1 /*
2  * Copyright 2013 Red Hat Inc.
3  *
4  * Permission is hereby granted, free of charge, to any person obtaining a
5  * copy of this software and associated documentation files (the "Software"),
6  * to deal in the Software without restriction, including without limitation
7  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8  * and/or sell copies of the Software, and to permit persons to whom the
9  * Software is furnished to do so, subject to the following conditions:
10  *
11  * The above copyright notice and this permission notice shall be included in
12  * all copies or substantial portions of the Software.
13  *
14  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
15  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
16  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
17  * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
18  * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
19  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
20  * OTHER DEALINGS IN THE SOFTWARE.
21  *
22  * Authors: Ben Skeggs <bskeggs@redhat.com>
23  */
24 #include "gf100.h"
25 #include "ctxgf100.h"
26 
27 #include <subdev/bios.h>
28 #include <subdev/bios/P0260.h>
29 
30 #include <nvif/class.h>
31 
32 /*******************************************************************************
33  * PGRAPH register lists
34  ******************************************************************************/
35 
36 static const struct gf100_gr_init
37 gm107_gr_init_main_0[] = {
38 	{ 0x400080,   1, 0x04, 0x003003c2 },
39 	{ 0x400088,   1, 0x04, 0x0001bfe7 },
40 	{ 0x40008c,   1, 0x04, 0x00060000 },
41 	{ 0x400090,   1, 0x04, 0x00000030 },
42 	{ 0x40013c,   1, 0x04, 0x003901f3 },
43 	{ 0x400140,   1, 0x04, 0x00000100 },
44 	{ 0x400144,   1, 0x04, 0x00000000 },
45 	{ 0x400148,   1, 0x04, 0x00000110 },
46 	{ 0x400138,   1, 0x04, 0x00000000 },
47 	{ 0x400130,   2, 0x04, 0x00000000 },
48 	{ 0x400124,   1, 0x04, 0x00000002 },
49 	{}
50 };
51 
52 static const struct gf100_gr_init
53 gm107_gr_init_ds_0[] = {
54 	{ 0x405844,   1, 0x04, 0x00ffffff },
55 	{ 0x405850,   1, 0x04, 0x00000000 },
56 	{ 0x405900,   1, 0x04, 0x00000000 },
57 	{ 0x405908,   1, 0x04, 0x00000000 },
58 	{}
59 };
60 
61 const struct gf100_gr_init
62 gm107_gr_init_scc_0[] = {
63 	{ 0x40803c,   1, 0x04, 0x00000010 },
64 	{}
65 };
66 
67 static const struct gf100_gr_init
68 gm107_gr_init_sked_0[] = {
69 	{ 0x407010,   1, 0x04, 0x00000000 },
70 	{ 0x407040,   1, 0x04, 0x40440424 },
71 	{ 0x407048,   1, 0x04, 0x0000000a },
72 	{}
73 };
74 
75 const struct gf100_gr_init
76 gm107_gr_init_prop_0[] = {
77 	{ 0x418408,   1, 0x04, 0x00000000 },
78 	{ 0x4184a0,   1, 0x04, 0x00000000 },
79 	{}
80 };
81 
82 const struct gf100_gr_init
83 gm107_gr_init_setup_1[] = {
84 	{ 0x4188c8,   2, 0x04, 0x00000000 },
85 	{ 0x4188d0,   1, 0x04, 0x00010000 },
86 	{ 0x4188d4,   1, 0x04, 0x00010201 },
87 	{}
88 };
89 
90 const struct gf100_gr_init
91 gm107_gr_init_zcull_0[] = {
92 	{ 0x418910,   1, 0x04, 0x00010001 },
93 	{ 0x418914,   1, 0x04, 0x00000301 },
94 	{ 0x418918,   1, 0x04, 0x00800000 },
95 	{ 0x418930,   2, 0x04, 0x00000000 },
96 	{ 0x418980,   1, 0x04, 0x77777770 },
97 	{ 0x418984,   3, 0x04, 0x77777777 },
98 	{}
99 };
100 
101 const struct gf100_gr_init
102 gm107_gr_init_gpc_unk_1[] = {
103 	{ 0x418d00,   1, 0x04, 0x00000000 },
104 	{ 0x418f00,   1, 0x04, 0x00000400 },
105 	{ 0x418f08,   1, 0x04, 0x00000000 },
106 	{ 0x418e08,   1, 0x04, 0x00000000 },
107 	{}
108 };
109 
110 static const struct gf100_gr_init
111 gm107_gr_init_tpccs_0[] = {
112 	{ 0x419dc4,   1, 0x04, 0x00000000 },
113 	{ 0x419dc8,   1, 0x04, 0x00000501 },
114 	{ 0x419dd0,   1, 0x04, 0x00000000 },
115 	{ 0x419dd4,   1, 0x04, 0x00000100 },
116 	{ 0x419dd8,   1, 0x04, 0x00000001 },
117 	{ 0x419ddc,   1, 0x04, 0x00000002 },
118 	{ 0x419de0,   1, 0x04, 0x00000001 },
119 	{ 0x419d0c,   1, 0x04, 0x00000000 },
120 	{ 0x419d10,   1, 0x04, 0x00000014 },
121 	{}
122 };
123 
124 const struct gf100_gr_init
125 gm107_gr_init_tex_0[] = {
126 	{ 0x419ab0,   1, 0x04, 0x00000000 },
127 	{ 0x419ab8,   1, 0x04, 0x000000e7 },
128 	{ 0x419abc,   1, 0x04, 0x00000000 },
129 	{ 0x419acc,   1, 0x04, 0x000000ff },
130 	{ 0x419ac0,   1, 0x04, 0x00000000 },
131 	{ 0x419aa8,   2, 0x04, 0x00000000 },
132 	{ 0x419ad0,   2, 0x04, 0x00000000 },
133 	{ 0x419ae0,   2, 0x04, 0x00000000 },
134 	{ 0x419af0,   4, 0x04, 0x00000000 },
135 	{}
136 };
137 
138 static const struct gf100_gr_init
139 gm107_gr_init_pe_0[] = {
140 	{ 0x419900,   1, 0x04, 0x000000ff },
141 	{ 0x41980c,   1, 0x04, 0x00000010 },
142 	{ 0x419844,   1, 0x04, 0x00000000 },
143 	{ 0x419838,   1, 0x04, 0x000000ff },
144 	{ 0x419850,   1, 0x04, 0x00000004 },
145 	{ 0x419854,   2, 0x04, 0x00000000 },
146 	{ 0x419894,   3, 0x04, 0x00100401 },
147 	{}
148 };
149 
150 const struct gf100_gr_init
151 gm107_gr_init_l1c_0[] = {
152 	{ 0x419c98,   1, 0x04, 0x00000000 },
153 	{ 0x419cc0,   2, 0x04, 0x00000000 },
154 	{}
155 };
156 
157 static const struct gf100_gr_init
158 gm107_gr_init_sm_0[] = {
159 	{ 0x419e30,   1, 0x04, 0x000000ff },
160 	{ 0x419e00,   1, 0x04, 0x00000000 },
161 	{ 0x419ea0,   1, 0x04, 0x00000000 },
162 	{ 0x419ee4,   1, 0x04, 0x00000000 },
163 	{ 0x419ea4,   1, 0x04, 0x00000100 },
164 	{ 0x419ea8,   1, 0x04, 0x01000000 },
165 	{ 0x419ee8,   1, 0x04, 0x00000091 },
166 	{ 0x419eb4,   1, 0x04, 0x00000000 },
167 	{ 0x419ebc,   2, 0x04, 0x00000000 },
168 	{ 0x419edc,   1, 0x04, 0x000c1810 },
169 	{ 0x419ed8,   1, 0x04, 0x00000000 },
170 	{ 0x419ee0,   1, 0x04, 0x00000000 },
171 	{ 0x419f74,   1, 0x04, 0x00005155 },
172 	{ 0x419f80,   4, 0x04, 0x00000000 },
173 	{}
174 };
175 
176 static const struct gf100_gr_init
177 gm107_gr_init_l1c_1[] = {
178 	{ 0x419ccc,   2, 0x04, 0x00000000 },
179 	{ 0x419c80,   1, 0x04, 0x3f006022 },
180 	{ 0x419c88,   1, 0x04, 0x00000000 },
181 	{}
182 };
183 
184 static const struct gf100_gr_init
185 gm107_gr_init_pes_0[] = {
186 	{ 0x41be50,   1, 0x04, 0x000000ff },
187 	{ 0x41be04,   1, 0x04, 0x00000000 },
188 	{ 0x41be08,   1, 0x04, 0x00000004 },
189 	{ 0x41be0c,   1, 0x04, 0x00000008 },
190 	{ 0x41be10,   1, 0x04, 0x0e3b8bc7 },
191 	{ 0x41be14,   2, 0x04, 0x00000000 },
192 	{ 0x41be3c,   5, 0x04, 0x00100401 },
193 	{}
194 };
195 
196 const struct gf100_gr_init
197 gm107_gr_init_wwdx_0[] = {
198 	{ 0x41bfd4,   1, 0x04, 0x00800000 },
199 	{ 0x41bfdc,   1, 0x04, 0x00000000 },
200 	{}
201 };
202 
203 const struct gf100_gr_init
204 gm107_gr_init_cbm_0[] = {
205 	{ 0x41becc,   1, 0x04, 0x00000000 },
206 	{}
207 };
208 
209 static const struct gf100_gr_init
210 gm107_gr_init_be_0[] = {
211 	{ 0x408890,   1, 0x04, 0x000000ff },
212 	{ 0x40880c,   1, 0x04, 0x00000000 },
213 	{ 0x408850,   1, 0x04, 0x00000004 },
214 	{ 0x408878,   1, 0x04, 0x00c81603 },
215 	{ 0x40887c,   1, 0x04, 0x80543432 },
216 	{ 0x408880,   1, 0x04, 0x0010581e },
217 	{ 0x408884,   1, 0x04, 0x00001205 },
218 	{ 0x408974,   1, 0x04, 0x000000ff },
219 	{ 0x408910,   9, 0x04, 0x00000000 },
220 	{ 0x408950,   1, 0x04, 0x00000000 },
221 	{ 0x408954,   1, 0x04, 0x0000ffff },
222 	{ 0x408958,   1, 0x04, 0x00000034 },
223 	{ 0x40895c,   1, 0x04, 0x8531a003 },
224 	{ 0x408960,   1, 0x04, 0x0561985a },
225 	{ 0x408964,   1, 0x04, 0x04e15c4f },
226 	{ 0x408968,   1, 0x04, 0x02808833 },
227 	{ 0x40896c,   1, 0x04, 0x01f02438 },
228 	{ 0x408970,   1, 0x04, 0x00012c00 },
229 	{ 0x408984,   1, 0x04, 0x00000000 },
230 	{ 0x408988,   1, 0x04, 0x08040201 },
231 	{ 0x40898c,   1, 0x04, 0x80402010 },
232 	{}
233 };
234 
235 static const struct gf100_gr_init
236 gm107_gr_init_sm_1[] = {
237 	{ 0x419e5c,   1, 0x04, 0x00000000 },
238 	{ 0x419e58,   1, 0x04, 0x00000000 },
239 	{}
240 };
241 
242 static const struct gf100_gr_pack
243 gm107_gr_pack_mmio[] = {
244 	{ gm107_gr_init_main_0 },
245 	{ gk110_gr_init_fe_0 },
246 	{ gf100_gr_init_pri_0 },
247 	{ gf100_gr_init_rstr2d_0 },
248 	{ gf100_gr_init_pd_0 },
249 	{ gm107_gr_init_ds_0 },
250 	{ gm107_gr_init_scc_0 },
251 	{ gm107_gr_init_sked_0 },
252 	{ gk110_gr_init_cwd_0 },
253 	{ gm107_gr_init_prop_0 },
254 	{ gk208_gr_init_gpc_unk_0 },
255 	{ gf100_gr_init_setup_0 },
256 	{ gf100_gr_init_crstr_0 },
257 	{ gm107_gr_init_setup_1 },
258 	{ gm107_gr_init_zcull_0 },
259 	{ gf100_gr_init_gpm_0 },
260 	{ gm107_gr_init_gpc_unk_1 },
261 	{ gf100_gr_init_gcc_0 },
262 	{ gm107_gr_init_tpccs_0 },
263 	{ gm107_gr_init_tex_0 },
264 	{ gm107_gr_init_pe_0 },
265 	{ gm107_gr_init_l1c_0 },
266 	{ gf100_gr_init_mpc_0 },
267 	{ gm107_gr_init_sm_0 },
268 	{ gm107_gr_init_l1c_1 },
269 	{ gm107_gr_init_pes_0 },
270 	{ gm107_gr_init_wwdx_0 },
271 	{ gm107_gr_init_cbm_0 },
272 	{ gm107_gr_init_be_0 },
273 	{ gm107_gr_init_sm_1 },
274 	{}
275 };
276 
277 /*******************************************************************************
278  * PGRAPH engine/subdev functions
279  ******************************************************************************/
280 
281 void
282 gm107_gr_init_bios(struct gf100_gr *gr)
283 {
284 	static const struct {
285 		u32 ctrl;
286 		u32 data;
287 	} regs[] = {
288 		{ 0x419ed8, 0x419ee0 },
289 		{ 0x419ad0, 0x419ad4 },
290 		{ 0x419ae0, 0x419ae4 },
291 		{ 0x419af0, 0x419af4 },
292 		{ 0x419af8, 0x419afc },
293 	};
294 	struct nvkm_device *device = gr->base.engine.subdev.device;
295 	struct nvkm_bios *bios = device->bios;
296 	struct nvbios_P0260E infoE;
297 	struct nvbios_P0260X infoX;
298 	int E = -1, X;
299 	u8 ver, hdr;
300 
301 	while (nvbios_P0260Ep(bios, ++E, &ver, &hdr, &infoE)) {
302 		if (X = -1, E < ARRAY_SIZE(regs)) {
303 			nvkm_wr32(device, regs[E].ctrl, infoE.data);
304 			while (nvbios_P0260Xp(bios, ++X, &ver, &hdr, &infoX))
305 				nvkm_wr32(device, regs[E].data, infoX.data);
306 		}
307 	}
308 }
309 
310 int
311 gm107_gr_init(struct gf100_gr *gr)
312 {
313 	struct nvkm_device *device = gr->base.engine.subdev.device;
314 	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, gr->tpc_total);
315 	u32 data[TPC_MAX / 8] = {};
316 	u8  tpcnr[GPC_MAX];
317 	int gpc, tpc, ppc, rop;
318 	int i;
319 
320 	nvkm_wr32(device, GPC_BCAST(0x0880), 0x00000000);
321 	nvkm_wr32(device, GPC_BCAST(0x0890), 0x00000000);
322 	nvkm_wr32(device, GPC_BCAST(0x0894), 0x00000000);
323 	nvkm_wr32(device, GPC_BCAST(0x08b4), nvkm_memory_addr(gr->unk4188b4) >> 8);
324 	nvkm_wr32(device, GPC_BCAST(0x08b8), nvkm_memory_addr(gr->unk4188b8) >> 8);
325 
326 	gf100_gr_mmio(gr, gr->func->mmio);
327 
328 	gm107_gr_init_bios(gr);
329 
330 	nvkm_wr32(device, GPC_UNIT(0, 0x3018), 0x00000001);
331 
332 	memset(data, 0x00, sizeof(data));
333 	memcpy(tpcnr, gr->tpc_nr, sizeof(gr->tpc_nr));
334 	for (i = 0, gpc = -1; i < gr->tpc_total; i++) {
335 		do {
336 			gpc = (gpc + 1) % gr->gpc_nr;
337 		} while (!tpcnr[gpc]);
338 		tpc = gr->tpc_nr[gpc] - tpcnr[gpc]--;
339 
340 		data[i / 8] |= tpc << ((i % 8) * 4);
341 	}
342 
343 	nvkm_wr32(device, GPC_BCAST(0x0980), data[0]);
344 	nvkm_wr32(device, GPC_BCAST(0x0984), data[1]);
345 	nvkm_wr32(device, GPC_BCAST(0x0988), data[2]);
346 	nvkm_wr32(device, GPC_BCAST(0x098c), data[3]);
347 
348 	for (gpc = 0; gpc < gr->gpc_nr; gpc++) {
349 		nvkm_wr32(device, GPC_UNIT(gpc, 0x0914),
350 			gr->magic_not_rop_nr << 8 | gr->tpc_nr[gpc]);
351 		nvkm_wr32(device, GPC_UNIT(gpc, 0x0910), 0x00040000 |
352 			gr->tpc_total);
353 		nvkm_wr32(device, GPC_UNIT(gpc, 0x0918), magicgpc918);
354 	}
355 
356 	nvkm_wr32(device, GPC_BCAST(0x3fd4), magicgpc918);
357 	nvkm_wr32(device, GPC_BCAST(0x08ac), nvkm_rd32(device, 0x100800));
358 
359 	nvkm_wr32(device, 0x400500, 0x00010001);
360 
361 	nvkm_wr32(device, 0x400100, 0xffffffff);
362 	nvkm_wr32(device, 0x40013c, 0xffffffff);
363 	nvkm_wr32(device, 0x400124, 0x00000002);
364 	nvkm_wr32(device, 0x409c24, 0x000e0000);
365 
366 	nvkm_wr32(device, 0x404000, 0xc0000000);
367 	nvkm_wr32(device, 0x404600, 0xc0000000);
368 	nvkm_wr32(device, 0x408030, 0xc0000000);
369 	nvkm_wr32(device, 0x404490, 0xc0000000);
370 	nvkm_wr32(device, 0x406018, 0xc0000000);
371 	nvkm_wr32(device, 0x407020, 0x40000000);
372 	nvkm_wr32(device, 0x405840, 0xc0000000);
373 	nvkm_wr32(device, 0x405844, 0x00ffffff);
374 	nvkm_mask(device, 0x419cc0, 0x00000008, 0x00000008);
375 
376 	for (gpc = 0; gpc < gr->gpc_nr; gpc++) {
377 		for (ppc = 0; ppc < 2 /* gr->ppc_nr[gpc] */; ppc++)
378 			nvkm_wr32(device, PPC_UNIT(gpc, ppc, 0x038), 0xc0000000);
379 		nvkm_wr32(device, GPC_UNIT(gpc, 0x0420), 0xc0000000);
380 		nvkm_wr32(device, GPC_UNIT(gpc, 0x0900), 0xc0000000);
381 		nvkm_wr32(device, GPC_UNIT(gpc, 0x1028), 0xc0000000);
382 		nvkm_wr32(device, GPC_UNIT(gpc, 0x0824), 0xc0000000);
383 		for (tpc = 0; tpc < gr->tpc_nr[gpc]; tpc++) {
384 			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
385 			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
386 			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
387 			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
388 			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
389 			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x430), 0xc0000000);
390 			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x644), 0x00dffffe);
391 			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x64c), 0x00000005);
392 		}
393 		nvkm_wr32(device, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
394 		nvkm_wr32(device, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
395 	}
396 
397 	for (rop = 0; rop < gr->rop_nr; rop++) {
398 		nvkm_wr32(device, ROP_UNIT(rop, 0x144), 0x40000000);
399 		nvkm_wr32(device, ROP_UNIT(rop, 0x070), 0x40000000);
400 		nvkm_wr32(device, ROP_UNIT(rop, 0x204), 0xffffffff);
401 		nvkm_wr32(device, ROP_UNIT(rop, 0x208), 0xffffffff);
402 	}
403 
404 	nvkm_wr32(device, 0x400108, 0xffffffff);
405 	nvkm_wr32(device, 0x400138, 0xffffffff);
406 	nvkm_wr32(device, 0x400118, 0xffffffff);
407 	nvkm_wr32(device, 0x400130, 0xffffffff);
408 	nvkm_wr32(device, 0x40011c, 0xffffffff);
409 	nvkm_wr32(device, 0x400134, 0xffffffff);
410 
411 	nvkm_wr32(device, 0x400054, 0x2c350f63);
412 
413 	gf100_gr_zbc_init(gr);
414 
415 	return gf100_gr_init_ctxctl(gr);
416 }
417 
418 #include "fuc/hubgm107.fuc5.h"
419 
420 static struct gf100_gr_ucode
421 gm107_gr_fecs_ucode = {
422 	.code.data = gm107_grhub_code,
423 	.code.size = sizeof(gm107_grhub_code),
424 	.data.data = gm107_grhub_data,
425 	.data.size = sizeof(gm107_grhub_data),
426 };
427 
428 #include "fuc/gpcgm107.fuc5.h"
429 
430 static struct gf100_gr_ucode
431 gm107_gr_gpccs_ucode = {
432 	.code.data = gm107_grgpc_code,
433 	.code.size = sizeof(gm107_grgpc_code),
434 	.data.data = gm107_grgpc_data,
435 	.data.size = sizeof(gm107_grgpc_data),
436 };
437 
438 static const struct gf100_gr_func
439 gm107_gr = {
440 	.init = gm107_gr_init,
441 	.mmio = gm107_gr_pack_mmio,
442 	.fecs.ucode = &gm107_gr_fecs_ucode,
443 	.gpccs.ucode = &gm107_gr_gpccs_ucode,
444 	.ppc_nr = 2,
445 	.grctx = &gm107_grctx,
446 	.sclass = {
447 		{ -1, -1, FERMI_TWOD_A },
448 		{ -1, -1, KEPLER_INLINE_TO_MEMORY_B },
449 		{ -1, -1, MAXWELL_A, &gf100_fermi },
450 		{ -1, -1, MAXWELL_COMPUTE_A },
451 		{}
452 	}
453 };
454 
455 int
456 gm107_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
457 {
458 	return gf100_gr_new_(&gm107_gr, device, index, pgr);
459 }
460