i915_gpu_error.c 40.3 KB
Newer Older
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
/*
 * Copyright (c) 2008 Intel Corporation
 *
 * 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 (including the next
 * paragraph) 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 OR COPYRIGHT HOLDERS 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:
 *    Eric Anholt <eric@anholt.net>
 *    Keith Packard <keithp@keithp.com>
 *    Mika Kuoppala <mika.kuoppala@intel.com>
 *
 */

#include <generated/utsrelease.h>
#include "i915_drv.h"

static const char *ring_str(int ring)
{
	switch (ring) {
	case RCS: return "render";
	case VCS: return "bsd";
	case BCS: return "blt";
	case VECS: return "vebox";
40
	case VCS2: return "bsd2";
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
	default: return "";
	}
}

static const char *pin_flag(int pinned)
{
	if (pinned > 0)
		return " P";
	else if (pinned < 0)
		return " p";
	else
		return "";
}

static const char *tiling_flag(int tiling)
{
	switch (tiling) {
	default:
	case I915_TILING_NONE: return "";
	case I915_TILING_X: return " X";
	case I915_TILING_Y: return " Y";
	}
}

static const char *dirty_flag(int dirty)
{
	return dirty ? " dirty" : "";
}

static const char *purgeable_flag(int purgeable)
{
	return purgeable ? " purgeable" : "";
}

static bool __i915_error_ok(struct drm_i915_error_state_buf *e)
{

	if (!e->err && WARN(e->bytes > (e->size - 1), "overflow")) {
		e->err = -ENOSPC;
		return false;
	}

	if (e->bytes == e->size - 1 || e->err)
		return false;

	return true;
}

static bool __i915_error_seek(struct drm_i915_error_state_buf *e,
			      unsigned len)
{
	if (e->pos + len <= e->start) {
		e->pos += len;
		return false;
	}

	/* First vsnprintf needs to fit in its entirety for memmove */
	if (len >= e->size) {
		e->err = -EIO;
		return false;
	}

	return true;
}

static void __i915_error_advance(struct drm_i915_error_state_buf *e,
				 unsigned len)
{
	/* If this is first printf in this window, adjust it so that
	 * start position matches start of the buffer
	 */

	if (e->pos < e->start) {
		const size_t off = e->start - e->pos;

		/* Should not happen but be paranoid */
		if (off > len || e->bytes) {
			e->err = -EIO;
			return;
		}

		memmove(e->buf, e->buf + off, len - off);
		e->bytes = len - off;
		e->pos = e->start;
		return;
	}

	e->bytes += len;
	e->pos += len;
}

static void i915_error_vprintf(struct drm_i915_error_state_buf *e,
			       const char *f, va_list args)
{
	unsigned len;

	if (!__i915_error_ok(e))
		return;

	/* Seek the first printf which is hits start position */
	if (e->pos < e->start) {
142 143 144
		va_list tmp;

		va_copy(tmp, args);
145 146 147 148
		len = vsnprintf(NULL, 0, f, tmp);
		va_end(tmp);

		if (!__i915_error_seek(e, len))
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
			return;
	}

	len = vsnprintf(e->buf + e->bytes, e->size - e->bytes, f, args);
	if (len >= e->size - e->bytes)
		len = e->size - e->bytes - 1;

	__i915_error_advance(e, len);
}

static void i915_error_puts(struct drm_i915_error_state_buf *e,
			    const char *str)
{
	unsigned len;

	if (!__i915_error_ok(e))
		return;

	len = strlen(str);

	/* Seek the first printf which is hits start position */
	if (e->pos < e->start) {
		if (!__i915_error_seek(e, len))
			return;
	}

	if (len >= e->size - e->bytes)
		len = e->size - e->bytes - 1;
	memcpy(e->buf + e->bytes, str, len);

	__i915_error_advance(e, len);
}

#define err_printf(e, ...) i915_error_printf(e, __VA_ARGS__)
#define err_puts(e, s) i915_error_puts(e, s)

static void print_error_buffers(struct drm_i915_error_state_buf *m,
				const char *name,
				struct drm_i915_error_buffer *err,
				int count)
{
190 191
	int i;

192
	err_printf(m, "  %s [%d]:\n", name, count);
193 194

	while (count--) {
195 196 197
		err_printf(m, "    %08x_%08x %8u %02x %02x [ ",
			   upper_32_bits(err->gtt_offset),
			   lower_32_bits(err->gtt_offset),
198 199
			   err->size,
			   err->read_domains,
200
			   err->write_domain);
201
		for (i = 0; i < I915_NUM_ENGINES; i++)
202 203 204
			err_printf(m, "%02x ", err->rseqno[i]);

		err_printf(m, "] %02x", err->wseqno);
205 206 207 208
		err_puts(m, pin_flag(err->pinned));
		err_puts(m, tiling_flag(err->tiling));
		err_puts(m, dirty_flag(err->dirty));
		err_puts(m, purgeable_flag(err->purgeable));
209
		err_puts(m, err->userptr ? " userptr" : "");
210 211
		err_puts(m, err->ring != -1 ? " " : "");
		err_puts(m, ring_str(err->ring));
212
		err_puts(m, i915_cache_level_str(m->i915, err->cache_level));
213 214 215 216 217 218 219 220 221 222 223

		if (err->name)
			err_printf(m, " (name: %d)", err->name);
		if (err->fence_reg != I915_FENCE_REG_NONE)
			err_printf(m, " (fence: %d)", err->fence_reg);

		err_puts(m, "\n");
		err++;
	}
}

224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241
static const char *hangcheck_action_to_str(enum intel_ring_hangcheck_action a)
{
	switch (a) {
	case HANGCHECK_IDLE:
		return "idle";
	case HANGCHECK_WAIT:
		return "wait";
	case HANGCHECK_ACTIVE:
		return "active";
	case HANGCHECK_KICK:
		return "kick";
	case HANGCHECK_HUNG:
		return "hung";
	}

	return "unknown";
}

242 243
static void i915_ring_error_state(struct drm_i915_error_state_buf *m,
				  struct drm_device *dev,
244 245
				  struct drm_i915_error_state *error,
				  int ring_idx)
246
{
247 248
	struct drm_i915_error_ring *ring = &error->ring[ring_idx];

249
	if (!ring->valid)
250 251
		return;

252
	err_printf(m, "%s command stream:\n", ring_str(ring_idx));
253 254 255 256 257
	err_printf(m, "  START: 0x%08x\n", ring->start);
	err_printf(m, "  HEAD:  0x%08x\n", ring->head);
	err_printf(m, "  TAIL:  0x%08x\n", ring->tail);
	err_printf(m, "  CTL:   0x%08x\n", ring->ctl);
	err_printf(m, "  HWS:   0x%08x\n", ring->hws);
258
	err_printf(m, "  ACTHD: 0x%08x %08x\n", (u32)(ring->acthd>>32), (u32)ring->acthd);
259 260 261
	err_printf(m, "  IPEIR: 0x%08x\n", ring->ipeir);
	err_printf(m, "  IPEHR: 0x%08x\n", ring->ipehr);
	err_printf(m, "  INSTDONE: 0x%08x\n", ring->instdone);
262
	if (INTEL_INFO(dev)->gen >= 4) {
263
		err_printf(m, "  BBADDR: 0x%08x %08x\n", (u32)(ring->bbaddr>>32), (u32)ring->bbaddr);
264 265
		err_printf(m, "  BB_STATE: 0x%08x\n", ring->bbstate);
		err_printf(m, "  INSTPS: 0x%08x\n", ring->instps);
266
	}
267
	err_printf(m, "  INSTPM: 0x%08x\n", ring->instpm);
268 269
	err_printf(m, "  FADDR: 0x%08x %08x\n", upper_32_bits(ring->faddr),
		   lower_32_bits(ring->faddr));
270
	if (INTEL_INFO(dev)->gen >= 6) {
271 272
		err_printf(m, "  RC PSMI: 0x%08x\n", ring->rc_psmi);
		err_printf(m, "  FAULT_REG: 0x%08x\n", ring->fault_reg);
273
		err_printf(m, "  SYNC_0: 0x%08x [last synced 0x%08x]\n",
274 275
			   ring->semaphore_mboxes[0],
			   ring->semaphore_seqno[0]);
276
		err_printf(m, "  SYNC_1: 0x%08x [last synced 0x%08x]\n",
277 278
			   ring->semaphore_mboxes[1],
			   ring->semaphore_seqno[1]);
279 280
		if (HAS_VEBOX(dev)) {
			err_printf(m, "  SYNC_2: 0x%08x [last synced 0x%08x]\n",
281 282
				   ring->semaphore_mboxes[2],
				   ring->semaphore_seqno[2]);
283
		}
284
	}
285 286 287 288 289 290 291 292 293 294 295 296 297
	if (USES_PPGTT(dev)) {
		err_printf(m, "  GFX_MODE: 0x%08x\n", ring->vm_info.gfx_mode);

		if (INTEL_INFO(dev)->gen >= 8) {
			int i;
			for (i = 0; i < 4; i++)
				err_printf(m, "  PDP%d: 0x%016llx\n",
					   i, ring->vm_info.pdp[i]);
		} else {
			err_printf(m, "  PP_DIR_BASE: 0x%08x\n",
				   ring->vm_info.pp_dir_base);
		}
	}
298
	err_printf(m, "  seqno: 0x%08x\n", ring->seqno);
299
	err_printf(m, "  last_seqno: 0x%08x\n", ring->last_seqno);
300 301 302
	err_printf(m, "  waiting: %s\n", yesno(ring->waiting));
	err_printf(m, "  ring->head: 0x%08x\n", ring->cpu_ring_head);
	err_printf(m, "  ring->tail: 0x%08x\n", ring->cpu_ring_tail);
303
	err_printf(m, "  hangcheck: %s [%d]\n",
304 305
		   hangcheck_action_to_str(ring->hangcheck_action),
		   ring->hangcheck_score);
306 307 308 309 310 311 312 313 314 315 316
}

void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...)
{
	va_list args;

	va_start(args, f);
	i915_error_vprintf(e, f, args);
	va_end(args);
}

317 318 319 320 321 322 323 324 325 326 327 328 329 330
static void print_error_obj(struct drm_i915_error_state_buf *m,
			    struct drm_i915_error_object *obj)
{
	int page, offset, elt;

	for (page = offset = 0; page < obj->page_count; page++) {
		for (elt = 0; elt < PAGE_SIZE/4; elt++) {
			err_printf(m, "%08x :  %08x\n", offset,
				   obj->pages[page][elt]);
			offset += 4;
		}
	}
}

331 332 333 334
int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
			    const struct i915_error_state_file_priv *error_priv)
{
	struct drm_device *dev = error_priv->dev;
335
	struct drm_i915_private *dev_priv = dev->dev_private;
336
	struct drm_i915_error_state *error = error_priv->error;
337
	struct drm_i915_error_object *obj;
338 339
	int i, j, offset, elt;
	int max_hangcheck_score;
340 341 342 343 344 345

	if (!error) {
		err_printf(m, "no error state collected\n");
		goto out;
	}

346
	err_printf(m, "%s\n", error->error_msg);
347 348 349
	err_printf(m, "Time: %ld s %ld us\n", error->time.tv_sec,
		   error->time.tv_usec);
	err_printf(m, "Kernel: " UTS_RELEASE "\n");
350 351 352 353 354 355 356 357 358 359 360 361 362 363
	max_hangcheck_score = 0;
	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
		if (error->ring[i].hangcheck_score > max_hangcheck_score)
			max_hangcheck_score = error->ring[i].hangcheck_score;
	}
	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
		if (error->ring[i].hangcheck_score == max_hangcheck_score &&
		    error->ring[i].pid != -1) {
			err_printf(m, "Active process (on ring %s): %s [%d]\n",
				   ring_str(i),
				   error->ring[i].comm,
				   error->ring[i].pid);
		}
	}
364
	err_printf(m, "Reset count: %u\n", error->reset_count);
365
	err_printf(m, "Suspend count: %u\n", error->suspend_count);
366
	err_printf(m, "PCI ID: 0x%04x\n", dev->pdev->device);
367 368 369 370
	err_printf(m, "PCI Revision: 0x%02x\n", dev->pdev->revision);
	err_printf(m, "PCI Subsystem: %04x:%04x\n",
		   dev->pdev->subsystem_vendor,
		   dev->pdev->subsystem_device);
371
	err_printf(m, "IOMMU enabled?: %d\n", error->iommu);
372 373 374 375 376 377 378 379 380 381 382

	if (HAS_CSR(dev)) {
		struct intel_csr *csr = &dev_priv->csr;

		err_printf(m, "DMC loaded: %s\n",
			   yesno(csr->dmc_payload != NULL));
		err_printf(m, "DMC fw version: %d.%d\n",
			   CSR_VERSION_MAJOR(csr->version),
			   CSR_VERSION_MINOR(csr->version));
	}

383 384
	err_printf(m, "EIR: 0x%08x\n", error->eir);
	err_printf(m, "IER: 0x%08x\n", error->ier);
385 386 387 388 389 390
	if (INTEL_INFO(dev)->gen >= 8) {
		for (i = 0; i < 4; i++)
			err_printf(m, "GTIER gt %d: 0x%08x\n", i,
				   error->gtier[i]);
	} else if (HAS_PCH_SPLIT(dev) || IS_VALLEYVIEW(dev))
		err_printf(m, "GTIER: 0x%08x\n", error->gtier[0]);
391 392 393 394
	err_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
	err_printf(m, "FORCEWAKE: 0x%08x\n", error->forcewake);
	err_printf(m, "DERRMR: 0x%08x\n", error->derrmr);
	err_printf(m, "CCID: 0x%08x\n", error->ccid);
395
	err_printf(m, "Missed interrupts: 0x%08lx\n", dev_priv->gpu_error.missed_irq_rings);
396 397 398 399 400 401 402 403 404 405

	for (i = 0; i < dev_priv->num_fence_regs; i++)
		err_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);

	for (i = 0; i < ARRAY_SIZE(error->extra_instdone); i++)
		err_printf(m, "  INSTDONE_%d: 0x%08x\n", i,
			   error->extra_instdone[i]);

	if (INTEL_INFO(dev)->gen >= 6) {
		err_printf(m, "ERROR: 0x%08x\n", error->error);
406 407 408 409 410

		if (INTEL_INFO(dev)->gen >= 8)
			err_printf(m, "FAULT_TLB_DATA: 0x%08x 0x%08x\n",
				   error->fault_data1, error->fault_data0);

411 412 413 414 415 416
		err_printf(m, "DONE_REG: 0x%08x\n", error->done_reg);
	}

	if (INTEL_INFO(dev)->gen == 7)
		err_printf(m, "ERR_INT: 0x%08x\n", error->err_int);

417 418
	for (i = 0; i < ARRAY_SIZE(error->ring); i++)
		i915_ring_error_state(m, dev, error, i);
419

420 421 422
	for (i = 0; i < error->vm_count; i++) {
		err_printf(m, "vm[%d]\n", i);

423
		print_error_buffers(m, "Active",
424 425
				    error->active_bo[i],
				    error->active_bo_count[i]);
426 427

		print_error_buffers(m, "Pinned",
428 429 430
				    error->pinned_bo[i],
				    error->pinned_bo_count[i]);
	}
431 432

	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
433 434
		obj = error->ring[i].batchbuffer;
		if (obj) {
435
			err_puts(m, dev_priv->engine[i].name);
436 437 438 439
			if (error->ring[i].pid != -1)
				err_printf(m, " (submitted by %s [%d])",
					   error->ring[i].comm,
					   error->ring[i].pid);
440 441 442
			err_printf(m, " --- gtt_offset = 0x%08x %08x\n",
				   upper_32_bits(obj->gtt_offset),
				   lower_32_bits(obj->gtt_offset));
443 444 445 446 447 448
			print_error_obj(m, obj);
		}

		obj = error->ring[i].wa_batchbuffer;
		if (obj) {
			err_printf(m, "%s (w/a) --- gtt_offset = 0x%08x\n",
449
				   dev_priv->engine[i].name,
450
				   lower_32_bits(obj->gtt_offset));
451
			print_error_obj(m, obj);
452 453 454 455
		}

		if (error->ring[i].num_requests) {
			err_printf(m, "%s --- %d requests\n",
456
				   dev_priv->engine[i].name,
457 458 459 460 461 462 463 464 465 466 467
				   error->ring[i].num_requests);
			for (j = 0; j < error->ring[i].num_requests; j++) {
				err_printf(m, "  seqno 0x%08x, emitted %ld, tail 0x%08x\n",
					   error->ring[i].requests[j].seqno,
					   error->ring[i].requests[j].jiffies,
					   error->ring[i].requests[j].tail);
			}
		}

		if ((obj = error->ring[i].ringbuffer)) {
			err_printf(m, "%s --- ringbuffer = 0x%08x\n",
468
				   dev_priv->engine[i].name,
469
				   lower_32_bits(obj->gtt_offset));
470
			print_error_obj(m, obj);
471 472
		}

473
		if ((obj = error->ring[i].hws_page)) {
474 475 476 477 478 479 480
			u64 hws_offset = obj->gtt_offset;
			u32 *hws_page = &obj->pages[0][0];

			if (i915.enable_execlists) {
				hws_offset += LRC_PPHWSP_PN * PAGE_SIZE;
				hws_page = &obj->pages[LRC_PPHWSP_PN][0];
			}
481
			err_printf(m, "%s --- HW Status = 0x%08llx\n",
482
				   dev_priv->engine[i].name, hws_offset);
483 484 485 486
			offset = 0;
			for (elt = 0; elt < PAGE_SIZE/16; elt += 4) {
				err_printf(m, "[%04x] %08x %08x %08x %08x\n",
					   offset,
487 488 489 490
					   hws_page[elt],
					   hws_page[elt+1],
					   hws_page[elt+2],
					   hws_page[elt+3]);
491 492 493 494
					offset += 16;
			}
		}

495 496 497 498
		obj = error->ring[i].wa_ctx;
		if (obj) {
			u64 wa_ctx_offset = obj->gtt_offset;
			u32 *wa_ctx_page = &obj->pages[0][0];
499
			struct intel_engine_cs *engine = &dev_priv->engine[RCS];
500 501
			u32 wa_ctx_size = (engine->wa_ctx.indirect_ctx.size +
					   engine->wa_ctx.per_ctx.size);
502 503

			err_printf(m, "%s --- WA ctx batch buffer = 0x%08llx\n",
504
				   dev_priv->engine[i].name, wa_ctx_offset);
505 506 507 508 509 510 511 512 513 514 515 516
			offset = 0;
			for (elt = 0; elt < wa_ctx_size; elt += 4) {
				err_printf(m, "[%04x] %08x %08x %08x %08x\n",
					   offset,
					   wa_ctx_page[elt + 0],
					   wa_ctx_page[elt + 1],
					   wa_ctx_page[elt + 2],
					   wa_ctx_page[elt + 3]);
				offset += 16;
			}
		}

517
		if ((obj = error->ring[i].ctx)) {
518
			err_printf(m, "%s --- HW Context = 0x%08x\n",
519
				   dev_priv->engine[i].name,
520
				   lower_32_bits(obj->gtt_offset));
521
			print_error_obj(m, obj);
522 523 524
		}
	}

525
	if ((obj = error->semaphore_obj)) {
526 527
		err_printf(m, "Semaphore page = 0x%08x\n",
			   lower_32_bits(obj->gtt_offset));
528 529 530 531 532 533 534 535 536 537
		for (elt = 0; elt < PAGE_SIZE/16; elt += 4) {
			err_printf(m, "[%04x] %08x %08x %08x %08x\n",
				   elt * 4,
				   obj->pages[0][elt],
				   obj->pages[0][elt+1],
				   obj->pages[0][elt+2],
				   obj->pages[0][elt+3]);
		}
	}

538 539 540 541 542 543 544 545 546 547 548 549 550 551
	if (error->overlay)
		intel_overlay_print_error_state(m, error->overlay);

	if (error->display)
		intel_display_print_error_state(m, dev, error->display);

out:
	if (m->bytes == 0 && m->err)
		return m->err;

	return 0;
}

int i915_error_state_buf_init(struct drm_i915_error_state_buf *ebuf,
552
			      struct drm_i915_private *i915,
553 554 555
			      size_t count, loff_t pos)
{
	memset(ebuf, 0, sizeof(*ebuf));
556
	ebuf->i915 = i915;
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

	/* We need to have enough room to store any i915_error_state printf
	 * so that we can move it to start position.
	 */
	ebuf->size = count + 1 > PAGE_SIZE ? count + 1 : PAGE_SIZE;
	ebuf->buf = kmalloc(ebuf->size,
				GFP_TEMPORARY | __GFP_NORETRY | __GFP_NOWARN);

	if (ebuf->buf == NULL) {
		ebuf->size = PAGE_SIZE;
		ebuf->buf = kmalloc(ebuf->size, GFP_TEMPORARY);
	}

	if (ebuf->buf == NULL) {
		ebuf->size = 128;
		ebuf->buf = kmalloc(ebuf->size, GFP_TEMPORARY);
	}

	if (ebuf->buf == NULL)
		return -ENOMEM;

	ebuf->start = pos;

	return 0;
}

static void i915_error_object_free(struct drm_i915_error_object *obj)
{
	int page;

	if (obj == NULL)
		return;

	for (page = 0; page < obj->page_count; page++)
		kfree(obj->pages[page]);

	kfree(obj);
}

static void i915_error_state_free(struct kref *error_ref)
{
	struct drm_i915_error_state *error = container_of(error_ref,
							  typeof(*error), ref);
	int i;

	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
		i915_error_object_free(error->ring[i].batchbuffer);
604
		i915_error_object_free(error->ring[i].wa_batchbuffer);
605
		i915_error_object_free(error->ring[i].ringbuffer);
606
		i915_error_object_free(error->ring[i].hws_page);
607 608
		i915_error_object_free(error->ring[i].ctx);
		kfree(error->ring[i].requests);
609
		i915_error_object_free(error->ring[i].wa_ctx);
610 611
	}

612
	i915_error_object_free(error->semaphore_obj);
613 614 615 616

	for (i = 0; i < error->vm_count; i++)
		kfree(error->active_bo[i]);

617
	kfree(error->active_bo);
618 619 620
	kfree(error->active_bo_count);
	kfree(error->pinned_bo);
	kfree(error->pinned_bo_count);
621 622 623 624 625 626
	kfree(error->overlay);
	kfree(error->display);
	kfree(error);
}

static struct drm_i915_error_object *
627 628 629
i915_error_object_create(struct drm_i915_private *dev_priv,
			 struct drm_i915_gem_object *src,
			 struct i915_address_space *vm)
630
{
631
	struct i915_ggtt *ggtt = &dev_priv->ggtt;
632
	struct drm_i915_error_object *dst;
633
	struct i915_vma *vma = NULL;
634
	int num_pages;
635 636
	bool use_ggtt;
	int i = 0;
637
	u64 reloc_offset;
638 639 640 641

	if (src == NULL || src->pages == NULL)
		return NULL;

642 643
	num_pages = src->base.size >> PAGE_SHIFT;

644 645 646 647
	dst = kmalloc(sizeof(*dst) + num_pages * sizeof(u32 *), GFP_ATOMIC);
	if (dst == NULL)
		return NULL;

648 649 650 651
	if (i915_gem_obj_bound(src, vm))
		dst->gtt_offset = i915_gem_obj_offset(src, vm);
	else
		dst->gtt_offset = -1;
652 653

	reloc_offset = dst->gtt_offset;
654 655
	if (i915_is_ggtt(vm))
		vma = i915_gem_obj_to_ggtt(src);
656
	use_ggtt = (src->cache_level == I915_CACHE_NONE &&
657
		   vma && (vma->bound & GLOBAL_BIND) &&
658
		   reloc_offset + num_pages * PAGE_SIZE <= ggtt->mappable_end);
659 660 661 662 663

	/* Cannot access stolen address directly, try to use the aperture */
	if (src->stolen) {
		use_ggtt = true;

664
		if (!(vma && vma->bound & GLOBAL_BIND))
665 666 667
			goto unwind;

		reloc_offset = i915_gem_obj_ggtt_offset(src);
668
		if (reloc_offset + num_pages * PAGE_SIZE > ggtt->mappable_end)
669 670 671 672
			goto unwind;
	}

	/* Cannot access snooped pages through the aperture */
673 674
	if (use_ggtt && src->cache_level != I915_CACHE_NONE &&
	    !HAS_LLC(dev_priv))
675 676 677 678
		goto unwind;

	dst->page_count = num_pages;
	while (num_pages--) {
679 680 681 682 683 684 685 686
		unsigned long flags;
		void *d;

		d = kmalloc(PAGE_SIZE, GFP_ATOMIC);
		if (d == NULL)
			goto unwind;

		local_irq_save(flags);
687
		if (use_ggtt) {
688 689 690 691 692 693 694
			void __iomem *s;

			/* Simply ignore tiling or any overlapping fence.
			 * It's part of the error state, and this hopefully
			 * captures what the GPU read.
			 */

695
			s = io_mapping_map_atomic_wc(ggtt->mappable,
696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714
						     reloc_offset);
			memcpy_fromio(d, s, PAGE_SIZE);
			io_mapping_unmap_atomic(s);
		} else {
			struct page *page;
			void *s;

			page = i915_gem_object_get_page(src, i);

			drm_clflush_pages(&page, 1);

			s = kmap_atomic(page);
			memcpy(d, s, PAGE_SIZE);
			kunmap_atomic(s);

			drm_clflush_pages(&page, 1);
		}
		local_irq_restore(flags);

715
		dst->pages[i++] = d;
716 717 718 719 720 721 722 723 724 725 726
		reloc_offset += PAGE_SIZE;
	}

	return dst;

unwind:
	while (i--)
		kfree(dst->pages[i]);
	kfree(dst);
	return NULL;
}
727
#define i915_error_ggtt_object_create(dev_priv, src) \
728
	i915_error_object_create((dev_priv), (src), &(dev_priv)->ggtt.base)
729 730

static void capture_bo(struct drm_i915_error_buffer *err,
731
		       struct i915_vma *vma)
732
{
733
	struct drm_i915_gem_object *obj = vma->obj;
734
	int i;
735

736 737
	err->size = obj->base.size;
	err->name = obj->base.name;
738
	for (i = 0; i < I915_NUM_ENGINES; i++)
739
		err->rseqno[i] = i915_gem_request_get_seqno(obj->last_read_req[i]);
740
	err->wseqno = i915_gem_request_get_seqno(obj->last_write_req);
741
	err->gtt_offset = vma->node.start;
742 743 744 745
	err->read_domains = obj->base.read_domains;
	err->write_domain = obj->base.write_domain;
	err->fence_reg = obj->fence_reg;
	err->pinned = 0;
B
Ben Widawsky 已提交
746
	if (i915_gem_obj_is_pinned(obj))
747 748 749 750
		err->pinned = 1;
	err->tiling = obj->tiling_mode;
	err->dirty = obj->dirty;
	err->purgeable = obj->madv != I915_MADV_WILLNEED;
751
	err->userptr = obj->userptr.mm != NULL;
752
	err->ring = obj->last_write_req ?
753
			i915_gem_request_get_engine(obj->last_write_req)->id : -1;
754 755 756 757 758 759
	err->cache_level = obj->cache_level;
}

static u32 capture_active_bo(struct drm_i915_error_buffer *err,
			     int count, struct list_head *head)
{
B
Ben Widawsky 已提交
760
	struct i915_vma *vma;
761 762
	int i = 0;

763
	list_for_each_entry(vma, head, vm_link) {
764
		capture_bo(err++, vma);
765 766 767 768 769 770 771 772
		if (++i == count)
			break;
	}

	return i;
}

static u32 capture_pinned_bo(struct drm_i915_error_buffer *err,
773 774
			     int count, struct list_head *head,
			     struct i915_address_space *vm)
775 776
{
	struct drm_i915_gem_object *obj;
777 778
	struct drm_i915_error_buffer * const first = err;
	struct drm_i915_error_buffer * const last = err + count;
779 780

	list_for_each_entry(obj, head, global_list) {
781
		struct i915_vma *vma;
782

783
		if (err == last)
784
			break;
785

786
		list_for_each_entry(vma, &obj->vma_list, obj_link)
787
			if (vma->vm == vm && vma->pin_count > 0)
788
				capture_bo(err++, vma);
789 790
	}

791
	return err - first;
792 793
}

794 795 796 797 798 799 800 801 802 803
/* Generate a semi-unique error code. The code is not meant to have meaning, The
 * code's only purpose is to try to prevent false duplicated bug reports by
 * grossly estimating a GPU error state.
 *
 * TODO Ideally, hashing the batchbuffer would be a very nice way to determine
 * the hang if we could strip the GTT offset information from it.
 *
 * It's only a small step better than a random number in its current form.
 */
static uint32_t i915_error_generate_code(struct drm_i915_private *dev_priv,
804 805
					 struct drm_i915_error_state *error,
					 int *ring_id)
806 807 808 809 810 811 812 813 814
{
	uint32_t error_code = 0;
	int i;

	/* IPEHR would be an ideal way to detect errors, as it's the gross
	 * measure of "the command that hung." However, has some very common
	 * synchronization commands which almost always appear in the case
	 * strictly a client bug. Use instdone to differentiate those some.
	 */
815
	for (i = 0; i < I915_NUM_ENGINES; i++) {
816 817 818 819
		if (error->ring[i].hangcheck_action == HANGCHECK_HUNG) {
			if (ring_id)
				*ring_id = i;

820
			return error->ring[i].ipehr ^ error->ring[i].instdone;
821 822
		}
	}
823 824 825 826

	return error_code;
}

827 828 829 830 831 832
static void i915_gem_record_fences(struct drm_device *dev,
				   struct drm_i915_error_state *error)
{
	struct drm_i915_private *dev_priv = dev->dev_private;
	int i;

833 834
	if (IS_GEN3(dev) || IS_GEN2(dev)) {
		for (i = 0; i < dev_priv->num_fence_regs; i++)
835 836 837 838 839 840 841 842
			error->fence[i] = I915_READ(FENCE_REG(i));
	} else if (IS_GEN5(dev) || IS_GEN4(dev)) {
		for (i = 0; i < dev_priv->num_fence_regs; i++)
			error->fence[i] = I915_READ64(FENCE_REG_965_LO(i));
	} else if (INTEL_INFO(dev)->gen >= 6) {
		for (i = 0; i < dev_priv->num_fence_regs; i++)
			error->fence[i] = I915_READ64(FENCE_REG_GEN6_LO(i));
	}
843 844
}

845

846 847
static void gen8_record_semaphore_state(struct drm_i915_private *dev_priv,
					struct drm_i915_error_state *error,
848
					struct intel_engine_cs *engine,
849 850
					struct drm_i915_error_ring *ering)
{
851
	struct intel_engine_cs *to;
852
	enum intel_engine_id id;
853 854 855 856 857 858

	if (!i915_semaphore_is_enabled(dev_priv->dev))
		return;

	if (!error->semaphore_obj)
		error->semaphore_obj =
859 860
			i915_error_ggtt_object_create(dev_priv,
						      dev_priv->semaphore_obj);
861

862
	for_each_engine_id(to, dev_priv, id) {
863 864 865
		int idx;
		u16 signal_offset;
		u32 *tmp;
866

867
		if (engine == to)
868 869
			continue;

870
		signal_offset = (GEN8_SIGNAL_OFFSET(engine, id) & (PAGE_SIZE - 1))
871
				/ 4;
872
		tmp = error->semaphore_obj->pages[0];
873
		idx = intel_ring_sync_index(engine, to);
874 875

		ering->semaphore_mboxes[idx] = tmp[signal_offset];
876
		ering->semaphore_seqno[idx] = engine->semaphore.sync_seqno[idx];
877 878 879
	}
}

880
static void gen6_record_semaphore_state(struct drm_i915_private *dev_priv,
881
					struct intel_engine_cs *engine,
882 883
					struct drm_i915_error_ring *ering)
{
884 885 886 887
	ering->semaphore_mboxes[0] = I915_READ(RING_SYNC_0(engine->mmio_base));
	ering->semaphore_mboxes[1] = I915_READ(RING_SYNC_1(engine->mmio_base));
	ering->semaphore_seqno[0] = engine->semaphore.sync_seqno[0];
	ering->semaphore_seqno[1] = engine->semaphore.sync_seqno[1];
888

889
	if (HAS_VEBOX(dev_priv)) {
890
		ering->semaphore_mboxes[2] =
891 892
			I915_READ(RING_SYNC_2(engine->mmio_base));
		ering->semaphore_seqno[2] = engine->semaphore.sync_seqno[2];
893 894 895
	}
}

896
static void i915_record_ring_state(struct drm_device *dev,
897
				   struct drm_i915_error_state *error,
898
				   struct intel_engine_cs *engine,
899
				   struct drm_i915_error_ring *ering)
900 901 902 903
{
	struct drm_i915_private *dev_priv = dev->dev_private;

	if (INTEL_INFO(dev)->gen >= 6) {
904 905
		ering->rc_psmi = I915_READ(RING_PSMI_CTL(engine->mmio_base));
		ering->fault_reg = I915_READ(RING_FAULT_REG(engine));
906
		if (INTEL_INFO(dev)->gen >= 8)
907 908
			gen8_record_semaphore_state(dev_priv, error, engine,
						    ering);
909
		else
910
			gen6_record_semaphore_state(dev_priv, engine, ering);
911 912
	}

913
	if (INTEL_INFO(dev)->gen >= 4) {
914 915 916 917 918 919
		ering->faddr = I915_READ(RING_DMA_FADD(engine->mmio_base));
		ering->ipeir = I915_READ(RING_IPEIR(engine->mmio_base));
		ering->ipehr = I915_READ(RING_IPEHR(engine->mmio_base));
		ering->instdone = I915_READ(RING_INSTDONE(engine->mmio_base));
		ering->instps = I915_READ(RING_INSTPS(engine->mmio_base));
		ering->bbaddr = I915_READ(RING_BBADDR(engine->mmio_base));
920
		if (INTEL_INFO(dev)->gen >= 8) {
921 922
			ering->faddr |= (u64) I915_READ(RING_DMA_FADD_UDW(engine->mmio_base)) << 32;
			ering->bbaddr |= (u64) I915_READ(RING_BBADDR_UDW(engine->mmio_base)) << 32;
923
		}
924
		ering->bbstate = I915_READ(RING_BBSTATE(engine->mmio_base));
925
	} else {
926 927 928
		ering->faddr = I915_READ(DMA_FADD_I8XX);
		ering->ipeir = I915_READ(IPEIR);
		ering->ipehr = I915_READ(IPEHR);
929
		ering->instdone = I915_READ(GEN2_INSTDONE);
930 931
	}

932 933 934
	ering->waiting = waitqueue_active(&engine->irq_queue);
	ering->instpm = I915_READ(RING_INSTPM(engine->mmio_base));
	ering->acthd = intel_ring_get_active_head(engine);
935
	ering->seqno = engine->get_seqno(engine);
936
	ering->last_seqno = engine->last_submitted_seqno;
937 938 939 940
	ering->start = I915_READ_START(engine);
	ering->head = I915_READ_HEAD(engine);
	ering->tail = I915_READ_TAIL(engine);
	ering->ctl = I915_READ_CTL(engine);
941

942
	if (I915_NEED_GFX_HWS(dev)) {
943
		i915_reg_t mmio;
944 945

		if (IS_GEN7(dev)) {
946
			switch (engine->id) {
947 948 949 950 951 952 953 954 955 956 957 958 959 960
			default:
			case RCS:
				mmio = RENDER_HWS_PGA_GEN7;
				break;
			case BCS:
				mmio = BLT_HWS_PGA_GEN7;
				break;
			case VCS:
				mmio = BSD_HWS_PGA_GEN7;
				break;
			case VECS:
				mmio = VEBOX_HWS_PGA_GEN7;
				break;
			}
961 962
		} else if (IS_GEN6(engine->dev)) {
			mmio = RING_HWS_PGA_GEN6(engine->mmio_base);
963 964
		} else {
			/* XXX: gen8 returns to sanity */
965
			mmio = RING_HWS_PGA(engine->mmio_base);
966 967
		}

968
		ering->hws = I915_READ(mmio);
969 970
	}

971 972
	ering->hangcheck_score = engine->hangcheck.score;
	ering->hangcheck_action = engine->hangcheck.action;
973 974 975 976

	if (USES_PPGTT(dev)) {
		int i;

977
		ering->vm_info.gfx_mode = I915_READ(RING_MODE_GEN7(engine));
978

979 980
		if (IS_GEN6(dev))
			ering->vm_info.pp_dir_base =
981
				I915_READ(RING_PP_DIR_BASE_READ(engine));
982 983
		else if (IS_GEN7(dev))
			ering->vm_info.pp_dir_base =
984
				I915_READ(RING_PP_DIR_BASE(engine));
985
		else if (INTEL_INFO(dev)->gen >= 8)
986 987
			for (i = 0; i < 4; i++) {
				ering->vm_info.pdp[i] =
988
					I915_READ(GEN8_RING_PDP_UDW(engine, i));
989 990
				ering->vm_info.pdp[i] <<= 32;
				ering->vm_info.pdp[i] |=
991
					I915_READ(GEN8_RING_PDP_LDW(engine, i));
992 993
			}
	}
994 995 996
}


997
static void i915_gem_record_active_context(struct intel_engine_cs *engine,
998 999 1000
					   struct drm_i915_error_state *error,
					   struct drm_i915_error_ring *ering)
{
1001
	struct drm_i915_private *dev_priv = engine->dev->dev_private;
1002 1003 1004
	struct drm_i915_gem_object *obj;

	/* Currently render ring is the only HW context user */
1005
	if (engine->id != RCS || !error->ccid)
1006 1007 1008
		return;

	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
1009 1010 1011
		if (!i915_gem_obj_ggtt_bound(obj))
			continue;

1012
		if ((error->ccid & PAGE_MASK) == i915_gem_obj_ggtt_offset(obj)) {
1013
			ering->ctx = i915_error_ggtt_object_create(dev_priv, obj);
1014 1015 1016 1017 1018 1019 1020 1021
			break;
		}
	}
}

static void i915_gem_record_rings(struct drm_device *dev,
				  struct drm_i915_error_state *error)
{
1022 1023
	struct drm_i915_private *dev_priv = to_i915(dev);
	struct i915_ggtt *ggtt = &dev_priv->ggtt;
1024 1025 1026
	struct drm_i915_gem_request *request;
	int i, count;

1027
	for (i = 0; i < I915_NUM_ENGINES; i++) {
1028
		struct intel_engine_cs *engine = &dev_priv->engine[i];
1029
		struct intel_ringbuffer *rbuf;
1030

1031 1032
		error->ring[i].pid = -1;

1033
		if (engine->dev == NULL)
1034 1035 1036 1037
			continue;

		error->ring[i].valid = true;

1038
		i915_record_ring_state(dev, error, engine, &error->ring[i]);
1039

1040
		request = i915_gem_find_active_request(engine);
1041
		if (request) {
1042 1043 1044 1045
			struct i915_address_space *vm;

			vm = request->ctx && request->ctx->ppgtt ?
				&request->ctx->ppgtt->base :
1046
				&ggtt->base;
1047

1048 1049 1050 1051 1052 1053 1054
			/* We need to copy these to an anonymous buffer
			 * as the simplest method to avoid being overwritten
			 * by userspace.
			 */
			error->ring[i].batchbuffer =
				i915_error_object_create(dev_priv,
							 request->batch_obj,
1055
							 vm);
1056

1057
			if (HAS_BROKEN_CS_TLB(dev_priv))
1058 1059
				error->ring[i].wa_batchbuffer =
					i915_error_ggtt_object_create(dev_priv,
1060
							     engine->scratch.obj);
1061

1062
			if (request->pid) {
1063 1064 1065
				struct task_struct *task;

				rcu_read_lock();
1066
				task = pid_task(request->pid, PIDTYPE_PID);
1067 1068 1069 1070 1071 1072 1073
				if (task) {
					strcpy(error->ring[i].comm, task->comm);
					error->ring[i].pid = task->pid;
				}
				rcu_read_unlock();
			}
		}
1074

1075 1076 1077 1078 1079 1080 1081
		if (i915.enable_execlists) {
			/* TODO: This is only a small fix to keep basic error
			 * capture working, but we need to add more information
			 * for it to be useful (e.g. dump the context being
			 * executed).
			 */
			if (request)
1082
				rbuf = request->ctx->engine[engine->id].ringbuf;
1083
			else
1084
				rbuf = dev_priv->kernel_context->engine[engine->id].ringbuf;
1085
		} else
1086
			rbuf = engine->buffer;
1087 1088 1089 1090

		error->ring[i].cpu_ring_head = rbuf->head;
		error->ring[i].cpu_ring_tail = rbuf->tail;

1091
		error->ring[i].ringbuffer =
1092
			i915_error_ggtt_object_create(dev_priv, rbuf->obj);
1093

1094
		error->ring[i].hws_page =
1095 1096
			i915_error_ggtt_object_create(dev_priv,
						      engine->status_page.obj);
1097

1098
		if (engine->wa_ctx.obj) {
1099 1100
			error->ring[i].wa_ctx =
				i915_error_ggtt_object_create(dev_priv,
1101
							      engine->wa_ctx.obj);
1102 1103
		}

1104
		i915_gem_record_active_context(engine, error, &error->ring[i]);
1105 1106

		count = 0;
1107
		list_for_each_entry(request, &engine->request_list, list)
1108 1109 1110 1111
			count++;

		error->ring[i].num_requests = count;
		error->ring[i].requests =
D
Daniel Vetter 已提交
1112
			kcalloc(count, sizeof(*error->ring[i].requests),
1113 1114 1115 1116 1117 1118 1119
				GFP_ATOMIC);
		if (error->ring[i].requests == NULL) {
			error->ring[i].num_requests = 0;
			continue;
		}

		count = 0;
1120
		list_for_each_entry(request, &engine->request_list, list) {
1121 1122
			struct drm_i915_error_request *erq;

1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141
			if (count >= error->ring[i].num_requests) {
				/*
				 * If the ring request list was changed in
				 * between the point where the error request
				 * list was created and dimensioned and this
				 * point then just exit early to avoid crashes.
				 *
				 * We don't need to communicate that the
				 * request list changed state during error
				 * state capture and that the error state is
				 * slightly incorrect as a consequence since we
				 * are typically only interested in the request
				 * list state at the point of error state
				 * capture, not in any changes happening during
				 * the capture.
				 */
				break;
			}

1142 1143 1144
			erq = &error->ring[i].requests[count++];
			erq->seqno = request->seqno;
			erq->jiffies = request->emitted_jiffies;
1145
			erq->tail = request->postfix;
1146 1147 1148 1149
		}
	}
}

1150 1151 1152 1153 1154 1155 1156
/* FIXME: Since pin count/bound list is global, we duplicate what we capture per
 * VM.
 */
static void i915_gem_capture_vm(struct drm_i915_private *dev_priv,
				struct drm_i915_error_state *error,
				struct i915_address_space *vm,
				const int ndx)
1157
{
1158
	struct drm_i915_error_buffer *active_bo = NULL, *pinned_bo = NULL;
1159
	struct drm_i915_gem_object *obj;
1160
	struct i915_vma *vma;
1161 1162 1163
	int i;

	i = 0;
1164
	list_for_each_entry(vma, &vm->active_list, vm_link)
1165
		i++;
1166
	error->active_bo_count[ndx] = i;
1167 1168

	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
1169
		list_for_each_entry(vma, &obj->vma_list, obj_link)
1170
			if (vma->vm == vm && vma->pin_count > 0)
1171 1172
				i++;
	}
1173
	error->pinned_bo_count[ndx] = i - error->active_bo_count[ndx];
1174 1175

	if (i) {
D
Daniel Vetter 已提交
1176
		active_bo = kcalloc(i, sizeof(*active_bo), GFP_ATOMIC);
1177 1178
		if (active_bo)
			pinned_bo = active_bo + error->active_bo_count[ndx];
1179 1180
	}

1181 1182 1183 1184
	if (active_bo)
		error->active_bo_count[ndx] =
			capture_active_bo(active_bo,
					  error->active_bo_count[ndx],
1185
					  &vm->active_list);
1186

1187 1188 1189 1190
	if (pinned_bo)
		error->pinned_bo_count[ndx] =
			capture_pinned_bo(pinned_bo,
					  error->pinned_bo_count[ndx],
1191
					  &dev_priv->mm.bound_list, vm);
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211
	error->active_bo[ndx] = active_bo;
	error->pinned_bo[ndx] = pinned_bo;
}

static void i915_gem_capture_buffers(struct drm_i915_private *dev_priv,
				     struct drm_i915_error_state *error)
{
	struct i915_address_space *vm;
	int cnt = 0, i = 0;

	list_for_each_entry(vm, &dev_priv->vm_list, global_link)
		cnt++;

	error->active_bo = kcalloc(cnt, sizeof(*error->active_bo), GFP_ATOMIC);
	error->pinned_bo = kcalloc(cnt, sizeof(*error->pinned_bo), GFP_ATOMIC);
	error->active_bo_count = kcalloc(cnt, sizeof(*error->active_bo_count),
					 GFP_ATOMIC);
	error->pinned_bo_count = kcalloc(cnt, sizeof(*error->pinned_bo_count),
					 GFP_ATOMIC);

1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230
	if (error->active_bo == NULL ||
	    error->pinned_bo == NULL ||
	    error->active_bo_count == NULL ||
	    error->pinned_bo_count == NULL) {
		kfree(error->active_bo);
		kfree(error->active_bo_count);
		kfree(error->pinned_bo);
		kfree(error->pinned_bo_count);

		error->active_bo = NULL;
		error->active_bo_count = NULL;
		error->pinned_bo = NULL;
		error->pinned_bo_count = NULL;
	} else {
		list_for_each_entry(vm, &dev_priv->vm_list, global_link)
			i915_gem_capture_vm(dev_priv, error, vm, i++);

		error->vm_count = cnt;
	}
1231 1232
}

1233 1234 1235
/* Capture all registers which don't fit into another category. */
static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
				   struct drm_i915_error_state *error)
1236
{
1237
	struct drm_device *dev = dev_priv->dev;
1238
	int i;
1239

1240 1241 1242 1243 1244 1245 1246
	/* General organization
	 * 1. Registers specific to a single generation
	 * 2. Registers which belong to multiple generations
	 * 3. Feature specific registers.
	 * 4. Everything else
	 * Please try to follow the order.
	 */
1247

1248 1249
	/* 1: Registers specific to a single generation */
	if (IS_VALLEYVIEW(dev)) {
1250
		error->gtier[0] = I915_READ(GTIER);
1251
		error->ier = I915_READ(VLV_IER);
1252
		error->forcewake = I915_READ_FW(FORCEWAKE_VLV);
1253
	}
1254

1255 1256
	if (IS_GEN7(dev))
		error->err_int = I915_READ(GEN7_ERR_INT);
1257

1258 1259 1260 1261 1262
	if (INTEL_INFO(dev)->gen >= 8) {
		error->fault_data0 = I915_READ(GEN8_FAULT_TLB_DATA0);
		error->fault_data1 = I915_READ(GEN8_FAULT_TLB_DATA1);
	}

1263
	if (IS_GEN6(dev)) {
1264
		error->forcewake = I915_READ_FW(FORCEWAKE);
1265 1266 1267
		error->gab_ctl = I915_READ(GAB_CTL);
		error->gfx_mode = I915_READ(GFX_MODE);
	}
1268

1269 1270
	/* 2: Registers which belong to multiple generations */
	if (INTEL_INFO(dev)->gen >= 7)
1271
		error->forcewake = I915_READ_FW(FORCEWAKE_MT);
1272 1273

	if (INTEL_INFO(dev)->gen >= 6) {
1274
		error->derrmr = I915_READ(DERRMR);
1275 1276 1277 1278
		error->error = I915_READ(ERROR_GEN6);
		error->done_reg = I915_READ(DONE_REG);
	}

1279
	/* 3: Feature specific registers */
1280 1281 1282 1283 1284 1285
	if (IS_GEN6(dev) || IS_GEN7(dev)) {
		error->gam_ecochk = I915_READ(GAM_ECOCHK);
		error->gac_eco = I915_READ(GAC_ECO_BITS);
	}

	/* 4: Everything else */
1286 1287 1288
	if (HAS_HW_CONTEXTS(dev))
		error->ccid = I915_READ(CCID);

1289 1290 1291 1292 1293
	if (INTEL_INFO(dev)->gen >= 8) {
		error->ier = I915_READ(GEN8_DE_MISC_IER);
		for (i = 0; i < 4; i++)
			error->gtier[i] = I915_READ(GEN8_GT_IER(i));
	} else if (HAS_PCH_SPLIT(dev)) {
1294
		error->ier = I915_READ(DEIER);
1295
		error->gtier[0] = I915_READ(GTIER);
1296 1297 1298 1299
	} else if (IS_GEN2(dev)) {
		error->ier = I915_READ16(IER);
	} else if (!IS_VALLEYVIEW(dev)) {
		error->ier = I915_READ(IER);
1300 1301 1302
	}
	error->eir = I915_READ(EIR);
	error->pgtbl_er = I915_READ(PGTBL_ER);
1303 1304

	i915_get_extra_instdone(dev, error->extra_instdone);
1305 1306
}

1307
static void i915_error_capture_msg(struct drm_device *dev,
1308
				   struct drm_i915_error_state *error,
1309
				   u32 engine_mask,
1310
				   const char *error_msg)
1311 1312 1313
{
	struct drm_i915_private *dev_priv = dev->dev_private;
	u32 ecode;
1314
	int ring_id = -1, len;
1315 1316 1317

	ecode = i915_error_generate_code(dev_priv, error, &ring_id);

1318
	len = scnprintf(error->error_msg, sizeof(error->error_msg),
1319 1320
			"GPU HANG: ecode %d:%d:0x%08x",
			INTEL_INFO(dev)->gen, ring_id, ecode);
1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331

	if (ring_id != -1 && error->ring[ring_id].pid != -1)
		len += scnprintf(error->error_msg + len,
				 sizeof(error->error_msg) - len,
				 ", in %s [%d]",
				 error->ring[ring_id].comm,
				 error->ring[ring_id].pid);

	scnprintf(error->error_msg + len, sizeof(error->error_msg) - len,
		  ", reason: %s, action: %s",
		  error_msg,
1332
		  engine_mask ? "reset" : "continue");
1333 1334
}

1335 1336 1337
static void i915_capture_gen_state(struct drm_i915_private *dev_priv,
				   struct drm_i915_error_state *error)
{
1338 1339 1340 1341
	error->iommu = -1;
#ifdef CONFIG_INTEL_IOMMU
	error->iommu = intel_iommu_gfx_mapped;
#endif
1342
	error->reset_count = i915_reset_count(&dev_priv->gpu_error);
1343
	error->suspend_count = dev_priv->suspend_count;
1344 1345
}

1346 1347 1348 1349 1350 1351 1352 1353 1354
/**
 * i915_capture_error_state - capture an error record for later analysis
 * @dev: drm device
 *
 * Should be called when an error is detected (either a hang or an error
 * interrupt) to capture error state from the time of the error.  Fills
 * out a structure which becomes available in debugfs for user level tools
 * to pick up.
 */
1355
void i915_capture_error_state(struct drm_device *dev, u32 engine_mask,
1356
			      const char *error_msg)
1357
{
1358
	static bool warned;
1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369
	struct drm_i915_private *dev_priv = dev->dev_private;
	struct drm_i915_error_state *error;
	unsigned long flags;

	/* Account for pipe specific data like PIPE*STAT */
	error = kzalloc(sizeof(*error), GFP_ATOMIC);
	if (!error) {
		DRM_DEBUG_DRIVER("out of memory, not capturing error state\n");
		return;
	}

1370 1371
	kref_init(&error->ref);

1372
	i915_capture_gen_state(dev_priv, error);
1373 1374 1375 1376
	i915_capture_reg_state(dev_priv, error);
	i915_gem_capture_buffers(dev_priv, error);
	i915_gem_record_fences(dev, error);
	i915_gem_record_rings(dev, error);
1377

1378 1379 1380 1381 1382
	do_gettimeofday(&error->time);

	error->overlay = intel_overlay_capture_error_state(dev);
	error->display = intel_display_capture_error_state(dev);

1383
	i915_error_capture_msg(dev, error, engine_mask, error_msg);
1384 1385
	DRM_INFO("%s\n", error->error_msg);

1386 1387 1388 1389 1390 1391 1392
	spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
	if (dev_priv->gpu_error.first_error == NULL) {
		dev_priv->gpu_error.first_error = error;
		error = NULL;
	}
	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);

1393
	if (error) {
1394
		i915_error_state_free(&error->ref);
1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405
		return;
	}

	if (!warned) {
		DRM_INFO("GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.\n");
		DRM_INFO("Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel\n");
		DRM_INFO("drm/i915 developers can then reassign to the right component if it's not a kernel issue.\n");
		DRM_INFO("The gpu crash dump is required to analyze gpu hangs, so please always attach it.\n");
		DRM_INFO("GPU crash dump saved to /sys/class/drm/card%d/error\n", dev->primary->index);
		warned = true;
	}
1406 1407 1408 1409 1410 1411 1412
}

void i915_error_state_get(struct drm_device *dev,
			  struct i915_error_state_file_priv *error_priv)
{
	struct drm_i915_private *dev_priv = dev->dev_private;

1413
	spin_lock_irq(&dev_priv->gpu_error.lock);
1414 1415 1416
	error_priv->error = dev_priv->gpu_error.first_error;
	if (error_priv->error)
		kref_get(&error_priv->error->ref);
1417
	spin_unlock_irq(&dev_priv->gpu_error.lock);
1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431

}

void i915_error_state_put(struct i915_error_state_file_priv *error_priv)
{
	if (error_priv->error)
		kref_put(&error_priv->error->ref, i915_error_state_free);
}

void i915_destroy_error_state(struct drm_device *dev)
{
	struct drm_i915_private *dev_priv = dev->dev_private;
	struct drm_i915_error_state *error;

1432
	spin_lock_irq(&dev_priv->gpu_error.lock);
1433 1434
	error = dev_priv->gpu_error.first_error;
	dev_priv->gpu_error.first_error = NULL;
1435
	spin_unlock_irq(&dev_priv->gpu_error.lock);
1436 1437 1438 1439 1440

	if (error)
		kref_put(&error->ref, i915_error_state_free);
}

1441
const char *i915_cache_level_str(struct drm_i915_private *i915, int type)
1442 1443 1444
{
	switch (type) {
	case I915_CACHE_NONE: return " uncached";
1445
	case I915_CACHE_LLC: return HAS_LLC(i915) ? " LLC" : " snooped";
1446
	case I915_CACHE_L3_LLC: return " L3+LLC";
1447
	case I915_CACHE_WT: return " WT";
1448 1449 1450 1451 1452 1453 1454 1455 1456 1457
	default: return "";
	}
}

/* NB: please notice the memset */
void i915_get_extra_instdone(struct drm_device *dev, uint32_t *instdone)
{
	struct drm_i915_private *dev_priv = dev->dev_private;
	memset(instdone, 0, sizeof(*instdone) * I915_NUM_INSTDONE_REG);

1458
	if (IS_GEN2(dev) || IS_GEN3(dev))
1459
		instdone[0] = I915_READ(GEN2_INSTDONE);
1460
	else if (IS_GEN4(dev) || IS_GEN5(dev) || IS_GEN6(dev)) {
1461
		instdone[0] = I915_READ(RING_INSTDONE(RENDER_RING_BASE));
1462
		instdone[1] = I915_READ(GEN4_INSTDONE1);
1463
	} else if (INTEL_INFO(dev)->gen >= 7) {
1464
		instdone[0] = I915_READ(RING_INSTDONE(RENDER_RING_BASE));
1465 1466 1467 1468 1469
		instdone[1] = I915_READ(GEN7_SC_INSTDONE);
		instdone[2] = I915_READ(GEN7_SAMPLER_INSTDONE);
		instdone[3] = I915_READ(GEN7_ROW_INSTDONE);
	}
}