DADY: %s\n", drdy, dgdy, dbdy, dady);
	1649
	1650	buffer+=sprintf(buffer, " ");
	1651	buffer+=sprintf(buffer, " S: %s, T: %s, W: %s\n", s, t, w);
	1652	buffer+=sprintf(buffer, " ");
	1653	buffer+=sprintf(buffer, " DSDX: %s, DTDX: %s, DWDX: %s\n", dsdx, dtdx, dwdx);
	1654	buffer+=sprintf(buffer, " ");
	1655	buffer+=sprintf(buffer, " DSDE: %s, DTDE: %s, DWDE: %s\n", dsde, dtde, dwde);
	1656	buffer+=sprintf(buffer, " ");
	1657	buffer+=sprintf(buffer, " DSDY: %s, DTDY: %s, DWDY: %s\n", dsdy, dtdy, dwdy);
	1658	break;
	1659	}
	1660	case 0x24:
	1661	case 0x25:
	1662	{
	1663	if (length < 16)
	1664	{
	1665	sprintf(buffer, "ERROR: Texture_Rectangle length = %d\n", length);
	1666	return;
	1667	}
	1668	cmd[2] = m_cmd_data[m_cmd_cur+2];
	1669	cmd[3] = m_cmd_data[m_cmd_cur+3];
	1670	sprintf(s, "%4.4f", (float)(INT16)((cmd[2] >> 16) & 0xffff) / 32.0f);
	1671	sprintf(t, "%4.4f", (float)(INT16)((cmd[2] >> 0) & 0xffff) / 32.0f);
	1672	sprintf(dsdx, "%4.4f", (float)(INT16)((cmd[3] >> 16) & 0xffff) / 1024.0f);
	1673	sprintf(dtdy, "%4.4f", (float)(INT16)((cmd[3] >> 16) & 0xffff) / 1024.0f);
	1674
	1675	if (command == 0x24)
	1676	sprintf(buffer, "Texture_Rectangle %d, %s, %s, %s, %s, %s, %s, %s, %s", tile, sh, th, sl, tl, s, t, dsdx, dtdy);
	1677	else
	1678	sprintf(buffer, "Texture_Rectangle_Flip %d, %s, %s, %s, %s, %s, %s, %s, %s", tile, sh, th, sl, tl, s, t, dsdx, dtdy);
	1679
	1680	break;
	1681	}
	1682	case 0x26: sprintf(buffer, "Sync_Load"); break;
	1683	case 0x27: sprintf(buffer, "Sync_Pipe"); break;
	1684	case 0x28: sprintf(buffer, "Sync_Tile"); break;
	1685	case 0x29: sprintf(buffer, "Sync_Full"); break;
	1686	case 0x2d: sprintf(buffer, "Set_Scissor %s, %s, %s, %s", sl, tl, sh, th); break;
	1687	case 0x2e: sprintf(buffer, "Set_Prim_Depth %04X, %04X", (cmd[1] >> 16) & 0xffff, cmd[1] & 0xffff); break;
	1688	case 0x2f: sprintf(buffer, "Set_Other_Modes %08X %08X", cmd[0], cmd[1]); break;
	1689	case 0x30: sprintf(buffer, "Load_TLUT %d, %s, %s, %s, %s", tile, sl, tl, sh, th); break;
	1690	case 0x32: sprintf(buffer, "Set_Tile_Size %d, %s, %s, %s, %s", tile, sl, tl, sh, th); break;
	1691	case 0x33: sprintf(buffer, "Load_Block %d, %03X, %03X, %03X, %03X", tile, (cmd[0] >> 12) & 0xfff, cmd[0] & 0xfff, (cmd[1] >> 12) & 0xfff, cmd[1] & 0xfff); break;
	1692	case 0x34: sprintf(buffer, "Load_Tile %d, %s, %s, %s, %s", tile, sl, tl, sh, th); break;
	1693	case 0x35: sprintf(buffer, "Set_Tile %d, %s, %s, %d, %04X", tile, format, size, ((cmd[0] >> 9) & 0x1ff) * 8, (cmd[0] & 0x1ff) * 8); break;
	1694	case 0x36: sprintf(buffer, "Fill_Rectangle %s, %s, %s, %s", sh, th, sl, tl); break;
	1695	case 0x37: sprintf(buffer, "Set_Fill_Color R: %d, G: %d, B: %d, A: %d", r, g, b, a); break;
	1696	case 0x38: sprintf(buffer, "Set_Fog_Color R: %d, G: %d, B: %d, A: %d", r, g, b, a); break;
	1697	case 0x39: sprintf(buffer, "Set_Blend_Color R: %d, G: %d, B: %d, A: %d", r, g, b, a); break;
	1698	case 0x3a: sprintf(buffer, "Set_Prim_Color %d, %d, R: %d, G: %d, B: %d, A: %d", (cmd[0] >> 8) & 0x1f, cmd[0] & 0xff, r, g, b, a); break;
	1699	case 0x3b: sprintf(buffer, "Set_Env_Color R: %d, G: %d, B: %d, A: %d", r, g, b, a); break;
	1700	case 0x3c: sprintf(buffer, "Set_Combine %08X %08X", cmd[0], cmd[1]); break;
	1701	case 0x3d: sprintf(buffer, "Set_Texture_Image %s, %s, %d, %08X", format, size, (cmd[0] & 0x1ff)+1, cmd[1]); break;
	1702	case 0x3e: sprintf(buffer, "Set_Mask_Image %08X", cmd[1]); break;
	1703	case 0x3f: sprintf(buffer, "Set_Color_Image %s, %s, %d, %08X", format, size, (cmd[0] & 0x1ff)+1, cmd[1]); break;
	1704	default: sprintf(buffer, "??? (%08X %08X)", cmd[0], cmd[1]); break;
	1705	}
	1706	}
	1707
	1708	/*****************************************************************************/
	1709
	1710	static UINT32 rightcvghex(UINT32 x, UINT32 fmask)
	1711	{
	1712	UINT32 stickybit = ((x >> 1) & 0x1fff) > 0;
	1713	UINT32 covered = ((x >> 14) & 3) + stickybit;
	1714	covered = (0xf0 >> covered) & 0xf;
	1715	return (covered & fmask);
	1716	}
	1717
	1718	static UINT32 leftcvghex(UINT32 x, UINT32 fmask)
	1719	{
	1720	UINT32 stickybit = ((x >> 1) & 0x1fff) > 0;
	1721	UINT32 covered = ((x >> 14) & 3) + stickybit;
	1722	covered = 0xf >> covered;
	1723	return (covered & fmask);
	1724	}
	1725
	1726	static INT32 CLIP(INT32 value,INT32 min,INT32 max)
	1727	{
	1728	if (value < min)
	1729	{
	1730	return min;
	1731	}
	1732	else if (value > max)
	1733	{
	1734	return max;
	1735	}
	1736	else
	1737	{
	1738	return value;
	1739	}
	1740	}
	1741
	1742	void n64_rdp::compute_cvg_noflip(extent_t Spans, INT32 majorx, INT32* minorx, INT32* majorxint, INT32* minorxint, INT32 scanline, INT32 yh, INT32 yl, INT32 base)
	1743	{
	1744	INT32 purgestart = 0xfff;
	1745	INT32 purgeend = 0;
	1746	bool writablescanline = !(scanline & ~0x3ff);
	1747	INT32 scanlinespx = scanline << 2;
	1748
	1749	if (!writablescanline) return;
	1750
	1751	for(int i = 0; i < 4; i++)
	1752	{
	1753	if (minorxint[i] < purgestart)
	1754	{
	1755	purgestart = minorxint[i];
	1756	}
	1757	if (majorxint[i] > purgeend)
	1758	{
	1759	purgeend = majorxint[i];
	1760	}
	1761	}
	1762
	1763	purgestart = CLIP(purgestart, 0, 1023);
	1764	purgeend = CLIP(purgeend, 0, 1023);
	1765	INT32 length = purgeend - purgestart;
	1766
	1767	if (length < 0) return;
	1768
	1769	rdp_span_aux userdata = (rdp_span_aux)Spans[scanline - base].userdata;
	1770	memset(&userdata->m_cvg[purgestart], 0, (length + 1) << 1);
	1771
	1772	for(int i = 0; i < 4; i++)
	1773	{
	1774	INT32 minorcur = minorx[i];
	1775	INT32 majorcur = majorx[i];
	1776	INT32 minorcurint = minorxint[i];
	1777	INT32 majorcurint = majorxint[i];
	1778	length = majorcurint - minorcurint;
	1779
	1780	INT32 fmask = (i & 1) ? 5 : 0xa;
	1781	INT32 maskshift = (i ^ 3) << 2;
	1782	INT32 fmaskshifted = fmask << maskshift;
	1783	INT32 fleft = CLIP(minorcurint + 1, 0, 647);
	1784	INT32 fright = CLIP(majorcurint - 1, 0, 647);
	1785	bool valid_y = ((scanlinespx + i) >= yh && (scanlinespx + i) < yl);
	1786	if (valid_y && length >= 0)
	1787	{
	1788	if (minorcurint != majorcurint)
	1789	{
	1790	if (!(minorcurint & ~0x3ff))
	1791	{
	1792	userdata->m_cvg[minorcurint] \|= (leftcvghex(minorcur, fmask) << maskshift);
	1793	}
	1794	if (!(majorcurint & ~0x3ff))
	1795	{
	1796	userdata->m_cvg[majorcurint] \|= (rightcvghex(majorcur, fmask) << maskshift);
	1797	}
	1798	}
	1799	else
	1800	{
	1801	if (!(majorcurint & ~0x3ff))
	1802	{
	1803	INT32 samecvg = leftcvghex(minorcur, fmask) & rightcvghex(majorcur, fmask);
	1804	userdata->m_cvg[majorcurint] \|= (samecvg << maskshift);
	1805	}
	1806	}
	1807	for (; fleft <= fright; fleft++)
	1808	{
	1809	userdata->m_cvg[fleft] \|= fmaskshifted;
	1810	}
	1811	}
	1812	}
	1813	}
	1814
	1815	void n64_rdp::compute_cvg_flip(extent_t Spans, INT32 majorx, INT32* minorx, INT32* majorxint, INT32* minorxint, INT32 scanline, INT32 yh, INT32 yl, INT32 base)
	1816	{
	1817	INT32 purgestart = 0xfff;
	1818	INT32 purgeend = 0;
	1819	bool writablescanline = !(scanline & ~0x3ff);
	1820	INT32 scanlinespx = scanline << 2;
	1821
	1822	if(!writablescanline) return;
	1823
	1824	for(int i = 0; i < 4; i++)
	1825	{
	1826	if (majorxint[i] < purgestart)
	1827	{
	1828	purgestart = majorxint[i];
	1829	}
	1830	if (minorxint[i] > purgeend)
	1831	{
	1832	purgeend = minorxint[i];
	1833	}
	1834	}
	1835
	1836	purgestart = CLIP(purgestart, 0, 1023);
	1837	purgeend = CLIP(purgeend, 0, 1023);
	1838
	1839	int length = purgeend - purgestart;
	1840
	1841	if (length < 0) return;
	1842
	1843	rdp_span_aux userdata = (rdp_span_aux)Spans[scanline - base].userdata;
	1844	memset(&userdata->m_cvg[purgestart], 0, (length + 1) << 1);
	1845
	1846	for(int i = 0; i < 4; i++)
	1847	{
	1848	INT32 minorcur = minorx[i];
	1849	INT32 majorcur = majorx[i];
	1850	INT32 minorcurint = minorxint[i];
	1851	INT32 majorcurint = majorxint[i];
	1852	length = minorcurint - majorcurint;
	1853
	1854	INT32 fmask = (i & 1) ? 5 : 0xa;
	1855	INT32 maskshift = (i ^ 3) << 2;
	1856	INT32 fmaskshifted = fmask << maskshift;
	1857	INT32 fleft = CLIP(majorcurint + 1, 0, 647);
	1858	INT32 fright = CLIP(minorcurint - 1, 0, 647);
	1859	bool valid_y = ((scanlinespx + i) >= yh && (scanlinespx + i) < yl);
	1860	if (valid_y && length >= 0)
	1861	{
	1862	if (minorcurint != majorcurint)
	1863	{
	1864	if (!(minorcurint & ~0x3ff))
	1865	{
	1866	userdata->m_cvg[minorcurint] \|= (rightcvghex(minorcur, fmask) << maskshift);
	1867	}
	1868	if (!(majorcurint & ~0x3ff))
	1869	{
	1870	userdata->m_cvg[majorcurint] \|= (leftcvghex(majorcur, fmask) << maskshift);
	1871	}
	1872	}
	1873	else
	1874	{
	1875	if (!(majorcurint & ~0x3ff))
	1876	{
	1877	INT32 samecvg = rightcvghex(minorcur, fmask) & leftcvghex(majorcur, fmask);
	1878	userdata->m_cvg[majorcurint] \|= (samecvg << maskshift);
	1879	}
	1880	}
	1881	for (; fleft <= fright; fleft++)
	1882	{
	1883	userdata->m_cvg[fleft] \|= fmaskshifted;
	1884	}
	1885	}
	1886	}
	1887	}
	1888
	1889	void n64_rdp::DrawTriangle(bool shade, bool texture, bool zbuffer, bool rect)
	1890	{
	1891	UINT32 *cmd_data = rect ? m_temp_rect_data : m_cmd_data;
	1892	UINT32 fifo_index = rect ? 0 : m_cmd_cur;
	1893	UINT32 w1 = cmd_data[fifo_index + 0];
	1894	UINT32 w2 = cmd_data[fifo_index + 1];
	1895
	1896	int flip = (w1 & 0x00800000) ? 1 : 0;
	1897	MiscState.MaxLevel = ((w1 >> 19) & 7);
	1898	int tilenum = (w1 >> 16) & 0x7;
	1899
	1900	int dsdiff = 0, dtdiff = 0, dwdiff = 0, drdiff = 0, dgdiff = 0, dbdiff = 0, dadiff = 0, dzdiff = 0;
	1901	int dsdeh = 0, dtdeh = 0, dwdeh = 0, drdeh = 0, dgdeh = 0, dbdeh = 0, dadeh = 0, dzdeh = 0;
	1902	int dsdxh = 0, dtdxh = 0, dwdxh = 0, drdxh = 0, dgdxh = 0, dbdxh = 0, dadxh = 0, dzdxh = 0;
	1903	int dsdyh = 0, dtdyh = 0, dwdyh = 0, drdyh = 0, dgdyh = 0, dbdyh = 0, dadyh = 0, dzdyh = 0;
	1904
	1905	INT32 maxxmx = 0; // maxxmx / minxhx very opaque names, consider re-naming
	1906	INT32 minxmx = 0;
	1907	INT32 maxxhx = 0;
	1908	INT32 minxhx = 0;
	1909
	1910	int shade_base = fifo_index + 8;
	1911	int texture_base = fifo_index + 8;
	1912	int zbuffer_base = fifo_index + 8;
	1913	if(shade)
	1914	{
	1915	texture_base += 16;
	1916	zbuffer_base += 16;
	1917	}
	1918	if(texture)
	1919	{
	1920	zbuffer_base += 16;
	1921	}
	1922
	1923	UINT32 w3 = cmd_data[fifo_index + 2];
	1924	UINT32 w4 = cmd_data[fifo_index + 3];
	1925	UINT32 w5 = cmd_data[fifo_index + 4];
	1926	UINT32 w6 = cmd_data[fifo_index + 5];
	1927	UINT32 w7 = cmd_data[fifo_index + 6];
	1928	UINT32 w8 = cmd_data[fifo_index + 7];
	1929
	1930	INT32 yl = (w1 & 0x3fff);
	1931	INT32 ym = ((w2 >> 16) & 0x3fff);
	1932	INT32 yh = ((w2 >> 0) & 0x3fff);
	1933	INT32 xl = (INT32)(w3 & 0x3fffffff);
	1934	INT32 xh = (INT32)(w5 & 0x3fffffff);
	1935	INT32 xm = (INT32)(w7 & 0x3fffffff);
	1936	// Inverse slopes in 16.16 format
	1937	INT32 dxldy = (INT32)(w4);
	1938	INT32 dxhdy = (INT32)(w6);
	1939	INT32 dxmdy = (INT32)(w8);
	1940
	1941	if (yl & 0x2000) yl \|= 0xffffc000;
	1942	if (ym & 0x2000) ym \|= 0xffffc000;
	1943	if (yh & 0x2000) yh \|= 0xffffc000;
	1944
	1945	if (xl & 0x20000000) xl \|= 0xc0000000;
	1946	if (xm & 0x20000000) xm \|= 0xc0000000;
	1947	if (xh & 0x20000000) xh \|= 0xc0000000;
	1948
	1949	int r = (cmd_data[shade_base+0 ] & 0xffff0000) \| ((cmd_data[shade_base+4 ] >> 16) & 0x0000ffff);
	1950	int g = ((cmd_data[shade_base+0 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+4 ] & 0x0000ffff);
	1951	int b = (cmd_data[shade_base+1 ] & 0xffff0000) \| ((cmd_data[shade_base+5 ] >> 16) & 0x0000ffff);
	1952	int a = ((cmd_data[shade_base+1 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+5 ] & 0x0000ffff);
	1953	int drdx = (cmd_data[shade_base+2 ] & 0xffff0000) \| ((cmd_data[shade_base+6 ] >> 16) & 0x0000ffff);
	1954	int dgdx = ((cmd_data[shade_base+2 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+6 ] & 0x0000ffff);
	1955	int dbdx = (cmd_data[shade_base+3 ] & 0xffff0000) \| ((cmd_data[shade_base+7 ] >> 16) & 0x0000ffff);
	1956	int dadx = ((cmd_data[shade_base+3 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+7 ] & 0x0000ffff);
	1957	int drde = (cmd_data[shade_base+8 ] & 0xffff0000) \| ((cmd_data[shade_base+12] >> 16) & 0x0000ffff);
	1958	int dgde = ((cmd_data[shade_base+8 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+12] & 0x0000ffff);
	1959	int dbde = (cmd_data[shade_base+9 ] & 0xffff0000) \| ((cmd_data[shade_base+13] >> 16) & 0x0000ffff);
	1960	int dade = ((cmd_data[shade_base+9 ] << 16) & 0xffff0000) \| (cmd_data[shade_base+13] & 0x0000ffff);
	1961	int drdy = (cmd_data[shade_base+10] & 0xffff0000) \| ((cmd_data[shade_base+14] >> 16) & 0x0000ffff);
	1962	int dgdy = ((cmd_data[shade_base+10] << 16) & 0xffff0000) \| (cmd_data[shade_base+14] & 0x0000ffff);
	1963	int dbdy = (cmd_data[shade_base+11] & 0xffff0000) \| ((cmd_data[shade_base+15] >> 16) & 0x0000ffff);
	1964	int dady = ((cmd_data[shade_base+11] << 16) & 0xffff0000) \| (cmd_data[shade_base+15] & 0x0000ffff);
	1965	int s = (cmd_data[texture_base+0 ] & 0xffff0000) \| ((cmd_data[texture_base+4 ] >> 16) & 0x0000ffff);
	1966	int t = ((cmd_data[texture_base+0 ] << 16) & 0xffff0000) \| (cmd_data[texture_base+4 ] & 0x0000ffff);
	1967	int w = (cmd_data[texture_base+1 ] & 0xffff0000) \| ((cmd_data[texture_base+5 ] >> 16) & 0x0000ffff);
	1968	int dsdx = (cmd_data[texture_base+2 ] & 0xffff0000) \| ((cmd_data[texture_base+6 ] >> 16) & 0x0000ffff);
	1969	int dtdx = ((cmd_data[texture_base+2 ] << 16) & 0xffff0000) \| (cmd_data[texture_base+6 ] & 0x0000ffff);
	1970	int dwdx = (cmd_data[texture_base+3 ] & 0xffff0000) \| ((cmd_data[texture_base+7 ] >> 16) & 0x0000ffff);
	1971	int dsde = (cmd_data[texture_base+8 ] & 0xffff0000) \| ((cmd_data[texture_base+12] >> 16) & 0x0000ffff);
	1972	int dtde = ((cmd_data[texture_base+8 ] << 16) & 0xffff0000) \| (cmd_data[texture_base+12] & 0x0000ffff);
	1973	int dwde = (cmd_data[texture_base+9 ] & 0xffff0000) \| ((cmd_data[texture_base+13] >> 16) & 0x0000ffff);
	1974	int dsdy = (cmd_data[texture_base+10] & 0xffff0000) \| ((cmd_data[texture_base+14] >> 16) & 0x0000ffff);
	1975	int dtdy = ((cmd_data[texture_base+10] << 16) & 0xffff0000) \| (cmd_data[texture_base+14] & 0x0000ffff);
	1976	int dwdy = (cmd_data[texture_base+11] & 0xffff0000) \| ((cmd_data[texture_base+15] >> 16) & 0x0000ffff);
	1977	int z = cmd_data[zbuffer_base+0];
	1978	int dzdx = cmd_data[zbuffer_base+1];
	1979	int dzde = cmd_data[zbuffer_base+2];
	1980	int dzdy = cmd_data[zbuffer_base+3];
	1981
	1982	int dzdy_dz = (dzdy >> 16) & 0xffff;
	1983	int dzdx_dz = (dzdx >> 16) & 0xffff;
	1984
	1985	extent_t Spans[2048];
	1986
	1987	SpanBase.m_span_drdy = drdy;
	1988	SpanBase.m_span_dgdy = dgdy;
	1989	SpanBase.m_span_dbdy = dbdy;
	1990	SpanBase.m_span_dady = dady;
	1991	SpanBase.m_span_dzdy = OtherModes.z_source_sel ? 0 : dzdy;
	1992
	1993	UINT32 temp_dzpix = ((dzdy_dz & 0x8000) ? ((~dzdy_dz) & 0x7fff) : dzdy_dz) + ((dzdx_dz & 0x8000) ? ((~dzdx_dz) & 0x7fff) : dzdx_dz);
	1994	SpanBase.m_span_dr = drdx & ~0x1f;
	1995	SpanBase.m_span_dg = dgdx & ~0x1f;
	1996	SpanBase.m_span_db = dbdx & ~0x1f;
	1997	SpanBase.m_span_da = dadx & ~0x1f;
	1998	SpanBase.m_span_ds = dsdx;
	1999	SpanBase.m_span_dt = dtdx;
	2000	SpanBase.m_span_dw = dwdx;
	2001	SpanBase.m_span_dz = OtherModes.z_source_sel ? 0 : dzdx;
	2002	SpanBase.m_span_dymax = 0;
	2003	SpanBase.m_span_dzpix = NormalizeDZPix(temp_dzpix & 0xffff) & 0xffff;
	2004
	2005	int xleft_inc = (dxmdy >> 2) & ~1;
	2006	int xright_inc = (dxhdy >> 2) & ~1;
	2007
	2008	int xright = xh & ~1;
	2009	int xleft = xm & ~1;
	2010
	2011	int sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0;
	2012	int do_offset = !(sign_dxhdy ^ (flip));
	2013
	2014	if (do_offset)
	2015	{
	2016	dsdeh = dsde >> 9; dsdyh = dsdy >> 9;
	2017	dtdeh = dtde >> 9; dtdyh = dtdy >> 9;
	2018	dwdeh = dwde >> 9; dwdyh = dwdy >> 9;
	2019	drdeh = drde >> 9; drdyh = drdy >> 9;
	2020	dgdeh = dgde >> 9; dgdyh = dgdy >> 9;
	2021	dbdeh = dbde >> 9; dbdyh = dbdy >> 9;
	2022	dadeh = dade >> 9; dadyh = dady >> 9;
	2023	dzdeh = dzde >> 9; dzdyh = dzdy >> 9;
	2024
	2025	dsdiff = (dsdeh << 8) + (dsdeh << 7) - (dsdyh << 8) - (dsdyh << 7);
	2026	dtdiff = (dtdeh << 8) + (dtdeh << 7) - (dtdyh << 8) - (dtdyh << 7);
	2027	dwdiff = (dwdeh << 8) + (dwdeh << 7) - (dwdyh << 8) - (dwdyh << 7);
	2028	drdiff = (drdeh << 8) + (drdeh << 7) - (drdyh << 8) - (drdyh << 7);
	2029	dgdiff = (dgdeh << 8) + (dgdeh << 7) - (dgdyh << 8) - (dgdyh << 7);
	2030	dbdiff = (dbdeh << 8) + (dbdeh << 7) - (dbdyh << 8) - (dbdyh << 7);
	2031	dadiff = (dadeh << 8) + (dadeh << 7) - (dadyh << 8) - (dadyh << 7);
	2032	dzdiff = (dzdeh << 8) + (dzdeh << 7) - (dzdyh << 8) - (dzdyh << 7);
	2033	}
	2034	else
	2035	{
	2036	dsdiff = dtdiff = dwdiff = drdiff = dgdiff = dbdiff = dadiff = dzdiff = 0;
	2037	}
	2038
	2039	dsdxh = dsdx >> 8;
	2040	dtdxh = dtdx >> 8;
	2041	dwdxh = dwdx >> 8;
	2042	drdxh = drdx >> 8;
	2043	dgdxh = dgdx >> 8;
	2044	dbdxh = dbdx >> 8;
	2045	dadxh = dadx >> 8;
	2046	dzdxh = dzdx >> 8;
	2047
	2048	INT32 ycur = yh & ~3;
	2049	INT32 ylfar = yl \| 3;
	2050	INT32 ldflag = (sign_dxhdy ^ flip) ? 0 : 3;
	2051	INT32 majorx[4];
	2052	INT32 minorx[4];
	2053	INT32 majorxint[4];
	2054	INT32 minorxint[4];
	2055	bool valid_y = true;
	2056
	2057	int xfrac = ((xright >> 8) & 0xff);
	2058
	2059	int clipy1 = Scissor.m_yh;
	2060	int clipy2 = Scissor.m_yl;
	2061
	2062	// Trivial reject
	2063	if((ycur >> 2) >= clipy2 && (ylfar >> 2) >= clipy2)
	2064	{
	2065	return;
	2066	}
	2067	if((ycur >> 2) < clipy1 && (ylfar >> 2) < clipy1)
	2068	{
	2069	return;
	2070	}
	2071
	2072	bool new_object = true;
	2073	rdp_poly_state *object = NULL;
	2074	bool valid = false;
	2075
	2076	if(flip)
	2077	{
	2078	for (int k = ycur; k <= ylfar; k++)
	2079	{
	2080	if (k == ym)
	2081	{
	2082	xleft = xl & ~1;
	2083	xleft_inc = (dxldy >> 2) & ~1;
	2084	}
	2085
	2086	int xstart = xleft >> 16; // 319
	2087	int xend = xright >> 16; // 0
	2088	int j = k >> 2;
	2089	int spanidx = (k - ycur) >> 2;
	2090	int spix = k & 3;
	2091	valid_y = !(k < yh \|\| k >= yl);
	2092
	2093	if (spanidx >= 0 && spanidx < 2048)
	2094	{
	2095	majorxint[spix] = xend; // 0
	2096	minorxint[spix] = xstart; // 319
	2097	majorx[spix] = xright; // 0x00000000
	2098	minorx[spix] = xleft; // 0x013f0000
	2099
	2100	if (spix == 0)
	2101	{
	2102	maxxmx = 0;
	2103	minxhx = 0xfff;
	2104	}
	2105
	2106	if (valid_y)
	2107	{
	2108	maxxmx = (xstart > maxxmx) ? xstart : maxxmx;
	2109	minxhx = (xend < minxhx) ? xend : minxhx;
	2110	}
	2111
	2112	if (spix == 0)
	2113	{
	2114	if(new_object)
	2115	{
	2116	object = &object_data_alloc();
	2117	memcpy(object->m_tmem, m_tmem, 0x1000);
	2118	new_object = false;
	2119	}
	2120
	2121	Spans[spanidx].userdata = (void)((UINT8)AuxBuf + AuxBufPtr);
	2122	AuxBufPtr += sizeof(rdp_span_aux);
	2123
	2124	if(AuxBufPtr >= EXTENT_AUX_COUNT)
	2125	{
	2126	fatalerror("n64_rdp::DrawTriangle: span aux buffer overflow\n");
	2127	}
	2128
	2129	rdp_span_aux userdata = (rdp_span_aux)Spans[spanidx].userdata;
	2130	valid = true;
	2131
	2132	userdata->m_tmem = object->m_tmem;
	2133
	2134	userdata->BlendColor = BlendColor;
	2135	userdata->PrimColor = PrimColor;
	2136	userdata->EnvColor = EnvColor;
	2137	userdata->FogColor = FogColor;
	2138	userdata->KeyScale = KeyScale;
	2139	userdata->LODFraction = LODFraction;
	2140	userdata->PrimLODFraction = PrimLODFraction;
	2141
	2142	userdata->ColorInputs.combiner_rgbsub_a_r[0] = userdata->ColorInputs.combiner_rgbsub_a_r[1] = &OneColor.i.r;
	2143	userdata->ColorInputs.combiner_rgbsub_a_g[0] = userdata->ColorInputs.combiner_rgbsub_a_g[1] = &OneColor.i.g;
	2144	userdata->ColorInputs.combiner_rgbsub_a_b[0] = userdata->ColorInputs.combiner_rgbsub_a_b[1] = &OneColor.i.b;
	2145	userdata->ColorInputs.combiner_rgbsub_b_r[0] = userdata->ColorInputs.combiner_rgbsub_b_r[1] = &OneColor.i.r;
	2146	userdata->ColorInputs.combiner_rgbsub_b_g[0] = userdata->ColorInputs.combiner_rgbsub_b_g[1] = &OneColor.i.g;
	2147	userdata->ColorInputs.combiner_rgbsub_b_b[0] = userdata->ColorInputs.combiner_rgbsub_b_b[1] = &OneColor.i.b;
	2148	userdata->ColorInputs.combiner_rgbmul_r[0] = userdata->ColorInputs.combiner_rgbmul_r[1] = &OneColor.i.r;
	2149	userdata->ColorInputs.combiner_rgbmul_g[0] = userdata->ColorInputs.combiner_rgbmul_g[1] = &OneColor.i.g;
	2150	userdata->ColorInputs.combiner_rgbmul_b[0] = userdata->ColorInputs.combiner_rgbmul_b[1] = &OneColor.i.b;
	2151	userdata->ColorInputs.combiner_rgbadd_r[0] = userdata->ColorInputs.combiner_rgbadd_r[1] = &OneColor.i.r;
	2152	userdata->ColorInputs.combiner_rgbadd_g[0] = userdata->ColorInputs.combiner_rgbadd_g[1] = &OneColor.i.g;
	2153	userdata->ColorInputs.combiner_rgbadd_b[0] = userdata->ColorInputs.combiner_rgbadd_b[1] = &OneColor.i.b;
	2154	userdata->ColorInputs.combiner_alphasub_a[0] = userdata->ColorInputs.combiner_alphasub_a[1] = &OneColor.i.a;
	2155	userdata->ColorInputs.combiner_alphasub_b[0] = userdata->ColorInputs.combiner_alphasub_b[1] = &OneColor.i.a;
	2156	userdata->ColorInputs.combiner_alphamul[0] = userdata->ColorInputs.combiner_alphamul[1] = &OneColor.i.a;
	2157	userdata->ColorInputs.combiner_alphaadd[0] = userdata->ColorInputs.combiner_alphaadd[1] = &OneColor.i.a;
	2158
	2159	userdata->ColorInputs.blender1a_r[0] = userdata->ColorInputs.blender1a_r[1] = &userdata->PixelColor.i.r;
	2160	userdata->ColorInputs.blender1a_g[0] = userdata->ColorInputs.blender1a_g[1] = &userdata->PixelColor.i.g;
	2161	userdata->ColorInputs.blender1a_b[0] = userdata->ColorInputs.blender1a_b[1] = &userdata->PixelColor.i.b;
	2162	userdata->ColorInputs.blender1b_a[0] = userdata->ColorInputs.blender1b_a[1] = &userdata->PixelColor.i.a;
	2163	userdata->ColorInputs.blender2a_r[0] = userdata->ColorInputs.blender2a_r[1] = &userdata->PixelColor.i.r;
	2164	userdata->ColorInputs.blender2a_g[0] = userdata->ColorInputs.blender2a_g[1] = &userdata->PixelColor.i.g;
	2165	userdata->ColorInputs.blender2a_b[0] = userdata->ColorInputs.blender2a_b[1] = &userdata->PixelColor.i.b;
	2166	userdata->ColorInputs.blender2b_a[0] = userdata->ColorInputs.blender2b_a[1] = &userdata->PixelColor.i.a;
	2167
	2168	// Setup blender data for this scanline
	2169	SetBlenderInput(0, 0, &userdata->ColorInputs.blender1a_r[0],
	2170	&userdata->ColorInputs.blender1a_g[0],
	2171	&userdata->ColorInputs.blender1a_b[0],
	2172	&userdata->ColorInputs.blender1b_a[0], OtherModes.blend_m1a_0, OtherModes.blend_m1b_0, userdata);
	2173	SetBlenderInput(0, 1, &userdata->ColorInputs.blender2a_r[0],
	2174	&userdata->ColorInputs.blender2a_g[0],
	2175	&userdata->ColorInputs.blender2a_b[0],
	2176	&userdata->ColorInputs.blender2b_a[0], OtherModes.blend_m2a_0, OtherModes.blend_m2b_0, userdata);
	2177	SetBlenderInput(1, 0, &userdata->ColorInputs.blender1a_r[1],
	2178	&userdata->ColorInputs.blender1a_g[1],
	2179	&userdata->ColorInputs.blender1a_b[1],
	2180	&userdata->ColorInputs.blender1b_a[1], OtherModes.blend_m1a_1, OtherModes.blend_m1b_1, userdata);
	2181	SetBlenderInput(1, 1, &userdata->ColorInputs.blender2a_r[1],
	2182	&userdata->ColorInputs.blender2a_g[1],
	2183	&userdata->ColorInputs.blender2a_b[1],
	2184	&userdata->ColorInputs.blender2b_a[1], OtherModes.blend_m2a_1, OtherModes.blend_m2b_1, userdata);
	2185
	2186	// Setup color combiner data for this scanline
	2187	SetSubAInputRGB(&userdata->ColorInputs.combiner_rgbsub_a_r[0], &userdata->ColorInputs.combiner_rgbsub_a_g[0], &userdata->ColorInputs.combiner_rgbsub_a_b[0], m_combine.sub_a_rgb0, userdata);
	2188	SetSubBInputRGB(&userdata->ColorInputs.combiner_rgbsub_b_r[0], &userdata->ColorInputs.combiner_rgbsub_b_g[0], &userdata->ColorInputs.combiner_rgbsub_b_b[0], m_combine.sub_b_rgb0, userdata);
	2189	SetMulInputRGB(&userdata->ColorInputs.combiner_rgbmul_r[0], &userdata->ColorInputs.combiner_rgbmul_g[0], &userdata->ColorInputs.combiner_rgbmul_b[0], m_combine.mul_rgb0, userdata);
	2190	SetAddInputRGB(&userdata->ColorInputs.combiner_rgbadd_r[0], &userdata->ColorInputs.combiner_rgbadd_g[0], &userdata->ColorInputs.combiner_rgbadd_b[0], m_combine.add_rgb0, userdata);
	2191	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_a[0], m_combine.sub_a_a0, userdata);
	2192	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_b[0], m_combine.sub_b_a0, userdata);
	2193	SetMulInputAlpha(&userdata->ColorInputs.combiner_alphamul[0], m_combine.mul_a0, userdata);
	2194	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphaadd[0], m_combine.add_a0, userdata);
	2195
	2196	SetSubAInputRGB(&userdata->ColorInputs.combiner_rgbsub_a_r[1], &userdata->ColorInputs.combiner_rgbsub_a_g[1], &userdata->ColorInputs.combiner_rgbsub_a_b[1], m_combine.sub_a_rgb1, userdata);
	2197	SetSubBInputRGB(&userdata->ColorInputs.combiner_rgbsub_b_r[1], &userdata->ColorInputs.combiner_rgbsub_b_g[1], &userdata->ColorInputs.combiner_rgbsub_b_b[1], m_combine.sub_b_rgb1, userdata);
	2198	SetMulInputRGB(&userdata->ColorInputs.combiner_rgbmul_r[1], &userdata->ColorInputs.combiner_rgbmul_g[1], &userdata->ColorInputs.combiner_rgbmul_b[1], m_combine.mul_rgb1, userdata);
	2199	SetAddInputRGB(&userdata->ColorInputs.combiner_rgbadd_r[1], &userdata->ColorInputs.combiner_rgbadd_g[1], &userdata->ColorInputs.combiner_rgbadd_b[1], m_combine.add_rgb1, userdata);
	2200	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_a[1], m_combine.sub_a_a1, userdata);
	2201	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_b[1], m_combine.sub_b_a1, userdata);
	2202	SetMulInputAlpha(&userdata->ColorInputs.combiner_alphamul[1], m_combine.mul_a1, userdata);
	2203	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphaadd[1], m_combine.add_a1, userdata);
	2204	}
	2205
	2206	if (spix == 3)
	2207	{
	2208	Spans[spanidx].startx = maxxmx;
	2209	Spans[spanidx].stopx = minxhx;
	2210	compute_cvg_flip(Spans, majorx, minorx, majorxint, minorxint, j, yh, yl, ycur >> 2);
	2211	}
	2212
	2213	if (spix == ldflag)
	2214	{
	2215	((rdp_span_aux*)Spans[spanidx].userdata)->m_unscissored_rx = xend;
	2216	xfrac = ((xright >> 8) & 0xff);
	2217	Spans[spanidx].param[SPAN_R].start = ((r >> 9) << 9) + drdiff - (xfrac * drdxh);
	2218	Spans[spanidx].param[SPAN_G].start = ((g >> 9) << 9) + dgdiff - (xfrac * dgdxh);
	2219	Spans[spanidx].param[SPAN_B].start = ((b >> 9) << 9) + dbdiff - (xfrac * dbdxh);
	2220	Spans[spanidx].param[SPAN_A].start = ((a >> 9) << 9) + dadiff - (xfrac * dadxh);
	2221	Spans[spanidx].param[SPAN_S].start = (((s >> 9) << 9) + dsdiff - (xfrac * dsdxh)) & ~0x1f;
	2222	Spans[spanidx].param[SPAN_T].start = (((t >> 9) << 9) + dtdiff - (xfrac * dtdxh)) & ~0x1f;
	2223	Spans[spanidx].param[SPAN_W].start = (((w >> 9) << 9) + dwdiff - (xfrac * dwdxh)) & ~0x1f;
	2224	Spans[spanidx].param[SPAN_Z].start = ((z >> 9) << 9) + dzdiff - (xfrac * dzdxh);
	2225	}
	2226	}
	2227
	2228	if (spix == 3)
	2229	{
	2230	r += drde;
	2231	g += dgde;
	2232	b += dbde;
	2233	a += dade;
	2234	s += dsde;
	2235	t += dtde;
	2236	w += dwde;
	2237	z += dzde;
	2238	}
	2239
	2240	xleft += xleft_inc;
	2241	xright += xright_inc;
	2242	}
	2243	}
	2244	else
	2245	{
	2246	for (int k = ycur; k <= ylfar; k++)
	2247	{
	2248	if (k == ym)
	2249	{
	2250	xleft = xl & ~1;
	2251	xleft_inc = (dxldy >> 2) & ~1;
	2252	}
	2253
	2254	int xstart = xleft >> 16;
	2255	int xend = xright >> 16;
	2256	int j = k >> 2;
	2257	int spanidx = j - (ycur >> 2);
	2258	int spix = k & 3;
	2259	valid_y = !(k < yh \|\| k >= yl);
	2260
	2261	if (k >= 0 && k < 0x1000)
	2262	{
	2263	majorxint[spix] = xend;
	2264	minorxint[spix] = xstart;
	2265	majorx[spix] = xright;
	2266	minorx[spix] = xleft;
	2267
	2268	if (spix == 0)
	2269	{
	2270	maxxhx = 0;
	2271	minxmx = 0xfff;
	2272	}
	2273
	2274	if (valid_y)
	2275	{
	2276	minxmx = (xstart < minxmx) ? xstart : minxmx;
	2277	maxxhx = (xend > maxxhx) ? xend : maxxhx;
	2278	}
	2279
	2280	if (spix == 0)
	2281	{
	2282	if(new_object)
	2283	{
	2284	object = &object_data_alloc();
	2285	memcpy(object->m_tmem, m_tmem, 0x1000);
	2286	new_object = false;
	2287	}
	2288
	2289	Spans[spanidx].userdata = (void)((UINT8)AuxBuf + AuxBufPtr);
	2290	valid = true;
	2291	AuxBufPtr += sizeof(rdp_span_aux);
	2292
	2293	if(AuxBufPtr >= EXTENT_AUX_COUNT)
	2294	{
	2295	fatalerror("n64_rdp::DrawTriangle: span aux buffer overflow\n");
	2296	}
	2297
	2298	rdp_span_aux userdata = (rdp_span_aux)Spans[spanidx].userdata;
	2299	userdata->m_tmem = object->m_tmem;
	2300
	2301	userdata->BlendColor = BlendColor;
	2302	userdata->PrimColor = PrimColor;
	2303	userdata->EnvColor = EnvColor;
	2304	userdata->FogColor = FogColor;
	2305	userdata->KeyScale = KeyScale;
	2306	userdata->LODFraction = LODFraction;
	2307	userdata->PrimLODFraction = PrimLODFraction;
	2308
	2309	userdata->ColorInputs.combiner_rgbsub_a_r[0] = userdata->ColorInputs.combiner_rgbsub_a_r[1] = &OneColor.i.r;
	2310	userdata->ColorInputs.combiner_rgbsub_a_g[0] = userdata->ColorInputs.combiner_rgbsub_a_g[1] = &OneColor.i.g;
	2311	userdata->ColorInputs.combiner_rgbsub_a_b[0] = userdata->ColorInputs.combiner_rgbsub_a_b[1] = &OneColor.i.b;
	2312	userdata->ColorInputs.combiner_rgbsub_b_r[0] = userdata->ColorInputs.combiner_rgbsub_b_r[1] = &OneColor.i.r;
	2313	userdata->ColorInputs.combiner_rgbsub_b_g[0] = userdata->ColorInputs.combiner_rgbsub_b_g[1] = &OneColor.i.g;
	2314	userdata->ColorInputs.combiner_rgbsub_b_b[0] = userdata->ColorInputs.combiner_rgbsub_b_b[1] = &OneColor.i.b;
	2315	userdata->ColorInputs.combiner_rgbmul_r[0] = userdata->ColorInputs.combiner_rgbmul_r[1] = &OneColor.i.r;
	2316	userdata->ColorInputs.combiner_rgbmul_g[0] = userdata->ColorInputs.combiner_rgbmul_g[1] = &OneColor.i.g;
	2317	userdata->ColorInputs.combiner_rgbmul_b[0] = userdata->ColorInputs.combiner_rgbmul_b[1] = &OneColor.i.b;
	2318	userdata->ColorInputs.combiner_rgbadd_r[0] = userdata->ColorInputs.combiner_rgbadd_r[1] = &OneColor.i.r;
	2319	userdata->ColorInputs.combiner_rgbadd_g[0] = userdata->ColorInputs.combiner_rgbadd_g[1] = &OneColor.i.g;
	2320	userdata->ColorInputs.combiner_rgbadd_b[0] = userdata->ColorInputs.combiner_rgbadd_b[1] = &OneColor.i.b;
	2321	userdata->ColorInputs.combiner_alphasub_a[0] = userdata->ColorInputs.combiner_alphasub_a[1] = &OneColor.i.a;
	2322	userdata->ColorInputs.combiner_alphasub_b[0] = userdata->ColorInputs.combiner_alphasub_b[1] = &OneColor.i.a;
	2323	userdata->ColorInputs.combiner_alphamul[0] = userdata->ColorInputs.combiner_alphamul[1] = &OneColor.i.a;
	2324	userdata->ColorInputs.combiner_alphaadd[0] = userdata->ColorInputs.combiner_alphaadd[1] = &OneColor.i.a;
	2325
	2326	userdata->ColorInputs.blender1a_r[0] = userdata->ColorInputs.blender1a_r[1] = &userdata->PixelColor.i.r;
	2327	userdata->ColorInputs.blender1a_g[0] = userdata->ColorInputs.blender1a_g[1] = &userdata->PixelColor.i.g;
	2328	userdata->ColorInputs.blender1a_b[0] = userdata->ColorInputs.blender1a_b[1] = &userdata->PixelColor.i.b;
	2329	userdata->ColorInputs.blender1b_a[0] = userdata->ColorInputs.blender1b_a[1] = &userdata->PixelColor.i.a;
	2330	userdata->ColorInputs.blender2a_r[0] = userdata->ColorInputs.blender2a_r[1] = &userdata->PixelColor.i.r;
	2331	userdata->ColorInputs.blender2a_g[0] = userdata->ColorInputs.blender2a_g[1] = &userdata->PixelColor.i.g;
	2332	userdata->ColorInputs.blender2a_b[0] = userdata->ColorInputs.blender2a_b[1] = &userdata->PixelColor.i.b;
	2333	userdata->ColorInputs.blender2b_a[0] = userdata->ColorInputs.blender2b_a[1] = &userdata->PixelColor.i.a;
	2334
	2335	// Setup blender data for this scanline
	2336	SetBlenderInput(0, 0, &userdata->ColorInputs.blender1a_r[0],
	2337	&userdata->ColorInputs.blender1a_g[0],
	2338	&userdata->ColorInputs.blender1a_b[0],
	2339	&userdata->ColorInputs.blender1b_a[0], OtherModes.blend_m1a_0, OtherModes.blend_m1b_0, userdata);
	2340	SetBlenderInput(0, 1, &userdata->ColorInputs.blender2a_r[0],
	2341	&userdata->ColorInputs.blender2a_g[0],
	2342	&userdata->ColorInputs.blender2a_b[0],
	2343	&userdata->ColorInputs.blender2b_a[0], OtherModes.blend_m2a_0, OtherModes.blend_m2b_0, userdata);
	2344	SetBlenderInput(1, 0, &userdata->ColorInputs.blender1a_r[1],
	2345	&userdata->ColorInputs.blender1a_g[1],
	2346	&userdata->ColorInputs.blender1a_b[1],
	2347	&userdata->ColorInputs.blender1b_a[1], OtherModes.blend_m1a_1, OtherModes.blend_m1b_1, userdata);
	2348	SetBlenderInput(1, 1, &userdata->ColorInputs.blender2a_r[1],
	2349	&userdata->ColorInputs.blender2a_g[1],
	2350	&userdata->ColorInputs.blender2a_b[1],
	2351	&userdata->ColorInputs.blender2b_a[1], OtherModes.blend_m2a_1, OtherModes.blend_m2b_1, userdata);
	2352
	2353	// Setup color combiner data for this scanline
	2354	SetSubAInputRGB(&userdata->ColorInputs.combiner_rgbsub_a_r[0], &userdata->ColorInputs.combiner_rgbsub_a_g[0], &userdata->ColorInputs.combiner_rgbsub_a_b[0], m_combine.sub_a_rgb0, userdata);
	2355	SetSubBInputRGB(&userdata->ColorInputs.combiner_rgbsub_b_r[0], &userdata->ColorInputs.combiner_rgbsub_b_g[0], &userdata->ColorInputs.combiner_rgbsub_b_b[0], m_combine.sub_b_rgb0, userdata);
	2356	SetMulInputRGB(&userdata->ColorInputs.combiner_rgbmul_r[0], &userdata->ColorInputs.combiner_rgbmul_g[0], &userdata->ColorInputs.combiner_rgbmul_b[0], m_combine.mul_rgb0, userdata);
	2357	SetAddInputRGB(&userdata->ColorInputs.combiner_rgbadd_r[0], &userdata->ColorInputs.combiner_rgbadd_g[0], &userdata->ColorInputs.combiner_rgbadd_b[0], m_combine.add_rgb0, userdata);
	2358	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_a[0], m_combine.sub_a_a0, userdata);
	2359	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_b[0], m_combine.sub_b_a0, userdata);
	2360	SetMulInputAlpha(&userdata->ColorInputs.combiner_alphamul[0], m_combine.mul_a0, userdata);
	2361	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphaadd[0], m_combine.add_a0, userdata);
	2362
	2363	SetSubAInputRGB(&userdata->ColorInputs.combiner_rgbsub_a_r[1], &userdata->ColorInputs.combiner_rgbsub_a_g[1], &userdata->ColorInputs.combiner_rgbsub_a_b[1], m_combine.sub_a_rgb1, userdata);
	2364	SetSubBInputRGB(&userdata->ColorInputs.combiner_rgbsub_b_r[1], &userdata->ColorInputs.combiner_rgbsub_b_g[1], &userdata->ColorInputs.combiner_rgbsub_b_b[1], m_combine.sub_b_rgb1, userdata);
	2365	SetMulInputRGB(&userdata->ColorInputs.combiner_rgbmul_r[1], &userdata->ColorInputs.combiner_rgbmul_g[1], &userdata->ColorInputs.combiner_rgbmul_b[1], m_combine.mul_rgb1, userdata);
	2366	SetAddInputRGB(&userdata->ColorInputs.combiner_rgbadd_r[1], &userdata->ColorInputs.combiner_rgbadd_g[1], &userdata->ColorInputs.combiner_rgbadd_b[1], m_combine.add_rgb1, userdata);
	2367	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_a[1], m_combine.sub_a_a1, userdata);
	2368	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphasub_b[1], m_combine.sub_b_a1, userdata);
	2369	SetMulInputAlpha(&userdata->ColorInputs.combiner_alphamul[1], m_combine.mul_a1, userdata);
	2370	SetSubInputAlpha(&userdata->ColorInputs.combiner_alphaadd[1], m_combine.add_a1, userdata);
	2371	}
	2372
	2373	if (spix == 3)
	2374	{
	2375	Spans[spanidx].startx = minxmx;
	2376	Spans[spanidx].stopx = maxxhx;
	2377	compute_cvg_noflip(Spans, majorx, minorx, majorxint, minorxint, j, yh, yl, ycur >> 2);
	2378	}
	2379
	2380	if (spix == ldflag)
	2381	{
	2382	((rdp_span_aux*)Spans[spanidx].userdata)->m_unscissored_rx = xend;
	2383	xfrac = ((xright >> 8) & 0xff);
	2384	Spans[spanidx].param[SPAN_R].start = ((r >> 9) << 9) + drdiff - (xfrac * drdxh);
	2385	Spans[spanidx].param[SPAN_G].start = ((g >> 9) << 9) + dgdiff - (xfrac * dgdxh);
	2386	Spans[spanidx].param[SPAN_B].start = ((b >> 9) << 9) + dbdiff - (xfrac * dbdxh);
	2387	Spans[spanidx].param[SPAN_A].start = ((a >> 9) << 9) + dadiff - (xfrac * dadxh);
	2388	Spans[spanidx].param[SPAN_S].start = (((s >> 9) << 9) + dsdiff - (xfrac * dsdxh)) & ~0x1f;
	2389	Spans[spanidx].param[SPAN_T].start = (((t >> 9) << 9) + dtdiff - (xfrac * dtdxh)) & ~0x1f;
	2390	Spans[spanidx].param[SPAN_W].start = (((w >> 9) << 9) + dwdiff - (xfrac * dwdxh)) & ~0x1f;
	2391	Spans[spanidx].param[SPAN_Z].start = ((z >> 9) << 9) + dzdiff - (xfrac * dzdxh);
	2392	}
	2393	}
	2394
	2395	if (spix == 3)
	2396	{
	2397	r += drde;
	2398	g += dgde;
	2399	b += dbde;
	2400	a += dade;
	2401	s += dsde;
	2402	t += dtde;
	2403	w += dwde;
	2404	z += dzde;
	2405	}
	2406	xleft += xleft_inc;
	2407	xright += xright_inc;
	2408	}
	2409	}
	2410
	2411	if(!new_object && valid)
	2412	{
	2413	RenderSpans(yh >> 2, yl >> 2, tilenum, flip ? true : false, Spans, rect, object);
	2414	}
	2415
	2416	//wait("DrawTriangle");
	2417	}
	2418
	2419	/*****************************************************************************/
	2420
	2421	////////////////////////
	2422	// FB ACCESSORS
	2423	////////////////////////
	2424	void n64_rdp::_Write16Bit_Cvg0_Blend(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2425	{
	2426	UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
	2427	UINT32 hb = fb;
	2428
	2429	UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
	2430	UINT32 finalcvg = 0;
	2431
	2432	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2433	{
	2434	finalcolor = RREADIDX16(fb) & 0xfffe;
	2435	CHECK16(fb);
	2436	}
	2437
	2438	finalcvg = userdata->CurrentPixCvg + userdata->CurrentMemCvg;
	2439	if (finalcvg & 8)
	2440	{
	2441	finalcvg = 7;
	2442	}
	2443	RWRITEIDX16(fb, finalcolor \| ((finalcvg >> 2) & 1));
	2444	HWRITEADDR8(hb, finalcvg & 3);
	2445	}
	2446
	2447	void n64_rdp::_Write16Bit_Cvg0_NoBlend(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2448	{
	2449	UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
	2450	UINT32 hb = fb;
	2451
	2452	UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
	2453	UINT32 finalcvg = 0;
	2454
	2455	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2456	{
	2457	finalcolor = RREADIDX16(fb) & 0xfffe;
	2458	CHECK16(fb);
	2459	}
	2460
	2461	finalcvg = (userdata->CurrentPixCvg - 1) & 7;
	2462	RWRITEIDX16(fb, finalcolor \| ((finalcvg >> 2) & 1));
	2463	HWRITEADDR8(hb, finalcvg & 3);
	2464	}
	2465
	2466	void n64_rdp::_Write16Bit_Cvg1(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2467	{
	2468	UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
	2469	UINT32 hb = fb;
	2470
	2471	UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
	2472	UINT32 finalcvg = 0;
	2473
	2474	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2475	{
	2476	finalcolor = RREADIDX16(fb) & 0xfffe;
	2477	CHECK16(fb);
	2478	}
	2479
	2480	finalcvg = (userdata->CurrentPixCvg + userdata->CurrentMemCvg) & 7;
	2481	RWRITEIDX16(fb, finalcolor \| ((finalcvg >> 2) & 1));
	2482	HWRITEADDR8(hb, finalcvg & 3);
	2483	}
	2484
	2485	void n64_rdp::_Write16Bit_Cvg2(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2486	{
	2487	UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
	2488	UINT32 hb = fb;
	2489
	2490	UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
	2491
	2492	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2493	{
	2494	finalcolor = RREADIDX16(fb) & 0xfffe;
	2495	CHECK16(fb);
	2496	}
	2497
	2498	RWRITEIDX16(fb, finalcolor \| 1);
	2499	HWRITEADDR8(hb, 3);
	2500	}
	2501
	2502	void n64_rdp::_Write16Bit_Cvg3(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2503	{
	2504	UINT32 fb = (object.MiscState.FBAddress >> 1) + curpixel;
	2505	UINT32 hb = fb;
	2506
	2507	UINT16 finalcolor = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1);
	2508
	2509	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2510	{
	2511	finalcolor = RREADIDX16(fb) & 0xfffe;
	2512	CHECK16(fb);
	2513	}
	2514
	2515	RWRITEIDX16(fb, finalcolor \| ((userdata->CurrentMemCvg >> 2) & 1));
	2516	HWRITEADDR8(hb, userdata->CurrentMemCvg & 3);
	2517	}
	2518
	2519	void n64_rdp::_Write32Bit_Cvg0_Blend(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2520	{
	2521	UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
	2522	UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
	2523	UINT32 finalcvg = 0;
	2524
	2525	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2526	{
	2527	finalcolor = RREADIDX32(fb) & 0xffffff00;
	2528	CHECK32(fb);
	2529	}
	2530
	2531	finalcvg = userdata->CurrentPixCvg + userdata->CurrentMemCvg;
	2532	if (finalcvg & 8)
	2533	{
	2534	finalcvg = 7;
	2535	}
	2536	RWRITEIDX32(fb, finalcolor \| (finalcvg << 5));
	2537	}
	2538
	2539	void n64_rdp::_Write32Bit_Cvg0_NoBlend(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2540	{
	2541	UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
	2542	UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
	2543	UINT32 finalcvg = 0;
	2544
	2545	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2546	{
	2547	finalcolor = RREADIDX32(fb) & 0xffffff00;
	2548	CHECK32(fb);
	2549	}
	2550
	2551	finalcvg = (userdata->CurrentPixCvg - 1) & 7;
	2552	RWRITEIDX32(fb, finalcolor \| (finalcvg << 5));
	2553	}
	2554
	2555	void n64_rdp::_Write32Bit_Cvg1(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2556	{
	2557	UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
	2558	UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
	2559	UINT32 finalcvg = 0;
	2560
	2561	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2562	{
	2563	finalcolor = RREADIDX32(fb) & 0xffffff00;
	2564	CHECK32(fb);
	2565	}
	2566
	2567	finalcvg = (userdata->CurrentPixCvg + userdata->CurrentMemCvg) & 7;
	2568	finalcolor \|= (finalcvg << 5);
	2569	RWRITEIDX32(fb, finalcolor);
	2570	}
	2571
	2572	void n64_rdp::_Write32Bit_Cvg2(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2573	{
	2574	UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
	2575	UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
	2576
	2577	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2578	{
	2579	finalcolor = RREADIDX32(fb) & 0xffffff00;
	2580	CHECK32(fb);
	2581	}
	2582
	2583	RWRITEIDX32(fb, finalcolor \| 0xE0);
	2584	}
	2585
	2586	void n64_rdp::_Write32Bit_Cvg3(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, rdp_span_aux *userdata, const rdp_poly_state &object)
	2587	{
	2588	UINT32 fb = (object.MiscState.FBAddress >> 2) + curpixel;
	2589	UINT32 finalcolor = (r << 24) \| (g << 16) \| (b << 8);//cvg as 3 MSBs of alpha channel;
	2590
	2591	if (object.OtherModes.color_on_cvg && !userdata->PreWrap)
	2592	{
	2593	finalcolor = RREADIDX32(fb) & 0xffffff00;
	2594	CHECK32(fb);
	2595	}
	2596
	2597	RWRITEIDX32(fb, finalcolor \| (userdata->CurrentMemCvg << 5));
	2598	}
	2599
	2600
	2601	void n64_rdp::_Read16Bit_ImgRead0(UINT32 curpixel, rdp_span_aux *userdata, const rdp_poly_state &object)
	2602	{
	2603	UINT16 fword = RREADIDX16((object.MiscState.FBAddress >> 1) + curpixel);
	2604	CHECK16((object.MiscState.FBAddress >> 1) + curpixel);
	2605	userdata->MemoryColor.i.r = GETHICOL(fword);
	2606	userdata->MemoryColor.i.g = GETMEDCOL(fword);
	2607	userdata->MemoryColor.i.b = GETLOWCOL(fword);
	2608	userdata->MemoryColor.i.a = 0xff;
	2609	userdata->CurrentMemCvg = 7;
	2610	}
	2611
	2612	void n64_rdp::_Read16Bit_ImgRead1(UINT32 curpixel, rdp_span_aux *userdata, const rdp_poly_state &object)
	2613	{
	2614	UINT16 fword = RREADIDX16((object.MiscState.FBAddress >> 1) + curpixel);
	2615	CHECK16((object.MiscState.FBAddress >> 1) + curpixel);
	2616	UINT8 hbyte = HREADADDR8((object.MiscState.FBAddress >> 1) + curpixel);
	2617	userdata->MemoryColor.i.r = GETHICOL(fword);
	2618	userdata->MemoryColor.i.g = GETMEDCOL(fword);
	2619	userdata->MemoryColor.i.b = GETLOWCOL(fword);
	2620	userdata->MemoryColor.i.a = userdata->CurrentMemCvg << 5;
	2621	userdata->CurrentMemCvg = ((fword & 1) << 2) \| (hbyte & 3);
	2622	}
	2623
	2624	void n64_rdp::_Read32Bit_ImgRead0(UINT32 curpixel, rdp_span_aux *userdata, const rdp_poly_state &object)
	2625	{
	2626	UINT32 mem = RREADIDX32((object.MiscState.FBAddress >> 2) + curpixel);
	2627	CHECK32((object.MiscState.FBAddress >> 2) + curpixel);
	2628	userdata->MemoryColor.i.r = (mem >> 24) & 0xff;
	2629	userdata->MemoryColor.i.g = (mem >> 16) & 0xff;
	2630	userdata->MemoryColor.i.b = (mem >> 8) & 0xff;
	2631	userdata->MemoryColor.i.a = 0xff;
	2632	userdata->CurrentMemCvg = 7;
	2633	}
	2634
	2635	void n64_rdp::_Read32Bit_ImgRead1(UINT32 curpixel, rdp_span_aux *userdata, const rdp_poly_state &object)
	2636	{
	2637	UINT32 mem = RREADIDX32((object.MiscState.FBAddress >> 2) + curpixel);
	2638	CHECK32((object.MiscState.FBAddress >> 2) + curpixel);
	2639	userdata->MemoryColor.i.r = (mem >> 24) & 0xff;
	2640	userdata->MemoryColor.i.g = (mem >> 16) & 0xff;
	2641	userdata->MemoryColor.i.b = (mem >> 8) & 0xff;
	2642	userdata->MemoryColor.i.a = (mem) & 0xff;
	2643	userdata->CurrentMemCvg = (mem >> 5) & 7;
	2644	}
	2645
	2646	void n64_rdp::_Copy16Bit(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, int CurrentPixCvg, const rdp_poly_state &object)
	2647	{
	2648	UINT16 val = ((r >> 3) << 11) \| ((g >> 3) << 6) \| ((b >> 3) << 1) \| ((CurrentPixCvg >> 2) & 1);
	2649	RWRITEIDX16((object.MiscState.FBAddress >> 1) + curpixel, val);
	2650	HWRITEADDR8((object.MiscState.FBAddress >> 1) + curpixel, CurrentPixCvg & 3);
	2651	CHECK16((object.MiscState.FBAddress >> 1) + curpixel);
	2652	}
	2653
	2654	void n64_rdp::_Copy32Bit(UINT32 curpixel, UINT32 r, UINT32 g, UINT32 b, int CurrentPixCvg, const rdp_poly_state &object)
	2655	{
	2656	UINT32 val = (r << 24) \| (g << 16) \| (b << 8) \| (CurrentPixCvg << 5);
	2657	RWRITEIDX32((object.MiscState.FBAddress >> 2) + curpixel, val);
	2658	CHECK32((object.MiscState.FBAddress >> 2) + curpixel);
	2659	}
	2660
	2661
	2662	void n64_rdp::_Fill16Bit(UINT32 curpixel, const rdp_poly_state &object)
	2663	{
	2664	UINT16 val;
	2665	if (curpixel & 1)
	2666	{
	2667	val = object.FillColor & 0xffff;
	2668	}
	2669	else
	2670	{
	2671	val = (object.FillColor >> 16) & 0xffff;
	2672	}
	2673	RWRITEIDX16((object.MiscState.FBAddress >> 1) + curpixel, val);
	2674	CHECK16((object.MiscState.FBAddress >> 1) + curpixel);
	2675	HWRITEADDR8((object.MiscState.FBAddress >> 1) + curpixel, ((val & 1) << 1) \| (val & 1));
	2676	}
	2677
	2678	void n64_rdp::_Fill32Bit(UINT32 curpixel, const rdp_poly_state &object)
	2679	{
	2680	UINT32 FillColor = object.FillColor;
	2681	RWRITEIDX32((object.MiscState.FBAddress >> 2) + curpixel, FillColor);
	2682	CHECK32((object.MiscState.FBAddress >> 2) + curpixel);
	2683	HWRITEADDR8((object.MiscState.FBAddress >> 1) + (curpixel << 1), (FillColor & 0x10000) ? 3 : 0);
	2684	HWRITEADDR8((object.MiscState.FBAddress >> 1) + (curpixel << 1) + 1, (FillColor & 0x1) ? 3 : 0);
	2685	}
	2686
	2687	////////////////////////
	2688	// RDP COMMANDS
	2689	////////////////////////
	2690	void n64_rdp::Triangle(bool shade, bool texture, bool zbuffer)
	2691	{
	2692	DrawTriangle(shade, texture, zbuffer, false);
	2693	m_pipe_clean = false;
	2694	//wait();
	2695	}
	2696
	2697	void n64_rdp::CmdTriangle(UINT32 w1, UINT32 w2)
	2698	{
	2699	Triangle(false, false, false);
	2700	}
	2701
	2702	void n64_rdp::CmdTriangleZ(UINT32 w1, UINT32 w2)
	2703	{
	2704	Triangle(false, false, true);
	2705	}
	2706
	2707	void n64_rdp::CmdTriangleT(UINT32 w1, UINT32 w2)
	2708	{
	2709	Triangle(false, true, false);
	2710	}
	2711
	2712	void n64_rdp::CmdTriangleTZ(UINT32 w1, UINT32 w2)
	2713	{
	2714	Triangle(false, true, true);
	2715	}
	2716
	2717	void n64_rdp::CmdTriangleS(UINT32 w1, UINT32 w2)
	2718	{
	2719	Triangle(true, false, false);
	2720	}
	2721
	2722	void n64_rdp::CmdTriangleSZ(UINT32 w1, UINT32 w2)
	2723	{
	2724	Triangle(true, false, true);
	2725	}
	2726
	2727	void n64_rdp::CmdTriangleST(UINT32 w1, UINT32 w2)
	2728	{
	2729	Triangle(true, true, false);
	2730	}
	2731
	2732	void n64_rdp::CmdTriangleSTZ(UINT32 w1, UINT32 w2)
	2733	{
	2734	Triangle(true, true, true);
	2735	}
	2736
	2737	void n64_rdp::CmdTexRect(UINT32 w1, UINT32 w2)
	2738	{
	2739	//if(m_pending_mode_block) { wait("Block on pending mode-change"); m_pending_mode_block = false; }
	2740
	2741	UINT32 *data = m_cmd_data + m_cmd_cur;
	2742
	2743	UINT32 w3 = data[2];
	2744	UINT32 w4 = data[3];
	2745
	2746	UINT32 tilenum = (w2 >> 24) & 0x7;
	2747	UINT32 xl = (w1 >> 12) & 0xfff;
	2748	UINT32 yl = (w1 >> 0) & 0xfff;
	2749	UINT32 xh = (w2 >> 12) & 0xfff;
	2750	UINT32 yh = (w2 >> 0) & 0xfff;
	2751	INT32 s = (w3 >> 16) & 0xffff;
	2752	INT32 t = (w3 >> 0) & 0xffff;
	2753	INT32 dsdx = (w4 >> 16) & 0xffff;
	2754	INT32 dtdy = (w4 >> 0) & 0xffff;
	2755
	2756	dsdx = SIGN16(dsdx);
	2757	dtdy = SIGN16(dtdy);
	2758
	2759	if (OtherModes.cycle_type == CYCLE_TYPE_FILL \|\| OtherModes.cycle_type == CYCLE_TYPE_COPY)
	2760	{
	2761	yl \|= 3;
	2762	}
	2763
	2764	UINT32 xlint = (xl >> 2) & 0x3ff;
	2765	UINT32 xhint = (xh >> 2) & 0x3ff;
	2766
	2767	UINT32* ewdata = m_temp_rect_data;
	2768	ewdata[0] = (0x24 << 24) \| ((0x80 \| tilenum) << 16) \| yl; // command, flipped, tile, yl
	2769	ewdata[1] = (yl << 16) \| yh; // ym, yh
	2770	ewdata[2] = (xlint << 16) \| ((xl & 3) << 14); // xl, xl frac
	2771	ewdata[3] = 0; // dxldy, dxldy frac
	2772	ewdata[4] = (xhint << 16) \| ((xh & 3) << 14); // xh, xh frac
	2773	ewdata[5] = 0; // dxhdy, dxhdy frac
	2774	ewdata[6] = (xlint << 16) \| ((xl & 3) << 14); // xm, xm frac
	2775	ewdata[7] = 0; // dxmdy, dxmdy frac
	2776	memset(&ewdata[8], 0, 16 * sizeof(UINT32)); // shade
	2777	ewdata[24] = (s << 16) \| t; // s, t
	2778	ewdata[25] = 0; // w
	2779	ewdata[26] = ((dsdx >> 5) << 16); // dsdx, dtdx
	2780	ewdata[27] = 0; // dwdx
	2781	ewdata[28] = 0; // s frac, t frac
	2782	ewdata[29] = 0; // w frac
	2783	ewdata[30] = ((dsdx & 0x1f) << 11) << 16; // dsdx frac, dtdx frac
	2784	ewdata[31] = 0; // dwdx frac
	2785	ewdata[32] = (dtdy >> 5) & 0xffff;//dsde, dtde
	2786	ewdata[33] = 0;//dwde
	2787	ewdata[34] = (dtdy >> 5) & 0xffff;//dsdy, dtdy
	2788	ewdata[35] = 0;//dwdy
	2789	ewdata[36] = (dtdy & 0x1f) << 11;//dsde frac, dtde frac
	2790	ewdata[37] = 0;//dwde frac
	2791	ewdata[38] = (dtdy & 0x1f) << 11;//dsdy frac, dtdy frac
	2792	ewdata[39] = 0;//dwdy frac
	2793	memset(&ewdata[40], 0, 4 * sizeof(UINT32));//depth
	2794
	2795	DrawTriangle(true, true, false, true);
	2796	}
	2797
	2798	void n64_rdp::CmdTexRectFlip(UINT32 w1, UINT32 w2)
	2799	{
	2800	//if(m_pending_mode_block) { wait("Block on pending mode-change"); m_pending_mode_block = false; }
	2801
	2802	UINT32 *data = m_cmd_data + m_cmd_cur;
	2803
	2804	UINT32 w3 = data[2];
	2805	UINT32 w4 = data[3];
	2806
	2807	UINT32 tilenum = (w2 >> 24) & 0x7;
	2808	UINT32 xl = (w1 >> 12) & 0xfff;
	2809	UINT32 yl = (w1 >> 0) & 0xfff;
	2810	UINT32 xh = (w2 >> 12) & 0xfff;
	2811	UINT32 yh = (w2 >> 0) & 0xfff;
	2812	INT32 s = (w3 >> 16) & 0xffff;
	2813	INT32 t = (w3 >> 0) & 0xffff;
	2814	INT32 dsdx = (w4 >> 16) & 0xffff;
	2815	INT32 dtdy = (w4 >> 0) & 0xffff;
	2816
	2817	dsdx = SIGN16(dsdx);
	2818	dtdy = SIGN16(dtdy);
	2819
	2820	if (OtherModes.cycle_type == CYCLE_TYPE_FILL \|\| OtherModes.cycle_type == CYCLE_TYPE_COPY)
	2821	{
	2822	yl \|= 3;
	2823	}
	2824
	2825	UINT32 xlint = (xl >> 2) & 0x3ff;
	2826	UINT32 xhint = (xh >> 2) & 0x3ff;
	2827
	2828	UINT32* ewdata = m_temp_rect_data;
	2829	ewdata[0] = (0x25 << 24) \| ((0x80 \| tilenum) << 16) \| yl;//command, flipped, tile, yl
	2830	ewdata[1] = (yl << 16) \| yh;//ym, yh
	2831	ewdata[2] = (xlint << 16) \| ((xl & 3) << 14);//xl, xl frac
	2832	ewdata[3] = 0;//dxldy, dxldy frac
	2833	ewdata[4] = (xhint << 16) \| ((xh & 3) << 14);//xh, xh frac
	2834	ewdata[5] = 0;//dxhdy, dxhdy frac
	2835	ewdata[6] = (xlint << 16) \| ((xl & 3) << 14);//xm, xm frac
	2836	ewdata[7] = 0;//dxmdy, dxmdy frac
	2837	memset(&ewdata[8], 0, 16 * sizeof(UINT32));//shade
	2838	ewdata[24] = (s << 16) \| t;//s, t
	2839	ewdata[25] = 0;//w
	2840	ewdata[26] = (dtdy >> 5) & 0xffff;//dsdx, dtdx
	2841	ewdata[27] = 0;//dwdx
	2842	ewdata[28] = 0;//s frac, t frac
	2843	ewdata[29] = 0;//w frac
	2844	ewdata[30] = ((dtdy & 0x1f) << 11);//dsdx frac, dtdx frac
	2845	ewdata[31] = 0;//dwdx frac
	2846	ewdata[32] = (dsdx >> 5) << 16;//dsde, dtde
	2847	ewdata[33] = 0;//dwde
	2848	ewdata[34] = (dsdx >> 5) << 16;//dsdy, dtdy
	2849	ewdata[35] = 0;//dwdy
	2850	ewdata[36] = (dsdx & 0x1f) << 27;//dsde frac, dtde frac
	2851	ewdata[37] = 0;//dwde frac
	2852	ewdata[38] = (dsdx & 0x1f) << 27;//dsdy frac, dtdy frac
	2853	ewdata[39] = 0;//dwdy frac
	2854	memset(&ewdata[40], 0, 4 * sizeof(UINT32));//depth
	2855
	2856	DrawTriangle(true, true, false, true);
	2857	}
	2858
	2859	void n64_rdp::CmdSyncLoad(UINT32 w1, UINT32 w2)
	2860	{
	2861	//wait("SyncLoad");
	2862	}
	2863
	2864	void n64_rdp::CmdSyncPipe(UINT32 w1, UINT32 w2)
	2865	{
	2866	//wait("SyncPipe");
	2867	}
	2868
	2869	void n64_rdp::CmdSyncTile(UINT32 w1, UINT32 w2)
	2870	{
	2871	//wait("SyncTile");
	2872	}
	2873
	2874	void n64_rdp::CmdSyncFull(UINT32 w1, UINT32 w2)
	2875	{
	2876	//wait("SyncFull");
	2877	dp_full_sync(*m_machine);
	2878	}
	2879
	2880	void n64_rdp::CmdSetKeyGB(UINT32 w1, UINT32 w2)
	2881	{
	2882	KeyScale.i.b = w2 & 0xff;
	2883	KeyScale.i.g = (w2 >> 16) & 0xff;
	2884	}
	2885
	2886	void n64_rdp::CmdSetKeyR(UINT32 w1, UINT32 w2)
	2887	{
	2888	KeyScale.i.r = w2 & 0xff;
	2889	}
	2890
	2891	void n64_rdp::CmdSetFillColor32(UINT32 w1, UINT32 w2)
	2892	{
	2893	//wait("SetFillColor");
	2894	FillColor = w2;
	2895	}
	2896
	2897	void n64_rdp::CmdSetConvert(UINT32 w1, UINT32 w2)
	2898	{
	2899	if(!m_pipe_clean) { m_pipe_clean = true; wait("SetConvert"); }
	2900	INT32 k0 = (w1 >> 13) & 0xff;
	2901	INT32 k1 = (w1 >> 4) & 0xff;
	2902	INT32 k2 = ((w1 & 7) << 5) \| ((w2 >> 27) & 0x1f);
	2903	INT32 k3 = (w2 >> 18) & 0xff;
	2904	INT32 k4 = (w2 >> 9) & 0xff;
	2905	INT32 k5 = w2 & 0xff;
	2906	k0 = ((w1 >> 21) & 1) ? (-(0x100 - k0)) : k0;
	2907	k1 = ((w1 >> 12) & 1) ? (-(0x100 - k1)) : k1;
	2908	k2 = (w1 & 0xf) ? (-(0x100 - k2)) : k2;
	2909	k3 = ((w2 >> 26) & 1) ? (-(0x100 - k3)) : k3;
	2910	k4 = ((w2 >> 17) & 1) ? (-(0x100 - k4)) : k4;
	2911	k5 = ((w2 >> 8) & 1) ? (-(0x100 - k5)) : k5;
	2912	SetYUVFactors(k0, k1, k2, k3, k4, k5);
	2913	}
	2914
	2915	void n64_rdp::CmdSetScissor(UINT32 w1, UINT32 w2)
	2916	{
	2917	Scissor.m_xh = ((w1 >> 12) & 0xfff) >> 2;
	2918	Scissor.m_yh = ((w1 >> 0) & 0xfff) >> 2;
	2919	Scissor.m_xl = ((w2 >> 12) & 0xfff) >> 2;
	2920	Scissor.m_yl = ((w2 >> 0) & 0xfff) >> 2;
	2921
	2922	// TODO: handle f & o?
	2923	}
	2924
	2925	void n64_rdp::CmdSetPrimDepth(UINT32 w1, UINT32 w2)
	2926	{
	2927	MiscState.PrimitiveZ = (UINT16)(w2 >> 16) & 0x7fff;
	2928	MiscState.PrimitiveDZ = (UINT16)(w1);
	2929	}
	2930
	2931	void n64_rdp::CmdSetOtherModes(UINT32 w1, UINT32 w2)
	2932	{
	2933	//wait("SetOtherModes");
	2934	OtherModes.cycle_type = (w1 >> 20) & 0x3; // 01
	2935	OtherModes.persp_tex_en = (w1 & 0x80000) ? 1 : 0; // 1
	2936	OtherModes.detail_tex_en = (w1 & 0x40000) ? 1 : 0; // 0
	2937	OtherModes.sharpen_tex_en = (w1 & 0x20000) ? 1 : 0; // 0
	2938	OtherModes.tex_lod_en = (w1 & 0x10000) ? 1 : 0; // 0
	2939	OtherModes.en_tlut = (w1 & 0x08000) ? 1 : 0; // 0
	2940	OtherModes.tlut_type = (w1 & 0x04000) ? 1 : 0; // 0
	2941	OtherModes.sample_type = (w1 & 0x02000) ? 1 : 0; // 1
	2942	OtherModes.mid_texel = (w1 & 0x01000) ? 1 : 0; // 0
	2943	OtherModes.bi_lerp0 = (w1 & 0x00800) ? 1 : 0; // 1
	2944	OtherModes.bi_lerp1 = (w1 & 0x00400) ? 1 : 0; // 1
	2945	OtherModes.convert_one = (w1 & 0x00200) ? 1 : 0; // 0
	2946	OtherModes.key_en = (w1 & 0x00100) ? 1 : 0; // 0
	2947	OtherModes.rgb_dither_sel = (w1 >> 6) & 0x3; // 00
	2948	OtherModes.alpha_dither_sel = (w1 >> 4) & 0x3; // 01
	2949	OtherModes.blend_m1a_0 = (w2 >> 30) & 0x3; // 11
	2950	OtherModes.blend_m1a_1 = (w2 >> 28) & 0x3; // 00
	2951	OtherModes.blend_m1b_0 = (w2 >> 26) & 0x3; // 10
	2952	OtherModes.blend_m1b_1 = (w2 >> 24) & 0x3; // 00
	2953	OtherModes.blend_m2a_0 = (w2 >> 22) & 0x3; // 00
	2954	OtherModes.blend_m2a_1 = (w2 >> 20) & 0x3; // 01
	2955	OtherModes.blend_m2b_0 = (w2 >> 18) & 0x3; // 00
	2956	OtherModes.blend_m2b_1 = (w2 >> 16) & 0x3; // 01
	2957	OtherModes.force_blend = (w2 >> 14) & 1; // 0
	2958	OtherModes.alpha_cvg_select = (w2 >> 13) & 1; // 1
	2959	OtherModes.cvg_times_alpha = (w2 >> 12) & 1; // 0
	2960	OtherModes.z_mode = (w2 >> 10) & 0x3; // 00
	2961	OtherModes.cvg_dest = (w2 >> 8) & 0x3; // 00
	2962	OtherModes.color_on_cvg = (w2 >> 7) & 1; // 0
	2963	OtherModes.image_read_en = (w2 >> 6) & 1; // 1
	2964	OtherModes.z_update_en = (w2 >> 5) & 1; // 1
	2965	OtherModes.z_compare_en = (w2 >> 4) & 1; // 1
	2966	OtherModes.antialias_en = (w2 >> 3) & 1; // 1
	2967	OtherModes.z_source_sel = (w2 >> 2) & 1; // 0
	2968	OtherModes.dither_alpha_en = (w2 >> 1) & 1; // 0
	2969	OtherModes.alpha_compare_en = (w2) & 1; // 0
	2970	}
	2971
	2972	void n64_rdp::CmdLoadTLUT(UINT32 w1, UINT32 w2)
	2973	{
	2974	//wait("LoadTLUT");
	2975	N64Tile* tile = m_tiles;
	2976
	2977	int tilenum = (w2 >> 24) & 0x7;
	2978	int sl = tile[tilenum].sl = ((w1 >> 12) & 0xfff);
	2979	int tl = tile[tilenum].tl = w1 & 0xfff;
	2980	int sh = tile[tilenum].sh = ((w2 >> 12) & 0xfff);
	2981	int th = tile[tilenum].th = w2 & 0xfff;
	2982
	2983	if (tl != th)
	2984	{
	2985	fatalerror("Load tlut: tl=%d, th=%d\n",tl,th);
	2986	}
	2987
	2988	int count = (sh >> 2) - (sl >> 2) + 1;
	2989	count <<= 2;
	2990
	2991	switch (MiscState.TISize)
	2992	{
	2993	case PIXEL_SIZE_16BIT:
	2994	{
	2995	if (tile[tilenum].tmem < 256)
	2996	{
	2997	fatalerror("rdp_load_tlut: loading tlut into low half at %d qwords\n",tile[tilenum].tmem);
	2998	}
	2999	UINT32 srcstart = (MiscState.TIAddress + (tl >> 2) * (MiscState.TIWidth << 1) + (sl >> 1)) >> 1;
	3000	UINT16 *dst = GetTMEM16();
	3001	UINT32 dststart = tile[tilenum].tmem << 2;
	3002
	3003	for (int i = 0; i < count; i += 4)
	3004	{
	3005	if (dststart < 2048)
	3006	{
	3007	dst[dststart] = U_RREADIDX16(srcstart);
	3008	dst[dststart + 1] = dst[dststart];
	3009	dst[dststart + 2] = dst[dststart];
	3010	dst[dststart + 3] = dst[dststart];
	3011	dststart += 4;
	3012	srcstart += 1;
	3013	}
	3014	}
	3015	break;
	3016	}
	3017	default: fatalerror("RDP: load_tlut: size = %d\n", MiscState.TISize);
	3018	}
	3019	}
	3020
	3021	void n64_rdp::CmdSetTileSize(UINT32 w1, UINT32 w2)
	3022	{
	3023	//wait("SetTileSize");
	3024
	3025	const int tilenum = (w2 >> 24) & 0x7;
	3026
	3027	m_tiles[tilenum].sl = (w1 >> 12) & 0xfff;
	3028	m_tiles[tilenum].tl = (w1 >> 0) & 0xfff;
	3029	m_tiles[tilenum].sh = (w2 >> 12) & 0xfff;
	3030	m_tiles[tilenum].th = (w2 >> 0) & 0xfff;
	3031	}
	3032
	3033	void n64_rdp::CmdLoadBlock(UINT32 w1, UINT32 w2)
	3034	{
	3035	//wait("LoadBlock");
	3036	N64Tile* tile = m_tiles;
	3037
	3038	int tilenum = (w2 >> 24) & 0x7;
	3039	UINT16* tc = GetTMEM16();
	3040
	3041	UINT16 sl, sh, tl;
	3042	tile[tilenum].sl = sl = ((w1 >> 12) & 0xfff);
	3043	tile[tilenum].tl = tl = ((w1 >> 0) & 0xfff);
	3044	tile[tilenum].sh = sh = ((w2 >> 12) & 0xfff);
	3045	UINT16 dxt = ((w2 >> 0) & 0xfff);
	3046
	3047	if (sh < sl)
	3048	{
	3049	fatalerror("load_block: sh < sl\n");
	3050	}
	3051
	3052	INT32 width = (sh - sl) + 1;
	3053
	3054	width = (width << MiscState.TISize) >> 1;
	3055	if (width & 7)
	3056	{
	3057	width = (width & ~7) + 8;
	3058	}
	3059	width >>= 3;
	3060
	3061	UINT32 tb = tile[tilenum].tmem << 2;
	3062
	3063	int tiwinwords = MiscState.TIWidth;
	3064	UINT32 slinwords = sl;
	3065
	3066	tiwinwords = (tiwinwords << MiscState.TISize) >> 2;
	3067	slinwords = (slinwords << MiscState.TISize) >> 2;
	3068
	3069	int ptr = 0, srcptr = 0;
	3070	UINT16 first, sec;
	3071	UINT32 src = (MiscState.TIAddress >> 1) + (tl * tiwinwords) + slinwords;
	3072
	3073	if (dxt != 0)
	3074	{
	3075	int j = 0;
	3076	int t = 0;
	3077	int oldt = 0;
	3078
	3079	if (tile[tilenum].size != PIXEL_SIZE_32BIT && tile[tilenum].format != FORMAT_YUV)
	3080	{
	3081	for (int i = 0; i < width; i ++)
	3082	{
	3083	oldt = t;
	3084	t = ((j >> 11) & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
	3085	if (t != oldt)
	3086	{
	3087	i += tile[tilenum].line;
	3088	}
	3089
	3090	ptr = tb + (i << 2);
	3091	srcptr = src + (i << 2);
	3092
	3093	tc[(ptr ^ t) & 0x7ff] = U_RREADIDX16(srcptr);
	3094	tc[((ptr + 1) ^ t) & 0x7ff] = U_RREADIDX16(srcptr + 1);
	3095	tc[((ptr + 2) ^ t) & 0x7ff] = U_RREADIDX16(srcptr + 2);
	3096	tc[((ptr + 3) ^ t) & 0x7ff] = U_RREADIDX16(srcptr + 3);
	3097	j += dxt;
	3098	}
	3099	}
	3100	else if (tile[tilenum].format == FORMAT_YUV)
	3101	{
	3102	for (int i = 0; i < width; i ++)
	3103	{
	3104	oldt = t;
	3105	t = ((j >> 11) & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
	3106	if (t != oldt)
	3107	{
	3108	i += tile[tilenum].line;
	3109	}
	3110
	3111	ptr = ((tb + (i << 1)) ^ t) & 0x3ff;
	3112	srcptr = src + (i << 2);
	3113
	3114	first = U_RREADIDX16(srcptr);
	3115	sec = U_RREADIDX16(srcptr + 1);
	3116	tc[ptr] = ((first >> 8) << 8) \| (sec >> 8);
	3117	tc[ptr \| 0x400] = ((first & 0xff) << 8) \| (sec & 0xff);
	3118
	3119	ptr = ((tb + (i << 1) + 1) ^ t) & 0x3ff;
	3120	first = U_RREADIDX16(srcptr + 2);
	3121	sec = U_RREADIDX16(srcptr + 3);
	3122	tc[ptr] = ((first >> 8) << 8) \| (sec >> 8);
	3123	tc[ptr \| 0x400] = ((first & 0xff) << 8) \| (sec & 0xff);
	3124
	3125	j += dxt;
	3126	}
	3127	}
	3128	else
	3129	{
	3130	for (int i = 0; i < width; i ++)
	3131	{
	3132	oldt = t;
	3133	t = ((j >> 11) & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
	3134	if (t != oldt)
	3135	i += tile[tilenum].line;
	3136
	3137	ptr = ((tb + (i << 1)) ^ t) & 0x3ff;
	3138	srcptr = src + (i << 2);
	3139	tc[ptr] = U_RREADIDX16(srcptr);
	3140	tc[ptr \| 0x400] = U_RREADIDX16(srcptr + 1);
	3141
	3142	ptr = ((tb + (i << 1) + 1) ^ t) & 0x3ff;
	3143	tc[ptr] = U_RREADIDX16(srcptr + 2);
	3144	tc[ptr \| 0x400] = U_RREADIDX16(srcptr + 3);
	3145
	3146	j += dxt;
	3147	}
	3148	}
	3149	tile[tilenum].th = tl + (j >> 11);
	3150	}
	3151	else
	3152	{
	3153	if (tile[tilenum].size != PIXEL_SIZE_32BIT && tile[tilenum].format != FORMAT_YUV)
	3154	{
	3155	for (int i = 0; i < width; i ++)
	3156	{
	3157	ptr = tb + (i << 2);
	3158	srcptr = src + (i << 2);
	3159	tc[(ptr ^ WORD_ADDR_XOR) & 0x7ff] = U_RREADIDX16(srcptr);
	3160	tc[((ptr + 1) ^ WORD_ADDR_XOR) & 0x7ff] = U_RREADIDX16(srcptr + 1);
	3161	tc[((ptr + 2) ^ WORD_ADDR_XOR) & 0x7ff] = U_RREADIDX16(srcptr + 2);
	3162	tc[((ptr + 3) ^ WORD_ADDR_XOR) & 0x7ff] = U_RREADIDX16(srcptr + 3);
	3163	}
	3164	}
	3165	else if (tile[tilenum].format == FORMAT_YUV)
	3166	{
	3167	for (int i = 0; i < width; i ++)
	3168	{
	3169	ptr = ((tb + (i << 1)) ^ WORD_ADDR_XOR) & 0x3ff;
	3170	srcptr = src + (i << 2);
	3171	first = U_RREADIDX16(srcptr);
	3172	sec = U_RREADIDX16(srcptr + 1);
	3173	tc[ptr] = ((first >> 8) << 8) \| (sec >> 8);//UV pair
	3174	tc[ptr \| 0x400] = ((first & 0xff) << 8) \| (sec & 0xff);
	3175
	3176	ptr = ((tb + (i << 1) + 1) ^ WORD_ADDR_XOR) & 0x3ff;
	3177	first = U_RREADIDX16(srcptr + 2);
	3178	sec = U_RREADIDX16(srcptr + 3);
	3179	tc[ptr] = ((first >> 8) << 8) \| (sec >> 8);
	3180	tc[ptr \| 0x400] = ((first & 0xff) << 8) \| (sec & 0xff);
	3181	}
	3182	}
	3183	else
	3184	{
	3185	for (int i = 0; i < width; i ++)
	3186	{
	3187	ptr = ((tb + (i << 1)) ^ WORD_ADDR_XOR) & 0x3ff;
	3188	srcptr = src + (i << 2);
	3189	tc[ptr] = U_RREADIDX16(srcptr);
	3190	tc[ptr \| 0x400] = U_RREADIDX16(srcptr + 1);
	3191
	3192	ptr = ((tb + (i << 1) + 1) ^ WORD_ADDR_XOR) & 0x3ff;
	3193	tc[ptr] = U_RREADIDX16(srcptr + 2);
	3194	tc[ptr \| 0x400] = U_RREADIDX16(srcptr + 3);
	3195	}
	3196	}
	3197	tile[tilenum].th = tl;
	3198	}
	3199	}
	3200
	3201	void n64_rdp::CmdLoadTile(UINT32 w1, UINT32 w2)
	3202	{
	3203	//wait("LoadTile");
	3204	N64Tile* tile = m_tiles;
	3205	int tilenum = (w2 >> 24) & 0x7;
	3206
	3207	tile[tilenum].sl = ((w1 >> 12) & 0xfff);
	3208	tile[tilenum].tl = ((w1 >> 0) & 0xfff);
	3209	tile[tilenum].sh = ((w2 >> 12) & 0xfff);
	3210	tile[tilenum].th = ((w2 >> 0) & 0xfff);
	3211
	3212	UINT16 sl = tile[tilenum].sl >> 2;
	3213	UINT16 tl = tile[tilenum].tl >> 2;
	3214	UINT16 sh = tile[tilenum].sh >> 2;
	3215	UINT16 th = tile[tilenum].th >> 2;
	3216
	3217	INT32 width = (sh - sl) + 1;
	3218	INT32 height = (th - tl) + 1;
	3219	/*
	3220	int topad;
	3221	if (MiscState.TISize < 3)
	3222	{
	3223	topad = (width * MiscState.TISize) & 0x7;
	3224	}
	3225	else
	3226	{
	3227	topad = (width << 2) & 0x7;
	3228	}
	3229	topad = 0; // ????
	3230	*/
	3231	switch (MiscState.TISize)
	3232	{
	3233	case PIXEL_SIZE_8BIT:
	3234	{
	3235	UINT32 src = MiscState.TIAddress;
	3236	UINT8 *tc = GetTMEM();
	3237	int tb = tile[tilenum].tmem << 3;
	3238
	3239	for (int j = 0; j < height; j++)
	3240	{
	3241	int tline = tb + ((tile[tilenum].line << 3) * j);
	3242	int s = ((j + tl) * MiscState.TIWidth) + sl;
	3243
	3244	int xorval8 = ((j & 1) ? BYTE_XOR_DWORD_SWAP : BYTE_ADDR_XOR);
	3245	for (int i = 0; i < width; i++)
	3246	{
	3247	tc[((tline + i) ^ xorval8) & 0xfff] = U_RREADADDR8(src + s + i);
	3248	}
	3249	}
	3250	break;
	3251	}
	3252	case PIXEL_SIZE_16BIT:
	3253	{
	3254	UINT32 src = MiscState.TIAddress >> 1;
	3255	UINT16 *tc = GetTMEM16();
	3256	UINT16 yuvword;
	3257
	3258	if (tile[tilenum].format != FORMAT_YUV)
	3259	{
	3260	for (int j = 0; j < height; j++)
	3261	{
	3262	int tb = tile[tilenum].tmem << 2;
	3263	int tline = tb + ((tile[tilenum].line << 2) * j);
	3264	int s = ((j + tl) * MiscState.TIWidth) + sl;
	3265	int xorval16 = (j & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
	3266
	3267	for (int i = 0; i < width; i++)
	3268	{
	3269	UINT32 taddr = (tline + i) ^ xorval16;
	3270	tc[taddr & 0x7ff] = U_RREADIDX16(src + s + i);
	3271	}
	3272	}
	3273	}
	3274	else
	3275	{
	3276	for (int j = 0; j < height; j++)
	3277	{
	3278	int tb = tile[tilenum].tmem << 3;
	3279	int tline = tb + ((tile[tilenum].line << 3) * j);
	3280	int s = ((j + tl) * MiscState.TIWidth) + sl;
	3281	int xorval8 = (j & 1) ? BYTE_XOR_DWORD_SWAP : BYTE_ADDR_XOR;
	3282
	3283	for (int i = 0; i < width; i++)
	3284	{
	3285	UINT32 taddr = ((tline + i) ^ xorval8) & 0x7ff;
	3286	yuvword = U_RREADIDX16(src + s + i);
	3287	GetTMEM()[taddr] = yuvword >> 8;
	3288	GetTMEM()[taddr \| 0x800] = yuvword & 0xff;
	3289	}
	3290	}
	3291	}
	3292	break;
	3293	}
	3294	case PIXEL_SIZE_32BIT:
	3295	{
	3296	UINT32 src = MiscState.TIAddress >> 2;
	3297	UINT16 *tc16 = GetTMEM16();
	3298	int tb = (tile[tilenum].tmem << 2);
	3299
	3300	for (int j = 0; j < height; j++)
	3301	{
	3302	int tline = tb + ((tile[tilenum].line << 2) * j);
	3303
	3304	int s = ((j + tl) * MiscState.TIWidth) + sl;
	3305	int xorval32cur = (j & 1) ? WORD_XOR_DWORD_SWAP : WORD_ADDR_XOR;
	3306	for (int i = 0; i < width; i++)
	3307	{
	3308	UINT32 c = U_RREADIDX32(src + s + i);
	3309	UINT32 ptr = ((tline + i) ^ xorval32cur) & 0x3ff;
	3310	tc16[ptr] = c >> 16;
	3311	tc16[ptr \| 0x400] = c & 0xffff;
	3312	}
	3313	}
	3314	break;
	3315	}
	3316
	3317	default: fatalerror("RDP: load_tile: size = %d\n", MiscState.TISize);
	3318	}
	3319	}
	3320
	3321	void n64_rdp::CmdSetTile(UINT32 w1, UINT32 w2)
	3322	{
	3323	//wait("SetTile");
	3324
	3325	int tilenum = (w2 >> 24) & 0x7;
	3326	N64Tile* tex_tile = &m_tiles[tilenum];
	3327
	3328	tex_tile->format = (w1 >> 21) & 0x7;
	3329	tex_tile->size = (w1 >> 19) & 0x3;
	3330	tex_tile->line = (w1 >> 9) & 0x1ff;
	3331	tex_tile->tmem = (w1 >> 0) & 0x1ff;
	3332	tex_tile->palette = (w2 >> 20) & 0xf;
	3333	tex_tile->ct = (w2 >> 19) & 0x1;
	3334	tex_tile->mt = (w2 >> 18) & 0x1;
	3335	tex_tile->mask_t = (w2 >> 14) & 0xf;
	3336	tex_tile->shift_t = (w2 >> 10) & 0xf;
	3337	tex_tile->cs = (w2 >> 9) & 0x1;
	3338	tex_tile->ms = (w2 >> 8) & 0x1;
	3339	tex_tile->mask_s = (w2 >> 4) & 0xf;
	3340	tex_tile->shift_s = (w2 >> 0) & 0xf;
	3341
	3342	if (tex_tile->format == FORMAT_I && tex_tile->size > PIXEL_SIZE_8BIT)
	3343	{
	3344	tex_tile->format = FORMAT_RGBA; // Used by Supercross 2000 (in-game)
	3345	}
	3346	if (tex_tile->format == FORMAT_CI && tex_tile->size > PIXEL_SIZE_8BIT)
	3347	{
	3348	tex_tile->format = FORMAT_RGBA; // Used by Clay Fighter - Sculptor's Cut
	3349	}
	3350
	3351	if (tex_tile->format == FORMAT_RGBA && tex_tile->size < PIXEL_SIZE_16BIT)
	3352	{
	3353	tex_tile->format = FORMAT_CI; // Used by Exterem-G2, Madden Football 64, and Rat Attack
	3354	}
	3355
	3356	//m_pending_mode_block = true;
	3357	}
	3358
	3359	void n64_rdp::CmdFillRect(UINT32 w1, UINT32 w2)
	3360	{
	3361	//if(m_pending_mode_block) { wait("Block on pending mode-change"); m_pending_mode_block = false; }
	3362
	3363	UINT32 xl = (w1 >> 12) & 0xfff;
	3364	UINT32 yl = (w1 >> 0) & 0xfff;
	3365	UINT32 xh = (w2 >> 12) & 0xfff;
	3366	UINT32 yh = (w2 >> 0) & 0xfff;
	3367
	3368	if (OtherModes.cycle_type == CYCLE_TYPE_FILL \|\| OtherModes.cycle_type == CYCLE_TYPE_COPY)
	3369	{
	3370	yl \|= 3;
	3371	}
	3372
	3373	UINT32 xlint = (xl >> 2) & 0x3ff;
	3374	UINT32 xhint = (xh >> 2) & 0x3ff;
	3375
	3376	UINT32* ewdata = m_temp_rect_data;
	3377	ewdata[0] = (0x3680 << 16) \| yl;//command, flipped, tile, yl
	3378	ewdata[1] = (yl << 16) \| yh;//ym, yh
	3379	ewdata[2] = (xlint << 16) \| ((xl & 3) << 14);//xl, xl frac
	3380	ewdata[3] = 0;//dxldy, dxldy frac
	3381	ewdata[4] = (xhint << 16) \| ((xh & 3) << 14);//xh, xh frac
	3382	ewdata[5] = 0;//dxhdy, dxhdy frac
	3383	ewdata[6] = (xlint << 16) \| ((xl & 3) << 14);//xm, xm frac
	3384	ewdata[7] = 0;//dxmdy, dxmdy frac
	3385	memset(&ewdata[8], 0, 36 * sizeof(UINT32));//shade, texture, depth
	3386
	3387	DrawTriangle(false, false, false, true);
	3388	}
	3389
	3390	void n64_rdp::CmdSetFogColor(UINT32 w1, UINT32 w2)
	3391	{
	3392	FogColor.c = w2;
	3393	}
	3394
	3395	void n64_rdp::CmdSetBlendColor(UINT32 w1, UINT32 w2)
	3396	{
	3397	BlendColor.c = w2;
	3398	}
	3399
	3400	void n64_rdp::CmdSetPrimColor(UINT32 w1, UINT32 w2)
	3401	{
	3402	MiscState.MinLevel = (w1 >> 8) & 0x1f;
	3403	PrimLODFraction = w1 & 0xff;
	3404	PrimColor.c = w2;
	3405	}
	3406
	3407	void n64_rdp::CmdSetEnvColor(UINT32 w1, UINT32 w2)
	3408	{
	3409	EnvColor.c = w2;
	3410	}
	3411
	3412	void n64_rdp::CmdSetCombine(UINT32 w1, UINT32 w2)
	3413	{
	3414	m_combine.sub_a_rgb0 = (w1 >> 20) & 0xf;
	3415	m_combine.mul_rgb0 = (w1 >> 15) & 0x1f;
	3416	m_combine.sub_a_a0 = (w1 >> 12) & 0x7;
	3417	m_combine.mul_a0 = (w1 >> 9) & 0x7;
	3418	m_combine.sub_a_rgb1 = (w1 >> 5) & 0xf;
	3419	m_combine.mul_rgb1 = (w1 >> 0) & 0x1f;
	3420
	3421	m_combine.sub_b_rgb0 = (w2 >> 28) & 0xf;
	3422	m_combine.sub_b_rgb1 = (w2 >> 24) & 0xf;
	3423	m_combine.sub_a_a1 = (w2 >> 21) & 0x7;
	3424	m_combine.mul_a1 = (w2 >> 18) & 0x7;
	3425	m_combine.add_rgb0 = (w2 >> 15) & 0x7;
	3426	m_combine.sub_b_a0 = (w2 >> 12) & 0x7;
	3427	m_combine.add_a0 = (w2 >> 9) & 0x7;
	3428	m_combine.add_rgb1 = (w2 >> 6) & 0x7;
	3429	m_combine.sub_b_a1 = (w2 >> 3) & 0x7;
	3430	m_combine.add_a1 = (w2 >> 0) & 0x7;
	3431	}
	3432
	3433	void n64_rdp::CmdSetTextureImage(UINT32 w1, UINT32 w2)
	3434	{
	3435	MiscState.TIFormat = (w1 >> 21) & 0x7;
	3436	MiscState.TISize = (w1 >> 19) & 0x3;
	3437	MiscState.TIWidth = (w1 & 0x3ff) + 1;
	3438	MiscState.TIAddress = w2 & 0x01ffffff;
	3439	}
	3440
	3441	void n64_rdp::CmdSetMaskImage(UINT32 w1, UINT32 w2)
	3442	{
	3443	//wait("SetMaskImage");
	3444
	3445	MiscState.ZBAddress = w2 & 0x01ffffff;
	3446	}
	3447
	3448	void n64_rdp::CmdSetColorImage(UINT32 w1, UINT32 w2)
	3449	{
	3450	//wait("SetColorImage");
	3451
	3452	MiscState.FBFormat = (w1 >> 21) & 0x7;
	3453	MiscState.FBSize = (w1 >> 19) & 0x3;
	3454	MiscState.FBWidth = (w1 & 0x3ff) + 1;
	3455	MiscState.FBAddress = w2 & 0x01ffffff;
	3456
	3457	if (MiscState.FBFormat < 2 \|\| MiscState.FBFormat > 32) // Jet Force Gemini sets the format to 4, Intensity. Protection?
	3458	{
	3459	MiscState.FBFormat = 2;
	3460	}
	3461	}
	3462
	3463	UINT32 n64_rdp::AddRightCvg(UINT32 x, UINT32 k)
	3464	{
	3465	//#undef FULL_SUBPIXELS
	3466	#define FULL_SUBPIXELS
	3467	UINT32 coveredsubpixels=((x >> 14) & 3);
	3468	if (!(x & 0xffff))
	3469	{
	3470	return 0;
	3471	}
	3472	#ifdef FULL_SUBPIXELS
	3473	if (!coveredsubpixels)
	3474	{
	3475	return 0;
	3476	}
	3477	if (!(k & 1))
	3478	{
	3479	return (coveredsubpixels<3) ? 1 : 2;
	3480	}
	3481	else
	3482	{
	3483	return (coveredsubpixels<2) ? 0 : 1;
	3484	}
	3485	#endif
	3486	if (!(k & 1))
	3487	{
	3488	return (coveredsubpixels<2) ? 1 : 2;
	3489	}
	3490	else
	3491	{
	3492	if (coveredsubpixels<1)
	3493	{
	3494	return 0;
	3495	}
	3496	else if (coveredsubpixels<3)
	3497	{
	3498	return 1;
	3499	}
	3500	else
	3501	{
	3502	return 2;
	3503	}
	3504	}
	3505	}
	3506
	3507	UINT32 n64_rdp::AddLeftCvg(UINT32 x, UINT32 k)
	3508	{
	3509	UINT32 coveredsubpixels = 3 - ((x >> 14) & 3);
	3510	if (!(x & 0xffff))
	3511	{
	3512	return 2;
	3513	}
	3514	#ifdef FULL_SUBPIXELS
	3515	if (!coveredsubpixels)
	3516	{
	3517	return 0;
	3518	}
	3519	if (!(k & 1))
	3520	{
	3521	return (coveredsubpixels<2) ? 0 : 1;
	3522	}
	3523	else
	3524	{
	3525	return (coveredsubpixels<3) ? 1 : 2;
	3526	}
	3527	#endif
	3528	if (k & 1)
	3529	{
	3530	return (coveredsubpixels<2) ? 1 : 2;
	3531	}
	3532	else
	3533	{
	3534	if (coveredsubpixels < 1)
	3535	{
	3536	return 0;
	3537	}
	3538	else if (coveredsubpixels < 3)
	3539	{
	3540	return 1;
	3541	}
	3542	else
	3543	{
	3544	return 2;
	3545	}
	3546	}
	3547	}
	3548
	3549	/*****************************************************************************/
	3550
	3551	void n64_rdp::CmdInvalid(UINT32 w1, UINT32 w2)
	3552	{
	3553	fatalerror("n64_rdp::Invalid: %d, %08x %08x\n", (w1 >> 24) & 0x3f, w1, w2);
	3554	}
	3555
	3556	void n64_rdp::CmdNoOp(UINT32 w1, UINT32 w2)
	3557	{
	3558	// Do nothing
	3559	}
	3560
	3561
	3562	void n64_rdp::ProcessList()
	3563	{
	3564	INT32 length = m_end - m_current;
	3565
	3566	if(length < 0)
	3567	{
	3568	m_current = m_end;
	3569	return;
	3570	}
	3571
	3572	// load command data
	3573	for(int i = 0; i < length; i += 4)
	3574	{
	3575	m_cmd_data[m_cmd_ptr++] = ReadData((m_current & 0x1fffffff) + i);
	3576	}
	3577
	3578	m_current = m_end;
	3579
	3580	UINT32 cmd = (m_cmd_data[0] >> 24) & 0x3f;
	3581	UINT32 cmd_length = (m_cmd_ptr + 1) * 4;
	3582
	3583	SetStatusReg(GetStatusReg() &~ DP_STATUS_FREEZE);
	3584
	3585	// check if more data is needed
	3586	if (cmd_length < rdp_command_length[cmd])
	3587	{
	3588	return;
	3589	}
	3590
	3591	while (m_cmd_cur < m_cmd_ptr)
	3592	{
	3593	cmd = (m_cmd_data[m_cmd_cur] >> 24) & 0x3f;
	3594
	3595	if (((m_cmd_ptr - m_cmd_cur) * 4) < rdp_command_length[cmd])
	3596	{
	3597	return;
	3598	//fatalerror("rdp_process_list: not enough rdp command data: cur = %d, ptr = %d, expected = %d\n", m_cmd_cur, m_cmd_ptr, rdp_command_length[cmd]);
	3599	}
	3600
	3601	if (LOG_RDP_EXECUTION)
	3602	{
	3603	char string[4000];
	3604	Dasm(string);
	3605
	3606	fprintf(rdp_exec, "%08X: %08X %08X %s\n", m_start+(m_cmd_cur * 4), m_cmd_data[m_cmd_cur+0], m_cmd_data[m_cmd_cur+1], string);
	3607	fflush(rdp_exec);
	3608	}
	3609
	3610	// execute the command
	3611	UINT32 w1 = m_cmd_data[m_cmd_cur+0];
	3612	UINT32 w2 = m_cmd_data[m_cmd_cur+1];
	3613
	3614	switch(cmd)
	3615	{
	3616	case 0x00: CmdNoOp(w1, w2); break;
	3617
	3618	case 0x08: CmdTriangle(w1, w2); break;
	3619	case 0x09: CmdTriangleZ(w1, w2); break;
	3620	case 0x0a: CmdTriangleT(w1, w2); break;
	3621	case 0x0b: CmdTriangleTZ(w1, w2); break;
	3622	case 0x0c: CmdTriangleS(w1, w2); break;
	3623	case 0x0d: CmdTriangleSZ(w1, w2); break;
	3624	case 0x0e: CmdTriangleST(w1, w2); break;
	3625	case 0x0f: CmdTriangleSTZ(w1, w2); break;
	3626
	3627	case 0x24: CmdTexRect(w1, w2); break;
	3628	case 0x25: CmdTexRectFlip(w1, w2); break;
	3629
	3630	case 0x26: CmdSyncLoad(w1, w2); break;
	3631	case 0x27: CmdSyncPipe(w1, w2); break;
	3632	case 0x28: CmdSyncTile(w1, w2); break;
	3633	case 0x29: CmdSyncFull(w1, w2); break;
	3634
	3635	case 0x2a: CmdSetKeyGB(w1, w2); break;
	3636	case 0x2b: CmdSetKeyR(w1, w2); break;
	3637
	3638	case 0x2c: CmdSetConvert(w1, w2); break;
	3639	case 0x3c: CmdSetCombine(w1, w2); break;
	3640	case 0x2d: CmdSetScissor(w1, w2); break;
	3641	case 0x2e: CmdSetPrimDepth(w1, w2); break;
	3642	case 0x2f: CmdSetOtherModes(w1, w2); break;
	3643
	3644	case 0x30: CmdLoadTLUT(w1, w2); break;
	3645	case 0x33: CmdLoadBlock(w1, w2); break;
	3646	case 0x34: CmdLoadTile(w1, w2); break;
	3647
	3648	case 0x32: CmdSetTileSize(w1, w2); break;
	3649	case 0x35: CmdSetTile(w1, w2); break;
	3650
	3651	case 0x36: CmdFillRect(w1, w2); break;
	3652
	3653	case 0x37: CmdSetFillColor32(w1, w2); break;
	3654	case 0x38: CmdSetFogColor(w1, w2); break;
	3655	case 0x39: CmdSetBlendColor(w1, w2); break;
	3656	case 0x3a: CmdSetPrimColor(w1, w2); break;
	3657	case 0x3b: CmdSetEnvColor(w1, w2); break;
	3658
	3659	case 0x3d: CmdSetTextureImage(w1, w2); break;
	3660	case 0x3e: CmdSetMaskImage(w1, w2); break;
	3661	case 0x3f: CmdSetColorImage(w1, w2); break;
	3662	}
	3663
	3664	m_cmd_cur += rdp_command_length[cmd] / 4;
	3665	};
	3666	m_cmd_ptr = 0;
	3667	m_cmd_cur = 0;
	3668
	3669	m_start = m_current = m_end;
	3670	}
	3671
	3672	/*****************************************************************************/
	3673
	3674	n64_rdp::n64_rdp(n64_state &state) : poly_manager<UINT32, rdp_poly_state, 8, 32000>(state.machine())
	3675	{
	3676	AuxBufPtr = 0;
	3677	AuxBuf = NULL;
	3678	m_pipe_clean = true;
	3679
	3680	m_pending_mode_block = false;
	3681
	3682	m_cmd_ptr = 0;
	3683	m_cmd_cur = 0;
	3684
	3685	m_start = 0;
	3686	m_end = 0;
	3687	m_current = 0;
	3688	m_status = 0x88;
	3689
	3690	for (int i = 0; i < 8; i++)
	3691	{
	3692	m_tiles[i].num = i;
	3693	}
	3694
	3695	OneColor.c = 0xffffffff;
	3696	ZeroColor.c = 0x00000000;
	3697
	3698	m_tmem = NULL;
	3699
	3700	m_machine = NULL;
	3701
	3702	//memset(m_hidden_bits, 3, 8388608);
	3703
	3704	PrimLODFraction = 0;
	3705
	3706	for (int i = 0; i < 256; i++)
	3707	{
	3708	m_gamma_table[i] = sqrt((float)(i << 6));
	3709	m_gamma_table[i] <<= 1;
	3710	}
	3711
	3712	for (int i = 0; i < 0x4000; i++)
	3713	{
	3714	m_gamma_dither_table[i] = sqrt((float)i);
	3715	m_gamma_dither_table[i] <<= 1;
	3716	}
	3717
	3718	z_build_com_table();
	3719
	3720	for (int i = 0; i < 0x4000; i++)
	3721	{
	3722	UINT32 exponent = (i >> 11) & 7;
	3723	UINT32 mantissa = i & 0x7ff;
	3724	z_complete_dec_table[i] = ((mantissa << z_dec_table[exponent].shift) + z_dec_table[exponent].add) & 0x3fffff;
	3725	}
	3726
	3727	precalc_cvmask_derivatives();
	3728
	3729	for(int i = 0; i < 0x200; i++)
	3730	{
	3731	switch((i >> 7) & 3)
	3732	{
	3733	case 0:
	3734	case 1:
	3735	s_special_9bit_clamptable[i] = i & 0xff;
	3736	break;
	3737	case 2:
	3738	s_special_9bit_clamptable[i] = 0xff;
	3739	break;
	3740	case 3:
	3741	s_special_9bit_clamptable[i] = 0;
	3742	break;
	3743	}
	3744	}
	3745
	3746	for(int i = 0; i < 32; i++)
	3747	{
	3748	ReplicatedRGBA[i] = (i << 3) \| ((i >> 2) & 7);
	3749	}
	3750
	3751	_Write[0] = &n64_rdp::_Write16Bit_Cvg0_NoBlend;
	3752	_Write[1] = &n64_rdp::_Write16Bit_Cvg0_Blend;
	3753	_Write[2] = &n64_rdp::_Write16Bit_Cvg1;
	3754	_Write[3] = &n64_rdp::_Write16Bit_Cvg1;
	3755	_Write[4] = &n64_rdp::_Write16Bit_Cvg2;
	3756	_Write[5] = &n64_rdp::_Write16Bit_Cvg2;
	3757	_Write[6] = &n64_rdp::_Write16Bit_Cvg3;
	3758	_Write[7] = &n64_rdp::_Write16Bit_Cvg3;
	3759	_Write[8] = &n64_rdp::_Write32Bit_Cvg0_NoBlend;
	3760	_Write[9] = &n64_rdp::_Write32Bit_Cvg0_Blend;
	3761	_Write[10] = &n64_rdp::_Write32Bit_Cvg1;
	3762	_Write[11] = &n64_rdp::_Write32Bit_Cvg1;
	3763	_Write[12] = &n64_rdp::_Write32Bit_Cvg2;
	3764	_Write[13] = &n64_rdp::_Write32Bit_Cvg2;
	3765	_Write[14] = &n64_rdp::_Write32Bit_Cvg3;
	3766	_Write[15] = &n64_rdp::_Write32Bit_Cvg3;
	3767
	3768	_Read[0] = &n64_rdp::_Read16Bit_ImgRead0;
	3769	_Read[1] = &n64_rdp::_Read16Bit_ImgRead1;
	3770	_Read[2] = &n64_rdp::_Read32Bit_ImgRead0;
	3771	_Read[3] = &n64_rdp::_Read32Bit_ImgRead1;
	3772
	3773	_Copy[0] = &n64_rdp::_Copy16Bit;
	3774	_Copy[1] = &n64_rdp::_Copy32Bit;
	3775
	3776	_Fill[0] = &n64_rdp::_Fill16Bit;
	3777	_Fill[1] = &n64_rdp::_Fill32Bit;
	3778	}
	3779
	3780	void n64_state::video_start()
	3781	{
	3782	m_rdp = auto_alloc(machine(), n64_rdp(*this));
	3783
	3784	m_rdp->SetMachine(machine());
	3785	m_rdp->InitInternalState();
	3786
	3787	m_rdp->Blender.SetMachine(machine());
	3788	m_rdp->Blender.SetProcessor(m_rdp);
	3789
	3790	m_rdp->TexPipe.SetMachine(machine());
	3791
	3792	m_rdp->AuxBuf = auto_alloc_array_clear(machine(), UINT8, EXTENT_AUX_COUNT);
	3793
	3794	if (LOG_RDP_EXECUTION)
	3795	{
	3796	rdp_exec = fopen("rdp_execute.txt", "wt");
	3797	}
	3798	}
	3799
	3800	UINT32 n64_state::screen_update_n64(screen_device &screen, bitmap_rgb32 &bitmap, const rectangle &cliprect)
	3801	{
	3802	n64_periphs *n64 = machine().device<n64_periphs>("rcp");
	3803	m_rdp->visarea = screen.visible_area();
	3804
	3805	//UINT16 frame_buffer = (UINT16)&rdram[(n64->vi_origin & 0xffffff) >> 2];
	3806	//UINT8 *cvg_buffer = &m_rdp.HiddenBits[((n64->vi_origin & 0xffffff) >> 2) >> 1];
	3807	//int vibuffering = ((n64->vi_control & 2) && fsaa && divot);
	3808
	3809	//vibuffering = 0; // Disabled for now
	3810
	3811	/*
	3812	if (vibuffering && ((n64->vi_control & 3) == 2))
	3813	{
	3814	if (frame_buffer)
	3815	{
	3816	for (j=0; j < vres; j++)
	3817	{
	3818	for (i=0; i < hres; i++)
	3819	{
	3820	UINT16 pix;
	3821	pix = frame_buffer[pixels ^ WORD_ADDR_XOR];
	3822	curpixel_cvg = ((pix & 1) << 2) \| (cvg_buffer[pixels ^ BYTE_ADDR_XOR] & 3); // Reuse of this variable
	3823	if (curpixel_cvg < 7 && i > 1 && j > 1 && i < (hres - 2) && j < (vres - 2) && fsaa)
	3824	{
	3825	newc = video_filter16(&frame_buffer[pixels ^ WORD_ADDR_XOR], &cvg_buffer[pixels ^ BYTE_ADDR_XOR], n64->vi_width);
	3826	ViBuffer[i][j] = newc;
	3827	}
	3828	else
	3829	{
	3830	newc.i.r = ((pix >> 8) & 0xf8) \| (pix >> 13);
	3831	newc.i.g = ((pix >> 3) & 0xf8) \| ((pix >> 8) & 0x07);
	3832	newc.i.b = ((pix << 2) & 0xf8) \| ((pix >> 3) & 0x07);
	3833	ViBuffer[i][j] = newc;
	3834	}
	3835	pixels++;
	3836	}
	3837	pixels += invisiblewidth;
	3838	}
	3839	}
	3840	}
	3841	*/
	3842
	3843	m_rdp->wait();
	3844	m_rdp->AuxBufPtr = 0;
	3845
	3846	if (n64->vi_blank)
	3847	{
	3848	bitmap.fill(0, m_rdp->visarea);
	3849	return 0;
	3850	}
	3851
	3852	m_rdp->VideoUpdate(n64, bitmap);
	3853
	3854	return 0;
	3855	}

https://github.com/mamedev/mame/commit/b4bda080b3a71e573d65c1255feb0b9603255a6a

199869 Revisions