summaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/nv50/nv50_vbo.c
blob: 54a2fa730a5a524da8c5c91cb94b32d4b34d33a4 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
/*
 * Copyright 2008 Ben Skeggs
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
 * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF
 * OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */

#include "pipe/p_context.h"
#include "pipe/p_state.h"
#include "util/u_inlines.h"
#include "util/u_format.h"

#include "nv50_context.h"

#define NV50_USING_LOATHED_EDGEFLAG(ctx) ((ctx)->vertprog->cfg.edgeflag_in < 16)

static INLINE unsigned
nv50_prim(unsigned mode)
{
	switch (mode) {
	case PIPE_PRIM_POINTS: return NV50TCL_VERTEX_BEGIN_POINTS;
	case PIPE_PRIM_LINES: return NV50TCL_VERTEX_BEGIN_LINES;
	case PIPE_PRIM_LINE_LOOP: return NV50TCL_VERTEX_BEGIN_LINE_LOOP;
	case PIPE_PRIM_LINE_STRIP: return NV50TCL_VERTEX_BEGIN_LINE_STRIP;
	case PIPE_PRIM_TRIANGLES: return NV50TCL_VERTEX_BEGIN_TRIANGLES;
	case PIPE_PRIM_TRIANGLE_STRIP:
		return NV50TCL_VERTEX_BEGIN_TRIANGLE_STRIP;
	case PIPE_PRIM_TRIANGLE_FAN: return NV50TCL_VERTEX_BEGIN_TRIANGLE_FAN;
	case PIPE_PRIM_QUADS: return NV50TCL_VERTEX_BEGIN_QUADS;
	case PIPE_PRIM_QUAD_STRIP: return NV50TCL_VERTEX_BEGIN_QUAD_STRIP;
	case PIPE_PRIM_POLYGON: return NV50TCL_VERTEX_BEGIN_POLYGON;
	case PIPE_PRIM_LINES_ADJACENCY:
		return NV50TCL_VERTEX_BEGIN_LINES_ADJACENCY;
	case PIPE_PRIM_LINE_STRIP_ADJACENCY:
		return NV50TCL_VERTEX_BEGIN_LINE_STRIP_ADJACENCY;
	case PIPE_PRIM_TRIANGLES_ADJACENCY:
		return NV50TCL_VERTEX_BEGIN_TRIANGLES_ADJACENCY;
	case PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY:
		return NV50TCL_VERTEX_BEGIN_TRIANGLE_STRIP_ADJACENCY;
	default:
		break;
	}

	NOUVEAU_ERR("invalid primitive type %d\n", mode);
	return NV50TCL_VERTEX_BEGIN_POINTS;
}

static INLINE uint32_t
nv50_vbo_type_to_hw(enum pipe_format format)
{
	const struct util_format_description *desc;

	desc = util_format_description(format);
	assert(desc);

	switch (desc->channel[0].type) {
	case UTIL_FORMAT_TYPE_FLOAT:
		return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_FLOAT;
	case UTIL_FORMAT_TYPE_UNSIGNED:
		if (desc->channel[0].normalized) {
			return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_UNORM;
		}
		return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_USCALED;
	case UTIL_FORMAT_TYPE_SIGNED:
		if (desc->channel[0].normalized) {
			return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_SNORM;
		}
		return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_SSCALED;
	/*
	case PIPE_FORMAT_TYPE_UINT:
		return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_UINT;
	case PIPE_FORMAT_TYPE_SINT:
		return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_SINT; */
	default:
		return 0;
	}
}

static INLINE uint32_t
nv50_vbo_size_to_hw(unsigned size, unsigned nr_c)
{
	static const uint32_t hw_values[] = {
		0, 0, 0, 0,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_8,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_8_8,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_8_8_8,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_8_8_8_8,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_16,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_16_16,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_16_16_16,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_16_16_16_16,
		0, 0, 0, 0,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_32,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_32_32,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_32_32_32,
		NV50TCL_VERTEX_ARRAY_ATTRIB_FORMAT_32_32_32_32 };

	/* we'd also have R11G11B10 and R10G10B10A2 */

	assert(nr_c > 0 && nr_c <= 4);

	if (size > 32)
		return 0;
	size >>= (3 - 2);

	return hw_values[size + (nr_c - 1)];
}

static INLINE uint32_t
nv50_vbo_vtxelt_to_hw(struct pipe_vertex_element *ve)
{
	uint32_t hw_type, hw_size;
	enum pipe_format pf = ve->src_format;
	const struct util_format_description *desc;
	unsigned size, nr_components;

	desc = util_format_description(pf);
	assert(desc);

	size = util_format_get_component_bits(pf, UTIL_FORMAT_COLORSPACE_RGB, 0);
	nr_components = util_format_get_nr_components(pf);

	hw_type = nv50_vbo_type_to_hw(pf);
	hw_size = nv50_vbo_size_to_hw(size, nr_components);

	if (!hw_type || !hw_size) {
		NOUVEAU_ERR("unsupported vbo format: %s\n", util_format_name(pf));
		abort();
		return 0x24e80000;
	}

	if (desc->swizzle[0] == UTIL_FORMAT_SWIZZLE_Z) /* BGRA */
		hw_size |= (1 << 31); /* no real swizzle bits :-( */

	return (hw_type | hw_size);
}

struct instance {
	struct nouveau_bo *bo;
	unsigned delta;
	unsigned stride;
	unsigned step;
	unsigned divisor;
};

static void
instance_init(struct nv50_context *nv50, struct instance *a, unsigned first)
{
	int i;

	for (i = 0; i < nv50->vtxelt->num_elements; i++) {
		struct pipe_vertex_element *ve = &nv50->vtxelt->pipe[i];
		struct pipe_vertex_buffer *vb;

		a[i].divisor = ve->instance_divisor;
		if (a[i].divisor) {
			vb = &nv50->vtxbuf[ve->vertex_buffer_index];

			a[i].bo = nouveau_bo(vb->buffer);
			a[i].stride = vb->stride;
			a[i].step = first % a[i].divisor;
			a[i].delta = vb->buffer_offset + ve->src_offset +
				     (first * a[i].stride);
		}
	}
}

static void
instance_step(struct nv50_context *nv50, struct instance *a)
{
	struct nouveau_channel *chan = nv50->screen->tesla->channel;
	struct nouveau_grobj *tesla = nv50->screen->tesla;
	int i;

	for (i = 0; i < nv50->vtxelt->num_elements; i++) {
		if (!a[i].divisor)
			continue;

		BEGIN_RING(chan, tesla,
			   NV50TCL_VERTEX_ARRAY_START_HIGH(i), 2);
		OUT_RELOCh(chan, a[i].bo, a[i].delta, NOUVEAU_BO_RD |
			   NOUVEAU_BO_VRAM | NOUVEAU_BO_GART);
		OUT_RELOCl(chan, a[i].bo, a[i].delta, NOUVEAU_BO_RD |
			   NOUVEAU_BO_VRAM | NOUVEAU_BO_GART);
		if (++a[i].step == a[i].divisor) {
			a[i].step = 0;
			a[i].delta += a[i].stride;
		}
	}
}

void
nv50_draw_arrays_instanced(struct pipe_context *pipe,
			   unsigned mode, unsigned start, unsigned count,
			   unsigned startInstance, unsigned instanceCount)
{
	struct nv50_context *nv50 = nv50_context(pipe);
	struct nouveau_channel *chan = nv50->screen->tesla->channel;
	struct nouveau_grobj *tesla = nv50->screen->tesla;
	struct instance a[16];
	unsigned prim = nv50_prim(mode);

	instance_init(nv50, a, startInstance);
	if (!nv50_state_validate(nv50, 10 + 16*3))
		return;

	BEGIN_RING(chan, tesla, NV50TCL_CB_ADDR, 2);
	OUT_RING  (chan, NV50_CB_AUX | (24 << 8));
	OUT_RING  (chan, startInstance);
	while (instanceCount--) {
		if (AVAIL_RING(chan) < (7 + 16*3)) {
			FIRE_RING(chan);
			if (!nv50_state_validate(nv50, 7 + 16*3)) {
				assert(0);
				return;
			}
		}
		instance_step(nv50, a);

		BEGIN_RING(chan, tesla, NV50TCL_VERTEX_BEGIN, 1);
		OUT_RING  (chan, prim);
		BEGIN_RING(chan, tesla, NV50TCL_VERTEX_BUFFER_FIRST, 2);
		OUT_RING  (chan, start);
		OUT_RING  (chan, count);
		BEGIN_RING(chan, tesla, NV50TCL_VERTEX_END, 1);
		OUT_RING  (chan, 0);

		prim |= (1 << 28);
	}
}

void
nv50_draw_arrays(struct pipe_context *pipe, unsigned mode, unsigned start,
		 unsigned count)
{
	nv50_draw_arrays_instanced(pipe, mode, start, count, 0, 1);
}

static INLINE boolean
nv50_draw_elements_inline_u08(struct nv50_context *nv50, uint8_t *map,
			      unsigned start, unsigned count)
{
	struct nouveau_channel *chan = nv50->screen->tesla->channel;
	struct nouveau_grobj *tesla = nv50->screen->tesla;

	map += start;

	if (count & 1) {
		BEGIN_RING(chan, tesla, NV50TCL_VB_ELEMENT_U32, 1);
		OUT_RING  (chan, map[0]);
		map++;
		count--;
	}

	while (count) {
		unsigned nr = count > 2046 ? 2046 : count;
		int i;

		BEGIN_RING_NI(chan, tesla, NV50TCL_VB_ELEMENT_U16, nr >> 1);
		for (i = 0; i < nr; i += 2)
			OUT_RING  (chan, (map[i + 1] << 16) | map[i]);

		count -= nr;
		map += nr;
	}
	return TRUE;
}

static INLINE boolean
nv50_draw_elements_inline_u16(struct nv50_context *nv50, uint16_t *map,
			      unsigned start, unsigned count)
{
	struct nouveau_channel *chan = nv50->screen->tesla->channel;
	struct nouveau_grobj *tesla = nv50->screen->tesla;

	map += start;

	if (count & 1) {
		BEGIN_RING(chan, tesla, NV50TCL_VB_ELEMENT_U32, 1);
		OUT_RING  (chan, map[0]);
		map++;
		count--;
	}

	while (count) {
		unsigned nr = count > 2046 ? 2046 : count;
		int i;

		BEGIN_RING_NI(chan, tesla, NV50TCL_VB_ELEMENT_U16, nr >> 1);
		for (i = 0; i < nr; i += 2)
			OUT_RING  (chan, (map[i + 1] << 16) | map[i]);

		count -= nr;
		map += nr;
	}
	return TRUE;
}

static INLINE boolean
nv50_draw_elements_inline_u32(struct nv50_context *nv50, uint32_t *map,
			      unsigned start, unsigned count)
{
	struct nouveau_channel *chan = nv50->screen->tesla->channel;
	struct nouveau_grobj *tesla = nv50->screen->tesla;

	map += start;

	while (count) {
		unsigned nr = count > 2047 ? 2047 : count;

		BEGIN_RING_NI(chan, tesla, NV50TCL_VB_ELEMENT_U32, nr);
		OUT_RINGp (chan, map, nr);

		count -= nr;
		map += nr;
	}
	return TRUE;
}

static void
nv50_draw_elements_inline(struct pipe_context *pipe,
			  struct pipe_buffer *indexBuffer, unsigned indexSize,
			  unsigned mode, unsigned start, unsigned count,
			  unsigned startInstance, unsigned instanceCount)
{
	struct pipe_screen *pscreen = pipe->screen;
	struct nv50_context *nv50 = nv50_context(pipe);
	struct nouveau_channel *chan = nv50->screen->tesla->channel;
	struct nouveau_grobj *tesla = nv50->screen->tesla;
	struct instance a[16];
	unsigned prim = nv50_prim(mode);
	void *map;

	map = pipe_buffer_map(pscreen, indexBuffer, PIPE_BUFFER_USAGE_CPU_READ);
	assert(map);
	if (!map)
		return;

	instance_init(nv50, a, startInstance);
	if (!nv50_state_validate(nv50, 0))
		return;

	BEGIN_RING(chan, tesla, NV50TCL_CB_ADDR, 2);
	OUT_RING  (chan, NV50_CB_AUX | (24 << 8));
	OUT_RING  (chan, startInstance);
	while (instanceCount--) {
		if (AVAIL_RING(chan) < (7 + 16*3)) {
			FIRE_RING(chan);
			if (!nv50_state_validate(nv50, 0)) {
				assert(0);
				return;
			}
		}
		instance_step(nv50, a);

		BEGIN_RING(chan, tesla, NV50TCL_VERTEX_BEGIN, 1);
		OUT_RING  (chan, prim);
		switch (indexSize) {
		case 1:
			nv50_draw_elements_inline_u08(nv50, map, start, count);
			break;
		case 2:
			nv50_draw_elements_inline_u16(nv50, map, start, count);
			break;
		case 4:
			nv50_draw_elements_inline_u32(nv50, map, start, count);
			break;
		default:
			assert(0);
			break;
		}
		BEGIN_RING(chan, tesla, NV50TCL_VERTEX_END, 1);
		OUT_RING  (chan, 0);

		prim |= (1 << 28);
	}

	pipe_buffer_unmap(pscreen, indexBuffer);
}

void
nv50_draw_elements_instanced(struct pipe_context *pipe,
			     struct pipe_buffer *indexBuffer,
			     unsigned indexSize,
			     unsigned mode, unsigned start, unsigned count,
			     unsigned startInstance, unsigned instanceCount)
{
	struct nv50_context *nv50 = nv50_context(pipe);
	struct nouveau_channel *chan = nv50->screen->tesla->channel;
	struct nouveau_grobj *tesla = nv50->screen->tesla;
	struct instance a[16];
	unsigned prim = nv50_prim(mode);

	if (indexSize == 1) {
		nv50_draw_elements_inline(pipe, indexBuffer, indexSize,
					  mode, start, count, startInstance,
					  instanceCount);
		return;
	}

	instance_init(nv50, a, startInstance);
	if (!nv50_state_validate(nv50, 13 + 16*3))
		return;

	BEGIN_RING(chan, tesla, NV50TCL_CB_ADDR, 2);
	OUT_RING  (chan, NV50_CB_AUX | (24 << 8));
	OUT_RING  (chan, startInstance);
	while (instanceCount--) {
		if (AVAIL_RING(chan) < (7 + 16*3)) {
			FIRE_RING(chan);
			if (!nv50_state_validate(nv50, 10 + 16*3)) {
				assert(0);
				return;
			}
		}
		instance_step(nv50, a);

		BEGIN_RING(chan, tesla, NV50TCL_VERTEX_BEGIN, 1);
		OUT_RING  (chan, prim);
		if (indexSize == 4) {
			BEGIN_RING(chan, tesla, NV50TCL_VB_ELEMENT_U32 | 0x30000, 0);
			OUT_RING  (chan, count);
			nouveau_pushbuf_submit(chan, nouveau_bo(indexBuffer),
					       start << 2, count << 2);
		} else
		if (indexSize == 2) {
			unsigned vb_start = (start & ~1);
			unsigned vb_end = (start + count + 1) & ~1;
			unsigned dwords = (vb_end - vb_start) >> 1;

			BEGIN_RING(chan, tesla, NV50TCL_VB_ELEMENT_U16_SETUP, 1);
			OUT_RING  (chan, ((start & 1) << 31) | count);
			BEGIN_RING(chan, tesla, NV50TCL_VB_ELEMENT_U16 | 0x30000, 0);
			OUT_RING  (chan, dwords);
			nouveau_pushbuf_submit(chan, nouveau_bo(indexBuffer),
					       vb_start << 1, dwords << 2);
			BEGIN_RING(chan, tesla, NV50TCL_VB_ELEMENT_U16_SETUP, 1);
			OUT_RING  (chan, 0);
		}
		BEGIN_RING(chan, tesla, NV50TCL_VERTEX_END, 1);
		OUT_RING  (chan, 0);

		prim |= (1 << 28);
	}
}

void
nv50_draw_elements(struct pipe_context *pipe,
		   struct pipe_buffer *indexBuffer, unsigned indexSize,
		   unsigned mode, unsigned start, unsigned count)
{
	nv50_draw_elements_instanced(pipe, indexBuffer, indexSize,
				     mode, start, count, 0, 1);
}

static INLINE boolean
nv50_vbo_static_attrib(struct nv50_context *nv50, unsigned attrib,
		       struct nouveau_stateobj **pso,
		       struct pipe_vertex_element *ve,
		       struct pipe_vertex_buffer *vb)

{
	struct nouveau_stateobj *so;
	struct nouveau_grobj *tesla = nv50->screen->tesla;
	struct nouveau_bo *bo = nouveau_bo(vb->buffer);
	float v[4];
	int ret;
	unsigned nr_components = util_format_get_nr_components(ve->src_format);

	ret = nouveau_bo_map(bo, NOUVEAU_BO_RD);
	if (ret)
		return FALSE;

	util_format_read_4f(ve->src_format, v, 0, (uint8_t *)bo->map +
			    (vb->buffer_offset + ve->src_offset), 0,
			    0, 0, 1, 1);
	so = *pso;
	if (!so)
		*pso = so = so_new(nv50->vtxelt->num_elements,
				   nv50->vtxelt->num_elements * 4, 0);

	switch (nr_components) {
	case 4:
		so_method(so, tesla, NV50TCL_VTX_ATTR_4F_X(attrib), 4);
		so_data  (so, fui(v[0]));
		so_data  (so, fui(v[1]));
		so_data  (so, fui(v[2]));
		so_data  (so, fui(v[3]));
		break;
	case 3:
		so_method(so, tesla, NV50TCL_VTX_ATTR_3F_X(attrib), 3);
		so_data  (so, fui(v[0]));
		so_data  (so, fui(v[1]));
		so_data  (so, fui(v[2]));
		break;
	case 2:
		so_method(so, tesla, NV50TCL_VTX_ATTR_2F_X(attrib), 2);
		so_data  (so, fui(v[0]));
		so_data  (so, fui(v[1]));
		break;
	case 1:
		if (attrib == nv50->vertprog->cfg.edgeflag_in) {
			so_method(so, tesla, NV50TCL_EDGEFLAG_ENABLE, 1);
			so_data  (so, v[0] ? 1 : 0);
		}
		so_method(so, tesla, NV50TCL_VTX_ATTR_1F(attrib), 1);
		so_data  (so, fui(v[0]));
		break;
	default:
		nouveau_bo_unmap(bo);
		return FALSE;
	}

	nouveau_bo_unmap(bo);
	return TRUE;
}

void
nv50_vtxelt_construct(struct nv50_vtxelt_stateobj *cso)
{
	unsigned i;

	for (i = 0; i < cso->num_elements; ++i) {
		struct pipe_vertex_element *ve = &cso->pipe[i];

		cso->hw[i] = nv50_vbo_vtxelt_to_hw(ve);
	}
}

struct nouveau_stateobj *
nv50_vbo_validate(struct nv50_context *nv50)
{
	struct nouveau_grobj *tesla = nv50->screen->tesla;
	struct nouveau_stateobj *vtxbuf, *vtxfmt, *vtxattr;
	unsigned i, n_ve;

	/* don't validate if Gallium took away our buffers */
	if (nv50->vtxbuf_nr == 0)
		return NULL;

	assert(!NV50_USING_LOATHED_EDGEFLAG(nv50));

	n_ve = MAX2(nv50->vtxelt->num_elements, nv50->state.vtxelt_nr);

	vtxattr = NULL;
	vtxbuf = so_new(n_ve * 2, n_ve * 5, nv50->vtxelt->num_elements * 4);
	vtxfmt = so_new(1, n_ve, 0);
	so_method(vtxfmt, tesla, NV50TCL_VERTEX_ARRAY_ATTRIB(0), n_ve);

	for (i = 0; i < nv50->vtxelt->num_elements; i++) {
		struct pipe_vertex_element *ve = &nv50->vtxelt->pipe[i];
		struct pipe_vertex_buffer *vb =
			&nv50->vtxbuf[ve->vertex_buffer_index];
		struct nouveau_bo *bo = nouveau_bo(vb->buffer);
		uint32_t hw = nv50->vtxelt->hw[i];

		if (!vb->stride &&
		    nv50_vbo_static_attrib(nv50, i, &vtxattr, ve, vb)) {
			so_data(vtxfmt, hw | (1 << 4));

			so_method(vtxbuf, tesla,
				  NV50TCL_VERTEX_ARRAY_FORMAT(i), 1);
			so_data  (vtxbuf, 0);
			continue;
		}

		so_data(vtxfmt, hw | i);

		so_method(vtxbuf, tesla, NV50TCL_VERTEX_ARRAY_FORMAT(i), 3);
		so_data  (vtxbuf, 0x20000000 |
			  (ve->instance_divisor ? 0 : vb->stride));
		so_reloc (vtxbuf, bo, vb->buffer_offset +
			  ve->src_offset, NOUVEAU_BO_VRAM | NOUVEAU_BO_GART |
			  NOUVEAU_BO_RD | NOUVEAU_BO_HIGH, 0, 0);
		so_reloc (vtxbuf, bo, vb->buffer_offset +
			  ve->src_offset, NOUVEAU_BO_VRAM | NOUVEAU_BO_GART |
			  NOUVEAU_BO_RD | NOUVEAU_BO_LOW, 0, 0);

		/* vertex array limits */
		so_method(vtxbuf, tesla, NV50TCL_VERTEX_ARRAY_LIMIT_HIGH(i), 2);
		so_reloc (vtxbuf, bo, vb->buffer->size - 1,
			  NOUVEAU_BO_VRAM | NOUVEAU_BO_GART | NOUVEAU_BO_RD |
			  NOUVEAU_BO_HIGH, 0, 0);
		so_reloc (vtxbuf, bo, vb->buffer->size - 1,
			  NOUVEAU_BO_VRAM | NOUVEAU_BO_GART | NOUVEAU_BO_RD |
			  NOUVEAU_BO_LOW, 0, 0);
	}
	for (; i < n_ve; ++i) {
		so_data  (vtxfmt, 0x7e080010);

		so_method(vtxbuf, tesla, NV50TCL_VERTEX_ARRAY_FORMAT(i), 1);
		so_data  (vtxbuf, 0);
	}
	nv50->state.vtxelt_nr = nv50->vtxelt->num_elements;

	so_ref (vtxbuf, &nv50->state.vtxbuf);
	so_ref (vtxattr, &nv50->state.vtxattr);
	so_ref (NULL, &vtxbuf);
	so_ref (NULL, &vtxattr);
	return vtxfmt;
}