summaryrefslogtreecommitdiff
path: root/src/gallium/drivers/nv40/nv40_draw.c
blob: a9a939af0c02e4332f102169aabafc5b86bc871d (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
#include "pipe/p_util.h"
#include "pipe/p_shader_tokens.h"

#include "draw/draw_context.h"
#include "draw/draw_vertex.h"
#include "draw/draw_pipe.h"

#include "nv40_context.h"
#define NV40_SHADER_NO_FUCKEDNESS
#include "nv40_shader.h"

/* Simple, but crappy, swtnl path, hopefully we wont need to hit this very
 * often at all.  Uses "quadro style" vertex submission + a fixed vertex
 * layout to avoid the need to generate a vertex program or vtxfmt.
 */

struct nv40_render_stage {
	struct draw_stage stage;
	struct nv40_context *nv40;
	unsigned prim;
};

static INLINE struct nv40_render_stage *
nv40_render_stage(struct draw_stage *stage)
{
	return (struct nv40_render_stage *)stage;
}

static INLINE void
nv40_render_vertex(struct nv40_context *nv40, const struct vertex_header *v)
{
	unsigned i;

	for (i = 0; i < nv40->swtnl.nr_attribs; i++) {
		unsigned idx = nv40->swtnl.draw[i];
		unsigned hw = nv40->swtnl.hw[i];

		switch (nv40->swtnl.emit[i]) {
		case EMIT_OMIT:
			break;
		case EMIT_1F:
			BEGIN_RING(curie, 0x1e40 + (hw * 4), 1);
			OUT_RING  (fui(v->data[idx][0]));
			break;
		case EMIT_2F:
			BEGIN_RING(curie, NV40TCL_VTX_ATTR_2F_X(hw), 2);
			OUT_RING  (fui(v->data[idx][0]));
			OUT_RING  (fui(v->data[idx][1]));
			break;
		case EMIT_3F:
			BEGIN_RING(curie, NV40TCL_VTX_ATTR_3F_X(hw), 3);
			OUT_RING  (fui(v->data[idx][0]));
			OUT_RING  (fui(v->data[idx][1]));
			OUT_RING  (fui(v->data[idx][2]));
			break;
		case EMIT_4F:
			BEGIN_RING(curie, NV40TCL_VTX_ATTR_4F_X(hw), 4);
			OUT_RING  (fui(v->data[idx][0]));
			OUT_RING  (fui(v->data[idx][1]));
			OUT_RING  (fui(v->data[idx][2]));
			OUT_RING  (fui(v->data[idx][3]));
			break;
		case EMIT_4UB:
			BEGIN_RING(curie, 0x1940 + (hw * 4), 1);
			OUT_RING  (pack_ub4(float_to_ubyte(v->data[idx][0]),
					    float_to_ubyte(v->data[idx][1]),
					    float_to_ubyte(v->data[idx][2]),
					    float_to_ubyte(v->data[idx][3])));
			break;
		default:
			assert(0);
			break;
		}
	}
}

static INLINE void
nv40_render_prim(struct draw_stage *stage, struct prim_header *prim,
	       unsigned mode, unsigned count)
{
	struct nv40_render_stage *rs = nv40_render_stage(stage);
	struct nv40_context *nv40 = rs->nv40;
	struct nouveau_pushbuf *pb = nv40->nvws->channel->pushbuf;
	unsigned i;

	/* Ensure there's room for 4xfloat32 + potentially 3 begin/end */
	if (pb->remaining < ((count * 20) + 6)) {
		if (rs->prim != NV40TCL_BEGIN_END_STOP) {
			NOUVEAU_ERR("AIII, missed flush\n");
			assert(0);
		}
		FIRE_RING(NULL);
		nv40_state_emit(nv40);
	}

	/* Switch primitive modes if necessary */
	if (rs->prim != mode) {
		if (rs->prim != NV40TCL_BEGIN_END_STOP) {
			BEGIN_RING(curie, NV40TCL_BEGIN_END, 1);
			OUT_RING  (NV40TCL_BEGIN_END_STOP);	
		}

		BEGIN_RING(curie, NV40TCL_BEGIN_END, 1);
		OUT_RING  (mode);
		rs->prim = mode;
	}

	/* Emit vertex data */
	for (i = 0; i < count; i++)
		nv40_render_vertex(nv40, prim->v[i]);

	/* If it's likely we'll need to empty the push buffer soon, finish
	 * off the primitive now.
	 */
	if (pb->remaining < ((count * 20) + 6)) {
		BEGIN_RING(curie, NV40TCL_BEGIN_END, 1);
		OUT_RING  (NV40TCL_BEGIN_END_STOP);
		rs->prim = NV40TCL_BEGIN_END_STOP;
	}
}

static void
nv40_render_point(struct draw_stage *draw, struct prim_header *prim)
{
	nv40_render_prim(draw, prim, NV40TCL_BEGIN_END_POINTS, 1);
}

static void
nv40_render_line(struct draw_stage *draw, struct prim_header *prim)
{
	nv40_render_prim(draw, prim, NV40TCL_BEGIN_END_LINES, 2);
}

static void
nv40_render_tri(struct draw_stage *draw, struct prim_header *prim)
{
	nv40_render_prim(draw, prim, NV40TCL_BEGIN_END_TRIANGLES, 3);
}

static void
nv40_render_flush(struct draw_stage *draw, unsigned flags)
{
	struct nv40_render_stage *rs = nv40_render_stage(draw);
	struct nv40_context *nv40 = rs->nv40;

	if (rs->prim != NV40TCL_BEGIN_END_STOP) {
		BEGIN_RING(curie, NV40TCL_BEGIN_END, 1);
		OUT_RING  (NV40TCL_BEGIN_END_STOP);
		rs->prim = NV40TCL_BEGIN_END_STOP;
	}
}

static void
nv40_render_reset_stipple_counter(struct draw_stage *draw)
{
}

static void
nv40_render_destroy(struct draw_stage *draw)
{
	FREE(draw);
}

static INLINE void
emit_mov(struct nv40_vertex_program *vp,
	 unsigned dst, unsigned src, unsigned vor, unsigned mask)
{
	struct nv40_vertex_program_exec *inst;

	vp->insns = realloc(vp->insns,
			    sizeof(struct nv40_vertex_program_exec) *
			    ++vp->nr_insns);
	inst = &vp->insns[vp->nr_insns - 1];

	inst->data[0] = 0x401f9c6c;
	inst->data[1] = 0x0040000d | (src << 8);
	inst->data[2] = 0x8106c083;
	inst->data[3] = 0x6041ff80 | (dst << 2) | (mask << 13);
	inst->const_index = -1;
	inst->has_branch_offset = FALSE;

	vp->ir |= (1 << src);
	if (vor != ~0)
		vp->or |= (1 << vor);
}

static struct nv40_vertex_program *
create_drawvp(struct nv40_context *nv40)
{
	struct nv40_vertex_program *vp = CALLOC_STRUCT(nv40_vertex_program);
	unsigned i;

	emit_mov(vp, NV40_VP_INST_DEST_POS, 0, ~0, 0xf);
	emit_mov(vp, NV40_VP_INST_DEST_COL0, 3, 0, 0xf);
	emit_mov(vp, NV40_VP_INST_DEST_COL1, 4, 1, 0xf);
	emit_mov(vp, NV40_VP_INST_DEST_BFC0, 3, 2, 0xf);
	emit_mov(vp, NV40_VP_INST_DEST_BFC1, 4, 3, 0xf);
	emit_mov(vp, NV40_VP_INST_DEST_FOGC, 5, 4, 0x8);
	for (i = 0; i < 8; i++)
		emit_mov(vp, NV40_VP_INST_DEST_TC(i), 8 + i, 14 + i, 0xf);

	vp->insns[vp->nr_insns - 1].data[3] |= 1;
	vp->translated = TRUE;
	return vp;
}

struct draw_stage *
nv40_draw_render_stage(struct nv40_context *nv40)
{
	struct nv40_render_stage *render = CALLOC_STRUCT(nv40_render_stage);

	if (!nv40->swtnl.vertprog)
		nv40->swtnl.vertprog = create_drawvp(nv40);

	render->nv40 = nv40;
	render->stage.draw = nv40->draw;
	render->stage.point = nv40_render_point;
	render->stage.line = nv40_render_line;
	render->stage.tri = nv40_render_tri;
	render->stage.flush = nv40_render_flush;
	render->stage.reset_stipple_counter = nv40_render_reset_stipple_counter;
	render->stage.destroy = nv40_render_destroy;

	return &render->stage;
}

boolean
nv40_draw_elements_swtnl(struct pipe_context *pipe,
			 struct pipe_buffer *idxbuf, unsigned idxbuf_size,
			 unsigned mode, unsigned start, unsigned count)
{
	struct nv40_context *nv40 = nv40_context(pipe);
	struct pipe_winsys *ws = pipe->winsys;
	unsigned i;
	void *map;

	if (!nv40_state_validate_swtnl(nv40))
		return FALSE;
	nv40->state.dirty &= ~(1ULL << NV40_STATE_VTXBUF);
	nv40_state_emit(nv40);

	for (i = 0; i < nv40->vtxbuf_nr; i++) {
		map = ws->buffer_map(ws, nv40->vtxbuf[i].buffer,
				     PIPE_BUFFER_USAGE_CPU_READ);
		draw_set_mapped_vertex_buffer(nv40->draw, i, map);
	}

	if (idxbuf) {
		map = ws->buffer_map(ws, idxbuf, PIPE_BUFFER_USAGE_CPU_READ);
		draw_set_mapped_element_buffer(nv40->draw, idxbuf_size, map);
	} else {
		draw_set_mapped_element_buffer(nv40->draw, 0, NULL);
	}

	if (nv40->constbuf[PIPE_SHADER_VERTEX]) {
		map = ws->buffer_map(ws, nv40->constbuf[PIPE_SHADER_VERTEX],
				     PIPE_BUFFER_USAGE_CPU_READ);
		draw_set_mapped_constant_buffer(nv40->draw, map);
	}

	draw_arrays(nv40->draw, mode, start, count);

	for (i = 0; i < nv40->vtxbuf_nr; i++)
		ws->buffer_unmap(ws, nv40->vtxbuf[i].buffer);

	if (idxbuf)
		ws->buffer_unmap(ws, idxbuf);

	if (nv40->constbuf[PIPE_SHADER_VERTEX])
		ws->buffer_unmap(ws, nv40->constbuf[PIPE_SHADER_VERTEX]);

	draw_flush(nv40->draw);
	pipe->flush(pipe, 0, NULL);

	return TRUE;
}

static INLINE void
emit_attrib(struct nv40_context *nv40, unsigned hw, unsigned emit,
	    unsigned semantic, unsigned index)
{
	unsigned draw_out = draw_find_vs_output(nv40->draw, semantic, index);
	unsigned a = nv40->swtnl.nr_attribs++;

	nv40->swtnl.hw[a] = hw;
	nv40->swtnl.emit[a] = emit;
	nv40->swtnl.draw[a] = draw_out;
}

static boolean
nv40_state_vtxfmt_validate(struct nv40_context *nv40)
{
	struct nv40_fragment_program *fp = nv40->fragprog;
	unsigned colour = 0, texcoords = 0, fog = 0, i;

	/* Determine needed fragprog inputs */
	for (i = 0; i < fp->info.num_inputs; i++) {
		switch (fp->info.input_semantic_name[i]) {
		case TGSI_SEMANTIC_POSITION:
			break;
		case TGSI_SEMANTIC_COLOR:
			colour |= (1 << fp->info.input_semantic_index[i]);
			break;
		case TGSI_SEMANTIC_GENERIC:
			texcoords |= (1 << fp->info.input_semantic_index[i]);
			break;
		case TGSI_SEMANTIC_FOG:
			fog = 1;
			break;
		default:
			assert(0);
		}
	}

	nv40->swtnl.nr_attribs = 0;

	/* Map draw vtxprog output to hw attribute IDs */
	for (i = 0; i < 2; i++) {
		if (!(colour & (1 << i)))
			continue;
		emit_attrib(nv40, 3 + i, EMIT_4UB, TGSI_SEMANTIC_COLOR, i);
	}

	for (i = 0; i < 8; i++) {
		if (!(texcoords & (1 << i)))
			continue;
		emit_attrib(nv40, 8 + i, EMIT_4F, TGSI_SEMANTIC_GENERIC, i);
	}

	if (fog) {
		emit_attrib(nv40, 5, EMIT_1F, TGSI_SEMANTIC_FOG, 0);
	}

	emit_attrib(nv40, 0, EMIT_3F, TGSI_SEMANTIC_POSITION, 0);

	return FALSE;
}

struct nv40_state_entry nv40_state_vtxfmt = {
	.validate = nv40_state_vtxfmt_validate,
	.dirty = {
		.pipe = NV40_NEW_ARRAYS | NV40_NEW_FRAGPROG,
		.hw = 0
	}
};