intel_gt.c 20.8 KB
Newer Older
1 2 3 4 5
// SPDX-License-Identifier: MIT
/*
 * Copyright © 2019 Intel Corporation
 */

6
#include "debugfs_gt.h"
7 8

#include "gem/i915_gem_lmem.h"
9
#include "i915_drv.h"
10
#include "intel_context.h"
11
#include "intel_gt.h"
12
#include "intel_gt_buffer_pool.h"
13
#include "intel_gt_clock_utils.h"
14
#include "intel_gt_pm.h"
15
#include "intel_gt_requests.h"
16
#include "intel_migrate.h"
17
#include "intel_mocs.h"
18
#include "intel_rc6.h"
19
#include "intel_renderstate.h"
20
#include "intel_rps.h"
21
#include "intel_uncore.h"
22
#include "intel_pm.h"
23
#include "shmem_utils.h"
24

25
void intel_gt_init_early(struct intel_gt *gt, struct drm_i915_private *i915)
26
{
27 28 29
	gt->i915 = i915;
	gt->uncore = &i915->uncore;

30 31 32
	spin_lock_init(&gt->irq_lock);

	INIT_LIST_HEAD(&gt->closed_vma);
33
	spin_lock_init(&gt->closed_lock);
34

35 36 37
	init_llist_head(&gt->watchdog.list);
	INIT_WORK(&gt->watchdog.work, intel_gt_watchdog_work);

38
	intel_gt_init_buffer_pool(gt);
39
	intel_gt_init_reset(gt);
40
	intel_gt_init_requests(gt);
41
	intel_gt_init_timelines(gt);
42
	intel_gt_pm_init_early(gt);
43

44
	intel_uc_init_early(&gt->uc);
45
	intel_rps_init_early(&gt->rps);
46
}
47

48 49 50 51 52 53 54
int intel_gt_probe_lmem(struct intel_gt *gt)
{
	struct drm_i915_private *i915 = gt->i915;
	struct intel_memory_region *mem;
	int id;
	int err;

M
Matthew Auld 已提交
55 56 57
	mem = intel_gt_setup_lmem(gt);
	if (mem == ERR_PTR(-ENODEV))
		mem = intel_gt_setup_fake_lmem(gt);
58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
	if (IS_ERR(mem)) {
		err = PTR_ERR(mem);
		if (err == -ENODEV)
			return 0;

		drm_err(&i915->drm,
			"Failed to setup region(%d) type=%d\n",
			err, INTEL_MEMORY_LOCAL);
		return err;
	}

	id = INTEL_REGION_LMEM;

	mem->id = id;

73 74
	intel_memory_region_set_name(mem, "local%u", mem->instance);

75 76 77 78 79 80 81
	GEM_BUG_ON(!HAS_REGION(i915, id));
	GEM_BUG_ON(i915->mm.regions[id]);
	i915->mm.regions[id] = mem;

	return 0;
}

82
void intel_gt_init_hw_early(struct intel_gt *gt, struct i915_ggtt *ggtt)
83
{
84
	gt->ggtt = ggtt;
85 86
}

87 88 89 90 91
static const struct intel_mmio_range icl_l3bank_steering_table[] = {
	{ 0x00B100, 0x00B3FF },
	{},
};

92 93 94 95 96 97 98
static u16 slicemask(struct intel_gt *gt, int count)
{
	u64 dss_mask = intel_sseu_get_subslices(&gt->info.sseu, 0);

	return intel_slicemask_from_dssmask(dss_mask, count);
}

99 100
int intel_gt_init_mmio(struct intel_gt *gt)
{
101 102
	struct drm_i915_private *i915 = gt->i915;

103 104
	intel_gt_init_clock_frequency(gt);

105
	intel_uc_init_mmio(&gt->uc);
106
	intel_sseu_info_init(gt);
107

108 109 110 111 112 113 114 115 116 117 118 119
	/*
	 * An mslice is unavailable only if both the meml3 for the slice is
	 * disabled *and* all of the DSS in the slice (quadrant) are disabled.
	 */
	if (HAS_MSLICES(i915))
		gt->info.mslice_mask =
			slicemask(gt, GEN_DSS_PER_MSLICE) |
			(intel_uncore_read(gt->uncore, GEN10_MIRROR_FUSE3) &
			 GEN12_MEML3_EN_MASK);

	if (GRAPHICS_VER(i915) >= 11 &&
		   GRAPHICS_VER_FULL(i915) < IP_VER(12, 50)) {
120 121 122 123
		gt->steering_table[L3BANK] = icl_l3bank_steering_table;
		gt->info.l3bank_mask =
			~intel_uncore_read(gt->uncore, GEN10_MIRROR_FUSE3) &
			GEN10_L3BANK_MASK;
124 125
	} else if (HAS_MSLICES(i915)) {
		MISSING_CASE(INTEL_INFO(i915)->platform);
126 127
	}

128 129 130
	return intel_engines_init_mmio(gt);
}

131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
static void init_unused_ring(struct intel_gt *gt, u32 base)
{
	struct intel_uncore *uncore = gt->uncore;

	intel_uncore_write(uncore, RING_CTL(base), 0);
	intel_uncore_write(uncore, RING_HEAD(base), 0);
	intel_uncore_write(uncore, RING_TAIL(base), 0);
	intel_uncore_write(uncore, RING_START(base), 0);
}

static void init_unused_rings(struct intel_gt *gt)
{
	struct drm_i915_private *i915 = gt->i915;

	if (IS_I830(i915)) {
		init_unused_ring(gt, PRB1_BASE);
		init_unused_ring(gt, SRB0_BASE);
		init_unused_ring(gt, SRB1_BASE);
		init_unused_ring(gt, SRB2_BASE);
		init_unused_ring(gt, SRB3_BASE);
151
	} else if (GRAPHICS_VER(i915) == 2) {
152 153
		init_unused_ring(gt, SRB0_BASE);
		init_unused_ring(gt, SRB1_BASE);
154
	} else if (GRAPHICS_VER(i915) == 3) {
155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170
		init_unused_ring(gt, PRB1_BASE);
		init_unused_ring(gt, PRB2_BASE);
	}
}

int intel_gt_init_hw(struct intel_gt *gt)
{
	struct drm_i915_private *i915 = gt->i915;
	struct intel_uncore *uncore = gt->uncore;
	int ret;

	gt->last_init_time = ktime_get();

	/* Double layer security blanket, see i915_gem_init() */
	intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);

171
	if (HAS_EDRAM(i915) && GRAPHICS_VER(i915) < 9)
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
		intel_uncore_rmw(uncore, HSW_IDICR, 0, IDIHASHMSK(0xf));

	if (IS_HASWELL(i915))
		intel_uncore_write(uncore,
				   MI_PREDICATE_RESULT_2,
				   IS_HSW_GT3(i915) ?
				   LOWER_SLICE_ENABLED : LOWER_SLICE_DISABLED);

	/* Apply the GT workarounds... */
	intel_gt_apply_workarounds(gt);
	/* ...and determine whether they are sticking. */
	intel_gt_verify_workarounds(gt, "init");

	intel_gt_init_swizzling(gt);

	/*
	 * At least 830 can leave some of the unused rings
	 * "active" (ie. head != tail) after resume which
	 * will prevent c3 entry. Makes sure all unused rings
	 * are totally idle.
	 */
	init_unused_rings(gt);

	ret = i915_ppgtt_init_hw(gt);
	if (ret) {
		DRM_ERROR("Enabling PPGTT failed (%d)\n", ret);
		goto out;
	}

	/* We can't enable contexts until all firmware is loaded */
	ret = intel_uc_init_hw(&gt->uc);
	if (ret) {
		i915_probe_error(i915, "Enabling uc failed (%d)\n", ret);
		goto out;
	}

	intel_mocs_init(gt);

out:
	intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
	return ret;
}

215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
static void rmw_set(struct intel_uncore *uncore, i915_reg_t reg, u32 set)
{
	intel_uncore_rmw(uncore, reg, 0, set);
}

static void rmw_clear(struct intel_uncore *uncore, i915_reg_t reg, u32 clr)
{
	intel_uncore_rmw(uncore, reg, clr, 0);
}

static void clear_register(struct intel_uncore *uncore, i915_reg_t reg)
{
	intel_uncore_rmw(uncore, reg, 0, 0);
}

230
static void gen6_clear_engine_error_register(struct intel_engine_cs *engine)
231 232 233 234 235 236 237 238 239 240 241 242 243
{
	GEN6_RING_FAULT_REG_RMW(engine, RING_FAULT_VALID, 0);
	GEN6_RING_FAULT_REG_POSTING_READ(engine);
}

void
intel_gt_clear_error_registers(struct intel_gt *gt,
			       intel_engine_mask_t engine_mask)
{
	struct drm_i915_private *i915 = gt->i915;
	struct intel_uncore *uncore = gt->uncore;
	u32 eir;

244
	if (GRAPHICS_VER(i915) != 2)
245 246
		clear_register(uncore, PGTBL_ER);

247
	if (GRAPHICS_VER(i915) < 4)
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
		clear_register(uncore, IPEIR(RENDER_RING_BASE));
	else
		clear_register(uncore, IPEIR_I965);

	clear_register(uncore, EIR);
	eir = intel_uncore_read(uncore, EIR);
	if (eir) {
		/*
		 * some errors might have become stuck,
		 * mask them.
		 */
		DRM_DEBUG_DRIVER("EIR stuck: 0x%08x, masking\n", eir);
		rmw_set(uncore, EMR, eir);
		intel_uncore_write(uncore, GEN2_IIR,
				   I915_MASTER_ERROR_INTERRUPT);
	}

265
	if (GRAPHICS_VER(i915) >= 12) {
266 267
		rmw_clear(uncore, GEN12_RING_FAULT_REG, RING_FAULT_VALID);
		intel_uncore_posting_read(uncore, GEN12_RING_FAULT_REG);
268
	} else if (GRAPHICS_VER(i915) >= 8) {
269 270
		rmw_clear(uncore, GEN8_RING_FAULT_REG, RING_FAULT_VALID);
		intel_uncore_posting_read(uncore, GEN8_RING_FAULT_REG);
271
	} else if (GRAPHICS_VER(i915) >= 6) {
272 273 274
		struct intel_engine_cs *engine;
		enum intel_engine_id id;

275
		for_each_engine_masked(engine, gt, engine_mask, id)
276
			gen6_clear_engine_error_register(engine);
277 278 279 280 281 282 283 284 285
	}
}

static void gen6_check_faults(struct intel_gt *gt)
{
	struct intel_engine_cs *engine;
	enum intel_engine_id id;
	u32 fault;

286
	for_each_engine(engine, gt, id) {
287 288
		fault = GEN6_RING_FAULT_REG_READ(engine);
		if (fault & RING_FAULT_VALID) {
289 290 291 292 293 294 295 296 297 298
			drm_dbg(&engine->i915->drm, "Unexpected fault\n"
				"\tAddr: 0x%08lx\n"
				"\tAddress space: %s\n"
				"\tSource ID: %d\n"
				"\tType: %d\n",
				fault & PAGE_MASK,
				fault & RING_FAULT_GTTSEL_MASK ?
				"GGTT" : "PPGTT",
				RING_FAULT_SRCID(fault),
				RING_FAULT_FAULT_TYPE(fault));
299 300 301 302 303 304 305
		}
	}
}

static void gen8_check_faults(struct intel_gt *gt)
{
	struct intel_uncore *uncore = gt->uncore;
306 307 308
	i915_reg_t fault_reg, fault_data0_reg, fault_data1_reg;
	u32 fault;

309
	if (GRAPHICS_VER(gt->i915) >= 12) {
310 311 312 313 314 315 316 317
		fault_reg = GEN12_RING_FAULT_REG;
		fault_data0_reg = GEN12_FAULT_TLB_DATA0;
		fault_data1_reg = GEN12_FAULT_TLB_DATA1;
	} else {
		fault_reg = GEN8_RING_FAULT_REG;
		fault_data0_reg = GEN8_FAULT_TLB_DATA0;
		fault_data1_reg = GEN8_FAULT_TLB_DATA1;
	}
318

319
	fault = intel_uncore_read(uncore, fault_reg);
320 321 322 323
	if (fault & RING_FAULT_VALID) {
		u32 fault_data0, fault_data1;
		u64 fault_addr;

324 325 326
		fault_data0 = intel_uncore_read(uncore, fault_data0_reg);
		fault_data1 = intel_uncore_read(uncore, fault_data1_reg);

327 328 329
		fault_addr = ((u64)(fault_data1 & FAULT_VA_HIGH_BITS) << 44) |
			     ((u64)fault_data0 << 12);

330 331 332 333 334 335 336 337 338 339 340
		drm_dbg(&uncore->i915->drm, "Unexpected fault\n"
			"\tAddr: 0x%08x_%08x\n"
			"\tAddress space: %s\n"
			"\tEngine ID: %d\n"
			"\tSource ID: %d\n"
			"\tType: %d\n",
			upper_32_bits(fault_addr), lower_32_bits(fault_addr),
			fault_data1 & FAULT_GTT_SEL ? "GGTT" : "PPGTT",
			GEN8_RING_FAULT_ENGINE_ID(fault),
			RING_FAULT_SRCID(fault),
			RING_FAULT_FAULT_TYPE(fault));
341 342 343 344 345 346 347 348
	}
}

void intel_gt_check_and_clear_faults(struct intel_gt *gt)
{
	struct drm_i915_private *i915 = gt->i915;

	/* From GEN8 onwards we only have one 'All Engine Fault Register' */
349
	if (GRAPHICS_VER(i915) >= 8)
350
		gen8_check_faults(gt);
351
	else if (GRAPHICS_VER(i915) >= 6)
352 353 354 355 356 357
		gen6_check_faults(gt);
	else
		return;

	intel_gt_clear_error_registers(gt, ALL_ENGINES);
}
358 359 360

void intel_gt_flush_ggtt_writes(struct intel_gt *gt)
{
361
	struct intel_uncore *uncore = gt->uncore;
362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384
	intel_wakeref_t wakeref;

	/*
	 * No actual flushing is required for the GTT write domain for reads
	 * from the GTT domain. Writes to it "immediately" go to main memory
	 * as far as we know, so there's no chipset flush. It also doesn't
	 * land in the GPU render cache.
	 *
	 * However, we do have to enforce the order so that all writes through
	 * the GTT land before any writes to the device, such as updates to
	 * the GATT itself.
	 *
	 * We also have to wait a bit for the writes to land from the GTT.
	 * An uncached read (i.e. mmio) seems to be ideal for the round-trip
	 * timing. This issue has only been observed when switching quickly
	 * between GTT writes and CPU reads from inside the kernel on recent hw,
	 * and it appears to only affect discrete GTT blocks (i.e. on LLC
	 * system agents we cannot reproduce this behaviour, until Cannonlake
	 * that was!).
	 */

	wmb();

385
	if (INTEL_INFO(gt->i915)->has_coherent_ggtt)
386 387
		return;

388
	intel_gt_chipset_flush(gt);
389

390
	with_intel_runtime_pm_if_in_use(uncore->rpm, wakeref) {
391
		unsigned long flags;
392

393
		spin_lock_irqsave(&uncore->lock, flags);
394 395
		intel_uncore_posting_read_fw(uncore,
					     RING_HEAD(RENDER_RING_BASE));
396
		spin_unlock_irqrestore(&uncore->lock, flags);
397 398
	}
}
399 400 401 402

void intel_gt_chipset_flush(struct intel_gt *gt)
{
	wmb();
403
	if (GRAPHICS_VER(gt->i915) < 6)
404 405
		intel_gtt_chipset_flush();
}
406

407 408
void intel_gt_driver_register(struct intel_gt *gt)
{
409
	intel_rps_driver_register(&gt->rps);
410 411

	debugfs_gt_register(gt);
412 413 414
}

static int intel_gt_init_scratch(struct intel_gt *gt, unsigned int size)
415 416 417 418 419 420
{
	struct drm_i915_private *i915 = gt->i915;
	struct drm_i915_gem_object *obj;
	struct i915_vma *vma;
	int ret;

421 422 423
	obj = i915_gem_object_create_lmem(i915, size, I915_BO_ALLOC_VOLATILE);
	if (IS_ERR(obj))
		obj = i915_gem_object_create_stolen(i915, size);
424
	if (IS_ERR(obj))
425 426
		obj = i915_gem_object_create_internal(i915, size);
	if (IS_ERR(obj)) {
427
		drm_err(&i915->drm, "Failed to allocate scratch page\n");
428 429 430 431 432 433 434 435 436
		return PTR_ERR(obj);
	}

	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
	if (IS_ERR(vma)) {
		ret = PTR_ERR(vma);
		goto err_unref;
	}

437
	ret = i915_ggtt_pin(vma, NULL, 0, PIN_HIGH);
438 439 440
	if (ret)
		goto err_unref;

441 442
	gt->scratch = i915_vma_make_unshrinkable(vma);

443 444 445 446 447 448 449
	return 0;

err_unref:
	i915_gem_object_put(obj);
	return ret;
}

450
static void intel_gt_fini_scratch(struct intel_gt *gt)
451 452 453
{
	i915_vma_unpin_and_release(&gt->scratch, 0);
}
454

455 456 457
static struct i915_address_space *kernel_vm(struct intel_gt *gt)
{
	if (INTEL_PPGTT(gt->i915) > INTEL_PPGTT_ALIASING)
458
		return &i915_ppgtt_create(gt)->vm;
459 460 461 462
	else
		return i915_vm_get(&gt->ggtt->vm);
}

463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483
static int __engines_record_defaults(struct intel_gt *gt)
{
	struct i915_request *requests[I915_NUM_ENGINES] = {};
	struct intel_engine_cs *engine;
	enum intel_engine_id id;
	int err = 0;

	/*
	 * As we reset the gpu during very early sanitisation, the current
	 * register state on the GPU should reflect its defaults values.
	 * We load a context onto the hw (with restore-inhibit), then switch
	 * over to a second context to save that default register state. We
	 * can then prime every new context with that state so they all start
	 * from the same default HW values.
	 */

	for_each_engine(engine, gt, id) {
		struct intel_renderstate so;
		struct intel_context *ce;
		struct i915_request *rq;

484 485 486
		/* We must be able to switch to something! */
		GEM_BUG_ON(!engine->kernel_context);

487 488 489 490 491 492
		ce = intel_context_create(engine);
		if (IS_ERR(ce)) {
			err = PTR_ERR(ce);
			goto out;
		}

493 494 495 496 497
		err = intel_renderstate_init(&so, ce);
		if (err)
			goto err;

		rq = i915_request_create(ce);
498 499
		if (IS_ERR(rq)) {
			err = PTR_ERR(rq);
500
			goto err_fini;
501 502 503 504 505 506 507 508 509 510 511 512 513
		}

		err = intel_engine_emit_ctx_wa(rq);
		if (err)
			goto err_rq;

		err = intel_renderstate_emit(&so, rq);
		if (err)
			goto err_rq;

err_rq:
		requests[id] = i915_request_get(rq);
		i915_request_add(rq);
514 515 516 517 518
err_fini:
		intel_renderstate_fini(&so, ce);
err:
		if (err) {
			intel_context_put(ce);
519
			goto out;
520
		}
521 522 523 524 525 526 527 528 529 530
	}

	/* Flush the default context image to memory, and enable powersaving. */
	if (intel_gt_wait_for_idle(gt, I915_GEM_IDLE_TIMEOUT) == -ETIME) {
		err = -EIO;
		goto out;
	}

	for (id = 0; id < ARRAY_SIZE(requests); id++) {
		struct i915_request *rq;
531
		struct file *state;
532 533 534 535 536

		rq = requests[id];
		if (!rq)
			continue;

537 538 539 540 541
		if (rq->fence.error) {
			err = -EIO;
			goto out;
		}

542
		GEM_BUG_ON(!test_bit(CONTEXT_ALLOC_BIT, &rq->context->flags));
543
		if (!rq->context->state)
544 545
			continue;

546 547 548 549
		/* Keep a copy of the state's backing pages; free the obj */
		state = shmem_create_from_object(rq->context->state->obj);
		if (IS_ERR(state)) {
			err = PTR_ERR(state);
550 551
			goto out;
		}
552
		rq->engine->default_state = state;
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
	}

out:
	/*
	 * If we have to abandon now, we expect the engines to be idle
	 * and ready to be torn-down. The quickest way we can accomplish
	 * this is by declaring ourselves wedged.
	 */
	if (err)
		intel_gt_set_wedged(gt);

	for (id = 0; id < ARRAY_SIZE(requests); id++) {
		struct intel_context *ce;
		struct i915_request *rq;

		rq = requests[id];
		if (!rq)
			continue;

		ce = rq->context;
		i915_request_put(rq);
		intel_context_put(ce);
	}
	return err;
}

static int __engines_verify_workarounds(struct intel_gt *gt)
{
	struct intel_engine_cs *engine;
	enum intel_engine_id id;
	int err = 0;

	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
		return 0;

	for_each_engine(engine, gt, id) {
		if (intel_engine_verify_workarounds(engine, "load"))
			err = -EIO;
	}

593 594 595 596
	/* Flush and restore the kernel context for safety */
	if (intel_gt_wait_for_idle(gt, I915_GEM_IDLE_TIMEOUT) == -ETIME)
		err = -EIO;

597 598 599 600 601
	return err;
}

static void __intel_gt_disable(struct intel_gt *gt)
{
602
	intel_gt_set_wedged_on_fini(gt);
603 604 605 606 607 608 609

	intel_gt_suspend_prepare(gt);
	intel_gt_suspend_late(gt);

	GEM_BUG_ON(intel_gt_pm_is_awake(gt));
}

610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628
int intel_gt_wait_for_idle(struct intel_gt *gt, long timeout)
{
	long remaining_timeout;

	/* If the device is asleep, we have no requests outstanding */
	if (!intel_gt_pm_is_awake(gt))
		return 0;

	while ((timeout = intel_gt_retire_requests_timeout(gt, timeout,
							   &remaining_timeout)) > 0) {
		cond_resched();
		if (signal_pending(current))
			return -EINTR;
	}

	return timeout ? timeout : intel_uc_wait_for_idle(&gt->uc,
							  remaining_timeout);
}

629 630 631 632
int intel_gt_init(struct intel_gt *gt)
{
	int err;

633
	err = i915_inject_probe_error(gt->i915, -ENODEV);
634 635 636
	if (err)
		return err;

637 638 639 640 641 642 643 644 645
	/*
	 * This is just a security blanket to placate dragons.
	 * On some systems, we very sporadically observe that the first TLBs
	 * used by the CS may be stale, despite us poking the TLB reset. If
	 * we hold the forcewake during initialisation these problems
	 * just magically go away.
	 */
	intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);

646 647
	err = intel_gt_init_scratch(gt,
				    GRAPHICS_VER(gt->i915) == 2 ? SZ_256K : SZ_4K);
648 649 650
	if (err)
		goto out_fw;

651 652
	intel_gt_pm_init(gt);

653 654 655
	gt->vm = kernel_vm(gt);
	if (!gt->vm) {
		err = -ENOMEM;
656
		goto err_pm;
657 658
	}

659 660 661 662
	err = intel_engines_init(gt);
	if (err)
		goto err_engines;

663 664 665
	err = intel_uc_init(&gt->uc);
	if (err)
		goto err_engines;
666 667 668

	err = intel_gt_resume(gt);
	if (err)
669
		goto err_uc_init;
670 671 672 673 674 675 676 677 678

	err = __engines_record_defaults(gt);
	if (err)
		goto err_gt;

	err = __engines_verify_workarounds(gt);
	if (err)
		goto err_gt;

679 680
	intel_uc_init_late(&gt->uc);

681 682 683 684
	err = i915_inject_probe_error(gt->i915, -EIO);
	if (err)
		goto err_gt;

685 686
	intel_migrate_init(&gt->migrate, gt);

687 688 689 690 691 692 693 694 695 696 697
	goto out_fw;
err_gt:
	__intel_gt_disable(gt);
	intel_uc_fini_hw(&gt->uc);
err_uc_init:
	intel_uc_fini(&gt->uc);
err_engines:
	intel_engines_release(gt);
	i915_vm_put(fetch_and_zero(&gt->vm));
err_pm:
	intel_gt_pm_fini(gt);
698
	intel_gt_fini_scratch(gt);
699 700 701 702
out_fw:
	if (err)
		intel_gt_set_wedged_on_init(gt);
	intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
703
	return err;
704 705 706 707
}

void intel_gt_driver_remove(struct intel_gt *gt)
{
708 709
	__intel_gt_disable(gt);

710
	intel_migrate_fini(&gt->migrate);
711
	intel_uc_driver_remove(&gt->uc);
712 713

	intel_engines_release(gt);
714 715 716 717
}

void intel_gt_driver_unregister(struct intel_gt *gt)
{
718 719
	intel_wakeref_t wakeref;

720
	intel_rps_driver_unregister(&gt->rps);
721 722 723 724 725 726 727

	/*
	 * Upon unregistering the device to prevent any new users, cancel
	 * all in-flight requests so that we can quickly unbind the active
	 * resources.
	 */
	intel_gt_set_wedged(gt);
728 729 730 731

	/* Scrub all HW state upon release */
	with_intel_runtime_pm(gt->uncore->rpm, wakeref)
		__intel_gt_reset(gt, ALL_ENGINES);
732 733 734 735
}

void intel_gt_driver_release(struct intel_gt *gt)
{
736 737 738 739 740 741
	struct i915_address_space *vm;

	vm = fetch_and_zero(&gt->vm);
	if (vm) /* FIXME being called twice on error paths :( */
		i915_vm_put(vm);

742
	intel_gt_pm_fini(gt);
743
	intel_gt_fini_scratch(gt);
744
	intel_gt_fini_buffer_pool(gt);
745 746
}

747
void intel_gt_driver_late_release(struct intel_gt *gt)
748
{
749 750 751
	/* We need to wait for inflight RCU frees to release their grip */
	rcu_barrier();

752
	intel_uc_driver_late_release(&gt->uc);
753
	intel_gt_fini_requests(gt);
754
	intel_gt_fini_reset(gt);
755
	intel_gt_fini_timelines(gt);
756
	intel_engines_free(gt);
757
}
758

759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805
/**
 * intel_gt_reg_needs_read_steering - determine whether a register read
 *     requires explicit steering
 * @gt: GT structure
 * @reg: the register to check steering requirements for
 * @type: type of multicast steering to check
 *
 * Determines whether @reg needs explicit steering of a specific type for
 * reads.
 *
 * Returns false if @reg does not belong to a register range of the given
 * steering type, or if the default (subslice-based) steering IDs are suitable
 * for @type steering too.
 */
static bool intel_gt_reg_needs_read_steering(struct intel_gt *gt,
					     i915_reg_t reg,
					     enum intel_steering_type type)
{
	const u32 offset = i915_mmio_reg_offset(reg);
	const struct intel_mmio_range *entry;

	if (likely(!intel_gt_needs_read_steering(gt, type)))
		return false;

	for (entry = gt->steering_table[type]; entry->end; entry++) {
		if (offset >= entry->start && offset <= entry->end)
			return true;
	}

	return false;
}

/**
 * intel_gt_get_valid_steering - determines valid IDs for a class of MCR steering
 * @gt: GT structure
 * @type: multicast register type
 * @sliceid: Slice ID returned
 * @subsliceid: Subslice ID returned
 *
 * Determines sliceid and subsliceid values that will steer reads
 * of a specific multicast register class to a valid value.
 */
static void intel_gt_get_valid_steering(struct intel_gt *gt,
					enum intel_steering_type type,
					u8 *sliceid, u8 *subsliceid)
{
	switch (type) {
806 807 808 809 810 811
	case L3BANK:
		GEM_DEBUG_WARN_ON(!gt->info.l3bank_mask); /* should be impossible! */

		*sliceid = 0;		/* unused */
		*subsliceid = __ffs(gt->info.l3bank_mask);
		break;
812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827
	case MSLICE:
		GEM_DEBUG_WARN_ON(!gt->info.mslice_mask); /* should be impossible! */

		*sliceid = __ffs(gt->info.mslice_mask);
		*subsliceid = 0;	/* unused */
		break;
	case LNCF:
		GEM_DEBUG_WARN_ON(!gt->info.mslice_mask); /* should be impossible! */

		/*
		 * An LNCF is always present if its mslice is present, so we
		 * can safely just steer to LNCF 0 in all cases.
		 */
		*sliceid = __ffs(gt->info.mslice_mask) << 1;
		*subsliceid = 0;	/* unused */
		break;
828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864
	default:
		MISSING_CASE(type);
		*sliceid = 0;
		*subsliceid = 0;
	}
}

/**
 * intel_gt_read_register_fw - reads a GT register with support for multicast
 * @gt: GT structure
 * @reg: register to read
 *
 * This function will read a GT register.  If the register is a multicast
 * register, the read will be steered to a valid instance (i.e., one that
 * isn't fused off or powered down by power gating).
 *
 * Returns the value from a valid instance of @reg.
 */
u32 intel_gt_read_register_fw(struct intel_gt *gt, i915_reg_t reg)
{
	int type;
	u8 sliceid, subsliceid;

	for (type = 0; type < NUM_STEERING_TYPES; type++) {
		if (intel_gt_reg_needs_read_steering(gt, reg, type)) {
			intel_gt_get_valid_steering(gt, type, &sliceid,
						    &subsliceid);
			return intel_uncore_read_with_mcr_steering_fw(gt->uncore,
								      reg,
								      sliceid,
								      subsliceid);
		}
	}

	return intel_uncore_read_fw(gt->uncore, reg);
}

865 866 867 868
void intel_gt_info_print(const struct intel_gt_info *info,
			 struct drm_printer *p)
{
	drm_printf(p, "available engines: %x\n", info->engine_mask);
869 870

	intel_sseu_dump(&info->sseu, p);
871
}