aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/nouveau/nv40_grctx.c
blob: 11b11c31f54388e9bc51ba36c77c65e874d632ed (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
/*
 * Copyright 2009 Red Hat Inc.
 *
 * 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 COPYRIGHT HOLDER(S) OR AUTHOR(S) 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.
 *
 * Authors: Ben Skeggs
 */

/* NVIDIA context programs handle a number of other conditions which are
 * not implemented in our versions.  It's not clear why NVIDIA context
 * programs have this code, nor whether it's strictly necessary for
 * correct operation.  We'll implement additional handling if/when we
 * discover it's necessary.
 *
 * - On context save, NVIDIA set 0x400314 bit 0 to 1 if the "3D state"
 *   flag is set, this gets saved into the context.
 * - On context save, the context program for all cards load nsource
 *   into a flag register and check for ILLEGAL_MTHD.  If it's set,
 *   opcode 0x60000d is called before resuming normal operation.
 * - Some context programs check more conditions than the above.  NV44
 *   checks: ((nsource & 0x0857) || (0x400718 & 0x0100) || (intr & 0x0001))
 *   and calls 0x60000d before resuming normal operation.
 * - At the very beginning of NVIDIA's context programs, flag 9 is checked
 *   and if true 0x800001 is called with count=0, pos=0, the flag is cleared
 *   and then the ctxprog is aborted.  It looks like a complicated NOP,
 *   its purpose is unknown.
 * - In the section of code that loads the per-vs state, NVIDIA check
 *   flag 10.  If it's set, they only transfer the small 0x300 byte block
 *   of state + the state for a single vs as opposed to the state for
 *   all vs units.  It doesn't seem likely that it'll occur in normal
 *   operation, especially seeing as it appears NVIDIA may have screwed
 *   up the ctxprogs for some cards and have an invalid instruction
 *   rather than a cp_lsr(ctx, dwords_for_1_vs_unit) instruction.
 * - There's a number of places where context offset 0 (where we place
 *   the PRAMIN offset of the context) is loaded into either 0x408000,
 *   0x408004 or 0x408008.  Not sure what's up there either.
 * - The ctxprogs for some cards save 0x400a00 again during the cleanup
 *   path for auto-loadctx.
 */

#define CP_FLAG_CLEAR                 0
#define CP_FLAG_SET                   1
#define CP_FLAG_SWAP_DIRECTION        ((0 * 32) + 0)
#define CP_FLAG_SWAP_DIRECTION_LOAD   0
#define CP_FLAG_SWAP_DIRECTION_SAVE   1
#define CP_FLAG_USER_SAVE             ((0 * 32) + 5)
#define CP_FLAG_USER_SAVE_NOT_PENDING 0
#define CP_FLAG_USER_SAVE_PENDING     1
#define CP_FLAG_USER_LOAD             ((0 * 32) + 6)
#define CP_FLAG_USER_LOAD_NOT_PENDING 0
#define CP_FLAG_USER_LOAD_PENDING     1
#define CP_FLAG_STATUS                ((3 * 32) + 0)
#define CP_FLAG_STATUS_IDLE           0
#define CP_FLAG_STATUS_BUSY           1
#define CP_FLAG_AUTO_SAVE             ((3 * 32) + 4)
#define CP_FLAG_AUTO_SAVE_NOT_PENDING 0
#define CP_FLAG_AUTO_SAVE_PENDING     1
#define CP_FLAG_AUTO_LOAD             ((3 * 32) + 5)
#define CP_FLAG_AUTO_LOAD_NOT_PENDING 0
#define CP_FLAG_AUTO_LOAD_PENDING     1
#define CP_FLAG_UNK54                 ((3 * 32) + 6)
#define CP_FLAG_UNK54_CLEAR           0
#define CP_FLAG_UNK54_SET             1
#define CP_FLAG_ALWAYS                ((3 * 32) + 8)
#define CP_FLAG_ALWAYS_FALSE          0
#define CP_FLAG_ALWAYS_TRUE           1
#define CP_FLAG_UNK57                 ((3 * 32) + 9)
#define CP_FLAG_UNK57_CLEAR           0
#define CP_FLAG_UNK57_SET             1

#define CP_CTX                   0x00100000
#define CP_CTX_COUNT             0x000fc000
#define CP_CTX_COUNT_SHIFT               14
#define CP_CTX_REG               0x00003fff
#define CP_LOAD_SR               0x00200000
#define CP_LOAD_SR_VALUE         0x000fffff
#define CP_BRA                   0x00400000
#define CP_BRA_IP                0x0000ff00
#define CP_BRA_IP_SHIFT                   8
#define CP_BRA_IF_CLEAR          0x00000080
#define CP_BRA_FLAG              0x0000007f
#define CP_WAIT                  0x00500000
#define CP_WAIT_SET              0x00000080
#define CP_WAIT_FLAG             0x0000007f
#define CP_SET                   0x00700000
#define CP_SET_1                 0x00000080
#define CP_SET_FLAG              0x0000007f
#define CP_NEXT_TO_SWAP          0x00600007
#define CP_NEXT_TO_CURRENT       0x00600009
#define CP_SET_CONTEXT_POINTER   0x0060000a
#define CP_END                   0x0060000e
#define CP_LOAD_MAGIC_UNK01      0x00800001 /* unknown */
#define CP_LOAD_MAGIC_NV44TCL    0x00800029 /* per-vs state (0x4497) */
#define CP_LOAD_MAGIC_NV40TCL    0x00800041 /* per-vs state (0x4097) */

#include "drmP.h"
#include "nouveau_drv.h"
#include "nouveau_grctx.h"

/* TODO:
 *  - get vs count from 0x1540
 *  - document unimplemented bits compared to nvidia
 *    - nsource handling
 *    - R0 & 0x0200 handling
 *    - single-vs handling
 *    - 400314 bit 0
 */

static int
nv40_graph_4097(struct drm_device *dev)
{
	struct drm_nouveau_private *dev_priv = dev->dev_private;

	if ((dev_priv->chipset & 0xf0) == 0x60)
		return 0;

	return !!(0x0baf & (1 << dev_priv->chipset));
}

static int
nv40_graph_vs_count(struct drm_device *dev)
{
	struct drm_nouveau_private *dev_priv = dev->dev_private;

	switch (dev_priv->chipset) {
	case 0x47:
	case 0x49:
	case 0x4b:
		return 8;
	case 0x40:
		return 6;
	case 0x41:
	case 0x42:
		return 5;
	case 0x43:
	case 0x44:
	case 0x46:
	case 0x4a:
		return 3;
	case 0x4c:
	case 0x4e:
	case 0x67:
	default:
		return 1;
	}
}


enum cp_label {
	cp_check_load = 1,
	cp_setup_auto_load,
	cp_setup_load,
	cp_setup_save,
	cp_swap_state,
	cp_swap_state3d_3_is_save,
	cp_prepare_exit,
	cp_exit,
};

static void
nv40_graph_construct_general(struct nouveau_grctx *ctx)
{
	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
	int i;

	cp_ctx(ctx, 0x4000a4, 1);
	gr_def(ctx, 0x4000a4, 0x00000008);
	cp_ctx(ctx, 0x400144, 58);
	gr_def(ctx, 0x400144, 0x00000001);
	cp_ctx(ctx, 0x400314, 1);
	gr_def(ctx, 0x400314, 0x00000000);
	cp_ctx(ctx, 0x400400, 10);
	cp_ctx(ctx, 0x400480, 10);
	cp_ctx(ctx, 0x400500, 19);
	gr_def(ctx, 0x400514, 0x00040000);
	gr_def(ctx, 0x400524, 0x55555555);
	gr_def(ctx, 0x400528, 0x55555555);
	gr_def(ctx, 0x40052c, 0x55555555);
	gr_def(ctx, 0x400530, 0x55555555);
	cp_ctx(ctx, 0x400560, 6);
	gr_def(ctx, 0x400568, 0x0000ffff);
	gr_def(ctx, 0x40056c, 0x0000ffff);
	cp_ctx(ctx, 0x40057c, 5);
	cp_ctx(ctx, 0x400710, 3);
	gr_def(ctx, 0x400710, 0x20010001);
	gr_def(ctx, 0x400714, 0x0f73ef00);
	cp_ctx(ctx, 0x400724, 1);
	gr_def(ctx, 0x400724, 0x02008821);
	cp_ctx(ctx, 0x400770, 3);
	if (dev_priv->chipset == 0x40) {
		cp_ctx(ctx, 0x400814, 4);
		cp_ctx(ctx, 0x400828, 5);
		cp_ctx(ctx, 0x400840, 5);
		gr_def(ctx, 0x400850, 0x00000040);
		cp_ctx(ctx, 0x400858, 4);
		gr_def(ctx, 0x400858, 0x00000040);
		gr_def(ctx, 0x40085c, 0x00000040);
		gr_def(ctx, 0x400864, 0x80000000);
		cp_ctx(ctx, 0x40086c, 9);
		gr_def(ctx, 0x40086c, 0x80000000);
		gr_def(ctx, 0x400870, 0x80000000);
		gr_def(ctx, 0x400874, 0x80000000);
		gr_def(ctx, 0x400878, 0x80000000);
		gr_def(ctx, 0x400888, 0x00000040);
		gr_def(ctx, 0x40088c, 0x80000000);
		cp_ctx(ctx, 0x4009c0, 8);
		gr_def(ctx, 0x4009cc, 0x80000000);
		gr_def(ctx, 0x4009dc, 0x80000000);
	} else {
		cp_ctx(ctx, 0x400840, 20);
		if (!nv40_graph_4097(ctx->dev)) {
			for (i = 0; i < 8; i++)
				gr_def(ctx, 0x400860 + (i * 4), 0x00000001);
		}
		gr_def(ctx, 0x400880, 0x00000040);
		gr_def(ctx, 0x400884, 0x00000040);
		gr_def(ctx, 0x400888, 0x00000040);
		cp_ctx(ctx, 0x400894, 11);
		gr_def(ctx, 0x400894, 0x00000040);
		if (nv40_graph_4097(ctx->dev)) {
			for (i = 0; i < 8; i++)
				gr_def(ctx, 0x4008a0 + (i * 4), 0x80000000);
		}
		cp_ctx(ctx, 0x4008e0, 2);
		cp_ctx(ctx, 0x4008f8, 2);
		if (dev_priv->chipset == 0x4c ||
		    (dev_priv->chipset & 0xf0) == 0x60)
			cp_ctx(ctx, 0x4009f8, 1);
	}
	cp_ctx(ctx, 0x400a00, 73);
	gr_def(ctx, 0x400b0c, 0x0b0b0b0c);
	cp_ctx(ctx, 0x401000, 4);
	cp_ctx(ctx, 0x405004, 1);
	switch (dev_priv->chipset) {
	case 0x47:
	case 0x49:
	case 0x4b:
		cp_ctx(ctx, 0x403448, 1);
		gr_def(ctx, 0x403448, 0x00001010);
		break;
	default:
		cp_ctx(ctx, 0x403440, 1);
		switch (dev_priv->chipset) {
		case 0x40:
			gr_def(ctx, 0x403440, 0x00000010);
			break;
		case 0x44:
		case 0x46:
		case 0x4a:
			gr_def(ctx, 0x403440, 0x00003010);
			break;
		case 0x41:
		case 0x42:
		case 0x43:
		case 0x4c:
		case 0x4e:
		case 0x67:
		default:
			gr_def(ctx, 0x403440, 0x00001010);
			break;
		}
		break;
	}
}

static void
nv40_graph_construct_state3d(struct nouveau_grctx *ctx)
{
	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
	int i;

	if (dev_priv->chipset == 0x40) {
		cp_ctx(ctx, 0x401880, 51);
		gr_def(ctx, 0x401940, 0x00000100);
	} else
	if (dev_priv->chipset == 0x46 || dev_priv->chipset == 0x47 ||
	    dev_priv->chipset == 0x49 || dev_priv->chipset == 0x4b) {
		cp_ctx(ctx, 0x401880, 32);
		for (i = 0; i < 16; i++)
			gr_def(ctx, 0x401880 + (i * 4), 0x00000111);
		if (dev_priv->chipset == 0x46)
			cp_ctx(ctx, 0x401900, 16);
		cp_ctx(ctx, 0x401940, 3);
	}
	cp_ctx(ctx, 0x40194c, 18);
	gr_def(ctx, 0x401954, 0x00000111);
	gr_def(ctx, 0x401958, 0x00080060);
	gr_def(ctx, 0x401974, 0x00000080);
	gr_def(ctx, 0x401978, 0xffff0000);
	gr_def(ctx, 0x40197c, 0x00000001);
	gr_def(ctx, 0x401990, 0x46400000);
	if (dev_priv->chipset == 0x40) {
		cp_ctx(ctx, 0x4019a0, 2);
		cp_ctx(ctx, 0x4019ac, 5);
	} else {
		cp_ctx(ctx, 0x4019a0, 1);
		cp_ctx(ctx, 0x4019b4, 3);
	}
	gr_def(ctx, 0x4019bc, 0xffff0000);
	switch (dev_priv->chipset) {
	case 0x46:
	case 0x47:
	case 0x49:
	case 0x4b:
		cp_ctx(ctx, 0x4019c0, 18);
		for (i = 0; i < 16; i++)
			gr_def(ctx, 0x4019c0 + (i * 4), 0x88888888);
		break;
	}
	cp_ctx(ctx, 0x401a08, 8);
	gr_def(ctx, 0x401a10, 0x0fff0000);
	gr_def(ctx, 0x401a14, 0x0fff0000);
	gr_def(ctx, 0x401a1c, 0x00011100);
	cp_ctx(ctx, 0x401a2c, 4);
	cp_ctx(ctx, 0x401a44, 26);
	for (i = 0; i < 16; i++)
		gr_def(ctx, 0x401a44 + (i * 4), 0x07ff0000);
	gr_def(ctx, 0x401a8c, 0x4b7fffff);
	if (dev_priv->chipset == 0x40) {
		cp_ctx(ctx, 0x401ab8, 3);
	} else {
		cp_ctx(ctx, 0x401ab8, 1);
		cp_ctx(ctx, 0x401ac0, 1);
	}
	cp_ctx(ctx, 0x401ad0, 8);
	gr_def(ctx, 0x401ad0, 0x30201000);
	gr_def(ctx, 0x401ad4, 0x70605040);
	gr_def(ctx, 0x401ad8, 0xb8a89888);
	gr_def(ctx, 0x401adc, 0xf8e8d8c8);
	cp_ctx(ctx, 0x401b10, dev_priv->chipset == 0x40 ? 2 : 1);
	gr_def(ctx, 0x401b10, 0x40100000);
	cp_ctx(ctx, 0x401b18, dev_priv->chipset == 0x40 ? 6 : 5);
	gr_def(ctx, 0x401b28, dev_priv->chipset == 0x40 ?
			      0x00000004 : 0x00000000);
	cp_ctx(ctx, 0x401b30, 25);
	gr_def(ctx, 0x401b34, 0x0000ffff);
	gr_def(ctx, 0x401b68, 0x435185d6);
	gr_def(ctx, 0x401b6c, 0x2155b699);
	gr_def(ctx, 0x401b70, 0xfedcba98);
	gr_def(ctx, 0x401b74, 0x00000098);
	gr_def(ctx, 0x401b84, 0xffffffff);
	gr_def(ctx, 0x401b88, 0x00ff7000);
	gr_def(ctx, 0x401b8c, 0x0000ffff);
	if (dev_priv->chipset != 0x44 && dev_priv->chipset != 0x4a &&
	    dev_priv->chipset != 0x4e)
		cp_ctx(ctx, 0x401b94, 1);
	cp_ctx(ctx, 0x401b98, 8);
	gr_def(ctx, 0x401b9c, 0x00ff0000);
	cp_ctx(ctx, 0x401bc0, 9);
	gr_def(ctx, 0x401be0, 0x00ffff00);
	cp_ctx(ctx, 0x401c00, 192);
	for (i = 0; i < 16; i++) { /* fragment texture units */
		gr_def(ctx, 0x401c40 + (i * 4), 0x00018488);
		gr_def(ctx, 0x401c80 + (i * 4), 0x00028202);
		gr_def(ctx, 0x401d00 + (i * 4), 0x0000aae4);
		gr_def(ctx, 0x401d40 + (i * 4), 0x01012000);
		gr_def(ctx, 0x401d80 + (i * 4), 0x00080008);
		gr_def(ctx, 0x401e00 + (i * 4), 0x00100008);
	}
	for (i = 0; i < 4; i++) { /* vertex texture units */
		gr_def(ctx, 0x401e90 + (i * 4), 0x0001bc80);
		gr_def(ctx, 0x401ea0 + (i * 4), 0x00000202);
		gr_def(ctx, 0x401ec0 + (i * 4), 0x00000008);
		gr_def(ctx, 0x401ee0 + (i * 4), 0x00080008);
	}
	cp_ctx(ctx, 0x400f5c, 3);
	gr_def(ctx, 0x400f5c, 0x00000002);
	cp_ctx(ctx, 0x400f84, 1);
}

static void
nv40_graph_construct_state3d_2(struct nouveau_grctx *ctx)
{
	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
	int i;

	cp_ctx(ctx, 0x402000, 1);
	cp_ctx(ctx, 0x402404, dev_priv->chipset == 0x40 ? 1 : 2);
	switch (dev_priv->chipset) {
	case 0x40:
		gr_def(ctx, 0x402404, 0x00000001);
		break;
	case 0x4c:
	case 0x4e:
	case 0x67:
		gr_def(ctx, 0x402404, 0x00000020);
		break;
	case 0x46:
	case 0x49:
	case 0x4b:
		gr_def(ctx, 0x402404, 0x00000421);
		break;
	default:
		gr_def(ctx, 0x402404, 0x00000021);
	}
	if (dev_priv->chipset != 0x40)
		gr_def(ctx, 0x402408, 0x030c30c3);
	switch (dev_priv->chipset) {
	case 0x44:
	case 0x46:
	case 0x4a:
	case 0x4c:
	case 0x4e:
	case 0x67:
		cp_ctx(ctx, 0x402440, 1);
		gr_def(ctx, 0x402440, 0x00011001);
		break;
	default:
		break;
	}
	cp_ctx(ctx, 0x402480, dev_priv->chipset == 0x40 ? 8 : 9);
	gr_def(ctx, 0x402488, 0x3e020200);
	gr_def(ctx, 0x40248c, 0x00ffffff);
	switch (dev_priv->chipset) {
	case 0x40:
		gr_def(ctx, 0x402490, 0x60103f00);
		break;
	case 0x47:
		gr_def(ctx, 0x402490, 0x40103f00);
		break;
	case 0x41:
	case 0x42:
	case 0x49:
	case 0x4b:
		gr_def(ctx, 0x402490, 0x20103f00);
		break;
	default:
		gr_def(ctx, 0x402490, 0x0c103f00);
		break;
	}
	gr_def(ctx, 0x40249c, dev_priv->chipset <= 0x43 ?
			      0x00020000 : 0x00040000);
	cp_ctx(ctx, 0x402500, 31);
	gr_def(ctx, 0x402530, 0x00008100);
	if (dev_priv->chipset == 0x40)
		cp_ctx(ctx, 0x40257c, 6);
	cp_ctx(ctx, 0x402594, 16);
	cp_ctx(ctx, 0x402800, 17);
	gr_def(ctx, 0x402800, 0x00000001);
	switch (dev_priv->chipset) {
	case 0x47:
	case 0x49:
	case 0x4b:
		cp_ctx(ctx, 0x402864, 1);
		gr_def(ctx, 0x402864, 0x00001001);
		cp_ctx(ctx, 0x402870, 3);
		gr_def(ctx, 0x402878, 0x00000003);
		if (dev_priv->chipset != 0x47) { /* belong at end!! */
			cp_ctx(ctx, 0x402900, 1);
			cp_ctx(ctx, 0x402940, 1);
			cp_ctx(ctx, 0x402980, 1);
			cp_ctx(ctx, 0x4029c0, 1);
			cp_ctx(ctx, 0x402a00, 1);
			cp_ctx(ctx, 0x402a40, 1);
			cp_ctx(ctx, 0x402a80, 1);
			cp_ctx(ctx, 0x402ac0, 1);
		}
		break;
	case 0x40:
		cp_ctx(ctx, 0x402844, 1);
		gr_def(ctx, 0x402844, 0x00000001);
		cp_ctx(ctx, 0x402850, 1);
		break;
	default:
		cp_ctx(ctx, 0x402844, 1);
		gr_def(ctx, 0x402844, 0x00001001);
		cp_ctx(ctx, 0x402850, 2);
		gr_def(ctx, 0x402854, 0x00000003);
		break;
	}

	cp_ctx(ctx, 0x402c00, 4);
	gr_def(ctx, 0x402c00, dev_priv->chipset == 0x40 ?
			      0x80800001 : 0x00888001);
	switch (dev_priv->chipset) {
	case 0x47:
	case 0x49:
	case 0x4b:
		cp_ctx(ctx, 0x402c20, 40);
		for (i = 0; i < 32; i++)
			gr_def(ctx, 0x402c40 + (i * 4), 0xffffffff);
		cp_ctx(ctx, 0x4030b8, 13);
		gr_def(ctx, 0x4030dc, 0x00000005);
		gr_def(ctx, 0x4030e8, 0x0000ffff);
		break;
	default:
		cp_ctx(ctx, 0x402c10, 4);
		if (dev_priv->chipset == 0x40)
			cp_ctx(ctx, 0x402c20, 36);
		else
		if (dev_priv->chipset <= 0x42)
			cp_ctx(ctx, 0x402c20, 24);
		else
		if (dev_priv->chipset <= 0x4a)
			cp_ctx(ctx, 0x402c20, 16);
		else
			cp_ctx(ctx, 0x402c20, 8);
		cp_ctx(ctx, 0x402cb0, dev_priv->chipset == 0x40 ? 12 : 13);
		gr_def(ctx, 0x402cd4, 0x00000005);
		if (dev_priv->chipset != 0x40)
			gr_def(ctx, 0x402ce0, 0x0000ffff);
		break;
	}

	cp_ctx(ctx, 0x403400, dev_priv->chipset == 0x40 ? 4 : 3);
	cp_ctx(ctx, 0x403410, dev_priv->chipset == 0x40 ? 4 : 3);
	cp_ctx(ctx, 0x403420, nv40_graph_vs_count(ctx->dev));
	for (i = 0; i < nv40_graph_vs_count(ctx->dev); i++)
		gr_def(ctx, 0x403420 + (i * 4), 0x00005555);

	if (dev_priv->chipset != 0x40) {
		cp_ctx(ctx, 0x403600, 1);
		gr_def(ctx, 0x403600, 0x00000001);
	}
	cp_ctx(ctx, 0x403800, 1);

	cp_ctx(ctx, 0x403c18, 1);
	gr_def(ctx, 0x403c18, 0x00000001);
	switch (dev_priv->chipset) {
	case 0x46:
	case 0x47:
	case 0x49:
	case 0x4b:
		cp_ctx(ctx, 0x405018, 1);
		gr_def(ctx, 0x405018, 0x08e00001);
		cp_ctx(ctx, 0x405c24, 1);
		gr_def(ctx, 0x405c24, 0x000e3000);
		break;
	}
	if (dev_priv->chipset != 0x4e)
		cp_ctx(ctx, 0x405800, 11);
	cp_ctx(ctx, 0x407000, 1);
}

static void
nv40_graph_construct_state3d_3(struct nouveau_grctx *ctx)
{
	int len = nv40_graph_4097(ctx->dev) ? 0x0684 : 0x0084;

	cp_out (ctx, 0x300000);
	cp_lsr (ctx, len - 4);
	cp_bra (ctx, SWAP_DIRECTION, SAVE, cp_swap_state3d_3_is_save);
	cp_lsr (ctx, len);
	cp_name(ctx, cp_swap_state3d_3_is_save);
	cp_out (ctx, 0x800001);

	ctx->ctxvals_pos += len;
}

static void
nv40_graph_construct_shader(struct nouveau_grctx *ctx)
{
	struct drm_device *dev = ctx->dev;
	struct drm_nouveau_private *dev_priv = dev->dev_private;
	struct nouveau_gpuobj *obj = ctx->data;
	int vs, vs_nr, vs_len, vs_nr_b0, vs_nr_b1, b0_offset, b1_offset;
	int offset, i;

	vs_nr    = nv40_graph_vs_count(ctx->dev);
	vs_nr_b0 = 363;
	vs_nr_b1 = dev_priv->chipset == 0x40 ? 128 : 64;
	if (dev_priv->chipset == 0x40) {
		b0_offset = 0x2200/4; /* 33a0 */
		b1_offset = 0x55a0/4; /* 1500 */
		vs_len = 0x6aa0/4;
	} else
	if (dev_priv->chipset == 0x41 || dev_priv->chipset == 0x42) {
		b0_offset = 0x2200/4; /* 2200 */
		b1_offset = 0x4400/4; /* 0b00 */
		vs_len = 0x4f00/4;
	} else {
		b0_offset = 0x1d40/4; /* 2200 */
		b1_offset = 0x3f40/4; /* 0b00 : 0a40 */
		vs_len = nv40_graph_4097(dev) ? 0x4a40/4 : 0x4980/4;
	}

	cp_lsr(ctx, vs_len * vs_nr + 0x300/4);
	cp_out(ctx, nv40_graph_4097(dev) ? 0x800041 : 0x800029);

	offset = ctx->ctxvals_pos;
	ctx->ctxvals_pos += (0x0300/4 + (vs_nr * vs_len));

	if (ctx->mode != NOUVEAU_GRCTX_VALS)
		return;

	offset += 0x0280/4;
	for (i = 0; i < 16; i++, offset += 2)
		nv_wo32(dev, obj, offset, 0x3f800000);

	for (vs = 0; vs < vs_nr; vs++, offset += vs_len) {
		for (i = 0; i < vs_nr_b0 * 6; i += 6)
			nv_wo32(dev, obj, offset + b0_offset + i, 0x00000001);
		for (i = 0; i < vs_nr_b1 * 4; i += 4)
			nv_wo32(dev, obj, offset + b1_offset + i, 0x3f800000);
	}
}

void
nv40_grctx_init(struct nouveau_grctx *ctx)
{
	/* decide whether we're loading/unloading the context */
	cp_bra (ctx, AUTO_SAVE, PENDING, cp_setup_save);
	cp_bra (ctx, USER_SAVE, PENDING, cp_setup_save);

	cp_name(ctx, cp_check_load);
	cp_bra (ctx, AUTO_LOAD, PENDING, cp_setup_auto_load);
	cp_bra (ctx, USER_LOAD, PENDING, cp_setup_load);
	cp_bra (ctx, ALWAYS, TRUE, cp_exit);

	/* setup for context load */
	cp_name(ctx, cp_setup_auto_load);
	cp_wait(ctx, STATUS, IDLE);
	cp_out (ctx, CP_NEXT_TO_SWAP);
	cp_name(ctx, cp_setup_load);
	cp_wait(ctx, STATUS, IDLE);
	cp_set (ctx, SWAP_DIRECTION, LOAD);
	cp_out (ctx, 0x00910880); /* ?? */
	cp_out (ctx, 0x00901ffe); /* ?? */
	cp_out (ctx, 0x01940000); /* ?? */
	cp_lsr (ctx, 0x20);
	cp_out (ctx, 0x0060000b); /* ?? */
	cp_wait(ctx, UNK57, CLEAR);
	cp_out (ctx, 0x0060000c); /* ?? */
	cp_bra (ctx, ALWAYS, TRUE, cp_swap_state);

	/* setup for context save */
	cp_name(ctx, cp_setup_save);
	cp_set (ctx, SWAP_DIRECTION, SAVE);

	/* general PGRAPH state */
	cp_name(ctx, cp_swap_state);
	cp_pos (ctx, 0x00020/4);
	nv40_graph_construct_general(ctx);
	cp_wait(ctx, STATUS, IDLE);

	/* 3D state, block 1 */
	cp_bra (ctx, UNK54, CLEAR, cp_prepare_exit);
	nv40_graph_construct_state3d(ctx);
	cp_wait(ctx, STATUS, IDLE);

	/* 3D state, block 2 */
	nv40_graph_construct_state3d_2(ctx);

	/* Some other block of "random" state */
	nv40_graph_construct_state3d_3(ctx);

	/* Per-vertex shader state */
	cp_pos (ctx, ctx->ctxvals_pos);
	nv40_graph_construct_shader(ctx);

	/* pre-exit state updates */
	cp_name(ctx, cp_prepare_exit);
	cp_bra (ctx, SWAP_DIRECTION, SAVE, cp_check_load);
	cp_bra (ctx, USER_SAVE, PENDING, cp_exit);
	cp_out (ctx, CP_NEXT_TO_CURRENT);

	cp_name(ctx, cp_exit);
	cp_set (ctx, USER_SAVE, NOT_PENDING);
	cp_set (ctx, USER_LOAD, NOT_PENDING);
	cp_out (ctx, CP_END);
}