amdgpu_device.c 55.9 KB
Newer Older
A
Alex Deucher 已提交
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
/*
 * Copyright 2008 Advanced Micro Devices, Inc.
 * Copyright 2008 Red Hat Inc.
 * Copyright 2009 Jerome Glisse.
 *
 * 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: Dave Airlie
 *          Alex Deucher
 *          Jerome Glisse
 */
#include <linux/console.h>
#include <linux/slab.h>
#include <linux/debugfs.h>
#include <drm/drmP.h>
#include <drm/drm_crtc_helper.h>
#include <drm/amdgpu_drm.h>
#include <linux/vgaarb.h>
#include <linux/vga_switcheroo.h>
#include <linux/efi.h>
#include "amdgpu.h"
#include "amdgpu_i2c.h"
#include "atom.h"
#include "amdgpu_atombios.h"
41
#include "amd_pcie.h"
42 43 44
#ifdef CONFIG_DRM_AMDGPU_CIK
#include "cik.h"
#endif
45
#include "vi.h"
A
Alex Deucher 已提交
46 47 48 49 50 51 52 53 54 55 56 57 58
#include "bif/bif_4_1_d.h"

static int amdgpu_debugfs_regs_init(struct amdgpu_device *adev);
static void amdgpu_debugfs_regs_cleanup(struct amdgpu_device *adev);

static const char *amdgpu_asic_name[] = {
	"BONAIRE",
	"KAVERI",
	"KABINI",
	"HAWAII",
	"MULLINS",
	"TOPAZ",
	"TONGA",
59
	"FIJI",
A
Alex Deucher 已提交
60
	"CARRIZO",
S
Samuel Li 已提交
61
	"STONEY",
A
Alex Deucher 已提交
62 63 64 65 66 67 68
	"LAST",
};

bool amdgpu_device_is_px(struct drm_device *dev)
{
	struct amdgpu_device *adev = dev->dev_private;

69
	if (adev->flags & AMD_IS_PX)
A
Alex Deucher 已提交
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
		return true;
	return false;
}

/*
 * MMIO register access helper functions.
 */
uint32_t amdgpu_mm_rreg(struct amdgpu_device *adev, uint32_t reg,
			bool always_indirect)
{
	if ((reg * 4) < adev->rmmio_size && !always_indirect)
		return readl(((void __iomem *)adev->rmmio) + (reg * 4));
	else {
		unsigned long flags;
		uint32_t ret;

		spin_lock_irqsave(&adev->mmio_idx_lock, flags);
		writel((reg * 4), ((void __iomem *)adev->rmmio) + (mmMM_INDEX * 4));
		ret = readl(((void __iomem *)adev->rmmio) + (mmMM_DATA * 4));
		spin_unlock_irqrestore(&adev->mmio_idx_lock, flags);

		return ret;
	}
}

void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
		    bool always_indirect)
{
	if ((reg * 4) < adev->rmmio_size && !always_indirect)
		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
	else {
		unsigned long flags;

		spin_lock_irqsave(&adev->mmio_idx_lock, flags);
		writel((reg * 4), ((void __iomem *)adev->rmmio) + (mmMM_INDEX * 4));
		writel(v, ((void __iomem *)adev->rmmio) + (mmMM_DATA * 4));
		spin_unlock_irqrestore(&adev->mmio_idx_lock, flags);
	}
}

u32 amdgpu_io_rreg(struct amdgpu_device *adev, u32 reg)
{
	if ((reg * 4) < adev->rio_mem_size)
		return ioread32(adev->rio_mem + (reg * 4));
	else {
		iowrite32((reg * 4), adev->rio_mem + (mmMM_INDEX * 4));
		return ioread32(adev->rio_mem + (mmMM_DATA * 4));
	}
}

void amdgpu_io_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
{

	if ((reg * 4) < adev->rio_mem_size)
		iowrite32(v, adev->rio_mem + (reg * 4));
	else {
		iowrite32((reg * 4), adev->rio_mem + (mmMM_INDEX * 4));
		iowrite32(v, adev->rio_mem + (mmMM_DATA * 4));
	}
}

/**
 * amdgpu_mm_rdoorbell - read a doorbell dword
 *
 * @adev: amdgpu_device pointer
 * @index: doorbell index
 *
 * Returns the value in the doorbell aperture at the
 * requested doorbell index (CIK).
 */
u32 amdgpu_mm_rdoorbell(struct amdgpu_device *adev, u32 index)
{
	if (index < adev->doorbell.num_doorbells) {
		return readl(adev->doorbell.ptr + index);
	} else {
		DRM_ERROR("reading beyond doorbell aperture: 0x%08x!\n", index);
		return 0;
	}
}

/**
 * amdgpu_mm_wdoorbell - write a doorbell dword
 *
 * @adev: amdgpu_device pointer
 * @index: doorbell index
 * @v: value to write
 *
 * Writes @v to the doorbell aperture at the
 * requested doorbell index (CIK).
 */
void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v)
{
	if (index < adev->doorbell.num_doorbells) {
		writel(v, adev->doorbell.ptr + index);
	} else {
		DRM_ERROR("writing beyond doorbell aperture: 0x%08x!\n", index);
	}
}

/**
 * amdgpu_invalid_rreg - dummy reg read function
 *
 * @adev: amdgpu device pointer
 * @reg: offset of register
 *
 * Dummy register read function.  Used for register blocks
 * that certain asics don't have (all asics).
 * Returns the value in the register.
 */
static uint32_t amdgpu_invalid_rreg(struct amdgpu_device *adev, uint32_t reg)
{
	DRM_ERROR("Invalid callback to read register 0x%04X\n", reg);
	BUG();
	return 0;
}

/**
 * amdgpu_invalid_wreg - dummy reg write function
 *
 * @adev: amdgpu device pointer
 * @reg: offset of register
 * @v: value to write to the register
 *
 * Dummy register read function.  Used for register blocks
 * that certain asics don't have (all asics).
 */
static void amdgpu_invalid_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
{
	DRM_ERROR("Invalid callback to write register 0x%04X with 0x%08X\n",
		  reg, v);
	BUG();
}

/**
 * amdgpu_block_invalid_rreg - dummy reg read function
 *
 * @adev: amdgpu device pointer
 * @block: offset of instance
 * @reg: offset of register
 *
 * Dummy register read function.  Used for register blocks
 * that certain asics don't have (all asics).
 * Returns the value in the register.
 */
static uint32_t amdgpu_block_invalid_rreg(struct amdgpu_device *adev,
					  uint32_t block, uint32_t reg)
{
	DRM_ERROR("Invalid callback to read register 0x%04X in block 0x%04X\n",
		  reg, block);
	BUG();
	return 0;
}

/**
 * amdgpu_block_invalid_wreg - dummy reg write function
 *
 * @adev: amdgpu device pointer
 * @block: offset of instance
 * @reg: offset of register
 * @v: value to write to the register
 *
 * Dummy register read function.  Used for register blocks
 * that certain asics don't have (all asics).
 */
static void amdgpu_block_invalid_wreg(struct amdgpu_device *adev,
				      uint32_t block,
				      uint32_t reg, uint32_t v)
{
	DRM_ERROR("Invalid block callback to write register 0x%04X in block 0x%04X with 0x%08X\n",
		  reg, block, v);
	BUG();
}

static int amdgpu_vram_scratch_init(struct amdgpu_device *adev)
{
	int r;

	if (adev->vram_scratch.robj == NULL) {
		r = amdgpu_bo_create(adev, AMDGPU_GPU_PAGE_SIZE,
249 250
				     PAGE_SIZE, true, AMDGPU_GEM_DOMAIN_VRAM,
				     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
251
				     NULL, NULL, &adev->vram_scratch.robj);
A
Alex Deucher 已提交
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
		if (r) {
			return r;
		}
	}

	r = amdgpu_bo_reserve(adev->vram_scratch.robj, false);
	if (unlikely(r != 0))
		return r;
	r = amdgpu_bo_pin(adev->vram_scratch.robj,
			  AMDGPU_GEM_DOMAIN_VRAM, &adev->vram_scratch.gpu_addr);
	if (r) {
		amdgpu_bo_unreserve(adev->vram_scratch.robj);
		return r;
	}
	r = amdgpu_bo_kmap(adev->vram_scratch.robj,
				(void **)&adev->vram_scratch.ptr);
	if (r)
		amdgpu_bo_unpin(adev->vram_scratch.robj);
	amdgpu_bo_unreserve(adev->vram_scratch.robj);

	return r;
}

static void amdgpu_vram_scratch_fini(struct amdgpu_device *adev)
{
	int r;

	if (adev->vram_scratch.robj == NULL) {
		return;
	}
	r = amdgpu_bo_reserve(adev->vram_scratch.robj, false);
	if (likely(r == 0)) {
		amdgpu_bo_kunmap(adev->vram_scratch.robj);
		amdgpu_bo_unpin(adev->vram_scratch.robj);
		amdgpu_bo_unreserve(adev->vram_scratch.robj);
	}
	amdgpu_bo_unref(&adev->vram_scratch.robj);
}

/**
 * amdgpu_program_register_sequence - program an array of registers.
 *
 * @adev: amdgpu_device pointer
 * @registers: pointer to the register array
 * @array_size: size of the register array
 *
 * Programs an array or registers with and and or masks.
 * This is a helper for setting golden registers.
 */
void amdgpu_program_register_sequence(struct amdgpu_device *adev,
				      const u32 *registers,
				      const u32 array_size)
{
	u32 tmp, reg, and_mask, or_mask;
	int i;

	if (array_size % 3)
		return;

	for (i = 0; i < array_size; i +=3) {
		reg = registers[i + 0];
		and_mask = registers[i + 1];
		or_mask = registers[i + 2];

		if (and_mask == 0xffffffff) {
			tmp = or_mask;
		} else {
			tmp = RREG32(reg);
			tmp &= ~and_mask;
			tmp |= or_mask;
		}
		WREG32(reg, tmp);
	}
}

void amdgpu_pci_config_reset(struct amdgpu_device *adev)
{
	pci_write_config_dword(adev->pdev, 0x7c, AMDGPU_ASIC_RESET_DATA);
}

/*
 * GPU doorbell aperture helpers function.
 */
/**
 * amdgpu_doorbell_init - Init doorbell driver information.
 *
 * @adev: amdgpu_device pointer
 *
 * Init doorbell driver information (CIK)
 * Returns 0 on success, error on failure.
 */
static int amdgpu_doorbell_init(struct amdgpu_device *adev)
{
	/* doorbell bar mapping */
	adev->doorbell.base = pci_resource_start(adev->pdev, 2);
	adev->doorbell.size = pci_resource_len(adev->pdev, 2);

	adev->doorbell.num_doorbells = min_t(u32, adev->doorbell.size / sizeof(u32), 
					     AMDGPU_DOORBELL_MAX_ASSIGNMENT+1);
	if (adev->doorbell.num_doorbells == 0)
		return -EINVAL;

	adev->doorbell.ptr = ioremap(adev->doorbell.base, adev->doorbell.num_doorbells * sizeof(u32));
	if (adev->doorbell.ptr == NULL) {
		return -ENOMEM;
	}
	DRM_INFO("doorbell mmio base: 0x%08X\n", (uint32_t)adev->doorbell.base);
	DRM_INFO("doorbell mmio size: %u\n", (unsigned)adev->doorbell.size);

	return 0;
}

/**
 * amdgpu_doorbell_fini - Tear down doorbell driver information.
 *
 * @adev: amdgpu_device pointer
 *
 * Tear down doorbell driver information (CIK)
 */
static void amdgpu_doorbell_fini(struct amdgpu_device *adev)
{
	iounmap(adev->doorbell.ptr);
	adev->doorbell.ptr = NULL;
}

/**
 * amdgpu_doorbell_get_kfd_info - Report doorbell configuration required to
 *                                setup amdkfd
 *
 * @adev: amdgpu_device pointer
 * @aperture_base: output returning doorbell aperture base physical address
 * @aperture_size: output returning doorbell aperture size in bytes
 * @start_offset: output returning # of doorbell bytes reserved for amdgpu.
 *
 * amdgpu and amdkfd share the doorbell aperture. amdgpu sets it up,
 * takes doorbells required for its own rings and reports the setup to amdkfd.
 * amdgpu reserved doorbells are at the start of the doorbell aperture.
 */
void amdgpu_doorbell_get_kfd_info(struct amdgpu_device *adev,
				phys_addr_t *aperture_base,
				size_t *aperture_size,
				size_t *start_offset)
{
	/*
	 * The first num_doorbells are used by amdgpu.
	 * amdkfd takes whatever's left in the aperture.
	 */
	if (adev->doorbell.size > adev->doorbell.num_doorbells * sizeof(u32)) {
		*aperture_base = adev->doorbell.base;
		*aperture_size = adev->doorbell.size;
		*start_offset = adev->doorbell.num_doorbells * sizeof(u32);
	} else {
		*aperture_base = 0;
		*aperture_size = 0;
		*start_offset = 0;
	}
}

/*
 * amdgpu_wb_*()
 * Writeback is the the method by which the the GPU updates special pages
 * in memory with the status of certain GPU events (fences, ring pointers,
 * etc.).
 */

/**
 * amdgpu_wb_fini - Disable Writeback and free memory
 *
 * @adev: amdgpu_device pointer
 *
 * Disables Writeback and frees the Writeback memory (all asics).
 * Used at driver shutdown.
 */
static void amdgpu_wb_fini(struct amdgpu_device *adev)
{
	if (adev->wb.wb_obj) {
		if (!amdgpu_bo_reserve(adev->wb.wb_obj, false)) {
			amdgpu_bo_kunmap(adev->wb.wb_obj);
			amdgpu_bo_unpin(adev->wb.wb_obj);
			amdgpu_bo_unreserve(adev->wb.wb_obj);
		}
		amdgpu_bo_unref(&adev->wb.wb_obj);
		adev->wb.wb = NULL;
		adev->wb.wb_obj = NULL;
	}
}

/**
 * amdgpu_wb_init- Init Writeback driver info and allocate memory
 *
 * @adev: amdgpu_device pointer
 *
 * Disables Writeback and frees the Writeback memory (all asics).
 * Used at driver startup.
 * Returns 0 on success or an -error on failure.
 */
static int amdgpu_wb_init(struct amdgpu_device *adev)
{
	int r;

	if (adev->wb.wb_obj == NULL) {
		r = amdgpu_bo_create(adev, AMDGPU_MAX_WB * 4, PAGE_SIZE, true,
454 455
				     AMDGPU_GEM_DOMAIN_GTT, 0,  NULL, NULL,
				     &adev->wb.wb_obj);
A
Alex Deucher 已提交
456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952
		if (r) {
			dev_warn(adev->dev, "(%d) create WB bo failed\n", r);
			return r;
		}
		r = amdgpu_bo_reserve(adev->wb.wb_obj, false);
		if (unlikely(r != 0)) {
			amdgpu_wb_fini(adev);
			return r;
		}
		r = amdgpu_bo_pin(adev->wb.wb_obj, AMDGPU_GEM_DOMAIN_GTT,
				&adev->wb.gpu_addr);
		if (r) {
			amdgpu_bo_unreserve(adev->wb.wb_obj);
			dev_warn(adev->dev, "(%d) pin WB bo failed\n", r);
			amdgpu_wb_fini(adev);
			return r;
		}
		r = amdgpu_bo_kmap(adev->wb.wb_obj, (void **)&adev->wb.wb);
		amdgpu_bo_unreserve(adev->wb.wb_obj);
		if (r) {
			dev_warn(adev->dev, "(%d) map WB bo failed\n", r);
			amdgpu_wb_fini(adev);
			return r;
		}

		adev->wb.num_wb = AMDGPU_MAX_WB;
		memset(&adev->wb.used, 0, sizeof(adev->wb.used));

		/* clear wb memory */
		memset((char *)adev->wb.wb, 0, AMDGPU_GPU_PAGE_SIZE);
	}

	return 0;
}

/**
 * amdgpu_wb_get - Allocate a wb entry
 *
 * @adev: amdgpu_device pointer
 * @wb: wb index
 *
 * Allocate a wb slot for use by the driver (all asics).
 * Returns 0 on success or -EINVAL on failure.
 */
int amdgpu_wb_get(struct amdgpu_device *adev, u32 *wb)
{
	unsigned long offset = find_first_zero_bit(adev->wb.used, adev->wb.num_wb);
	if (offset < adev->wb.num_wb) {
		__set_bit(offset, adev->wb.used);
		*wb = offset;
		return 0;
	} else {
		return -EINVAL;
	}
}

/**
 * amdgpu_wb_free - Free a wb entry
 *
 * @adev: amdgpu_device pointer
 * @wb: wb index
 *
 * Free a wb slot allocated for use by the driver (all asics)
 */
void amdgpu_wb_free(struct amdgpu_device *adev, u32 wb)
{
	if (wb < adev->wb.num_wb)
		__clear_bit(wb, adev->wb.used);
}

/**
 * amdgpu_vram_location - try to find VRAM location
 * @adev: amdgpu device structure holding all necessary informations
 * @mc: memory controller structure holding memory informations
 * @base: base address at which to put VRAM
 *
 * Function will place try to place VRAM at base address provided
 * as parameter (which is so far either PCI aperture address or
 * for IGP TOM base address).
 *
 * If there is not enough space to fit the unvisible VRAM in the 32bits
 * address space then we limit the VRAM size to the aperture.
 *
 * Note: We don't explicitly enforce VRAM start to be aligned on VRAM size,
 * this shouldn't be a problem as we are using the PCI aperture as a reference.
 * Otherwise this would be needed for rv280, all r3xx, and all r4xx, but
 * not IGP.
 *
 * Note: we use mc_vram_size as on some board we need to program the mc to
 * cover the whole aperture even if VRAM size is inferior to aperture size
 * Novell bug 204882 + along with lots of ubuntu ones
 *
 * Note: when limiting vram it's safe to overwritte real_vram_size because
 * we are not in case where real_vram_size is inferior to mc_vram_size (ie
 * note afected by bogus hw of Novell bug 204882 + along with lots of ubuntu
 * ones)
 *
 * Note: IGP TOM addr should be the same as the aperture addr, we don't
 * explicitly check for that thought.
 *
 * FIXME: when reducing VRAM size align new size on power of 2.
 */
void amdgpu_vram_location(struct amdgpu_device *adev, struct amdgpu_mc *mc, u64 base)
{
	uint64_t limit = (uint64_t)amdgpu_vram_limit << 20;

	mc->vram_start = base;
	if (mc->mc_vram_size > (adev->mc.mc_mask - base + 1)) {
		dev_warn(adev->dev, "limiting VRAM to PCI aperture size\n");
		mc->real_vram_size = mc->aper_size;
		mc->mc_vram_size = mc->aper_size;
	}
	mc->vram_end = mc->vram_start + mc->mc_vram_size - 1;
	if (limit && limit < mc->real_vram_size)
		mc->real_vram_size = limit;
	dev_info(adev->dev, "VRAM: %lluM 0x%016llX - 0x%016llX (%lluM used)\n",
			mc->mc_vram_size >> 20, mc->vram_start,
			mc->vram_end, mc->real_vram_size >> 20);
}

/**
 * amdgpu_gtt_location - try to find GTT location
 * @adev: amdgpu device structure holding all necessary informations
 * @mc: memory controller structure holding memory informations
 *
 * Function will place try to place GTT before or after VRAM.
 *
 * If GTT size is bigger than space left then we ajust GTT size.
 * Thus function will never fails.
 *
 * FIXME: when reducing GTT size align new size on power of 2.
 */
void amdgpu_gtt_location(struct amdgpu_device *adev, struct amdgpu_mc *mc)
{
	u64 size_af, size_bf;

	size_af = ((adev->mc.mc_mask - mc->vram_end) + mc->gtt_base_align) & ~mc->gtt_base_align;
	size_bf = mc->vram_start & ~mc->gtt_base_align;
	if (size_bf > size_af) {
		if (mc->gtt_size > size_bf) {
			dev_warn(adev->dev, "limiting GTT\n");
			mc->gtt_size = size_bf;
		}
		mc->gtt_start = (mc->vram_start & ~mc->gtt_base_align) - mc->gtt_size;
	} else {
		if (mc->gtt_size > size_af) {
			dev_warn(adev->dev, "limiting GTT\n");
			mc->gtt_size = size_af;
		}
		mc->gtt_start = (mc->vram_end + 1 + mc->gtt_base_align) & ~mc->gtt_base_align;
	}
	mc->gtt_end = mc->gtt_start + mc->gtt_size - 1;
	dev_info(adev->dev, "GTT: %lluM 0x%016llX - 0x%016llX\n",
			mc->gtt_size >> 20, mc->gtt_start, mc->gtt_end);
}

/*
 * GPU helpers function.
 */
/**
 * amdgpu_card_posted - check if the hw has already been initialized
 *
 * @adev: amdgpu_device pointer
 *
 * Check if the asic has been initialized (all asics).
 * Used at driver startup.
 * Returns true if initialized or false if not.
 */
bool amdgpu_card_posted(struct amdgpu_device *adev)
{
	uint32_t reg;

	/* then check MEM_SIZE, in case the crtcs are off */
	reg = RREG32(mmCONFIG_MEMSIZE);

	if (reg)
		return true;

	return false;

}

/**
 * amdgpu_boot_test_post_card - check and possibly initialize the hw
 *
 * @adev: amdgpu_device pointer
 *
 * Check if the asic is initialized and if not, attempt to initialize
 * it (all asics).
 * Returns true if initialized or false if not.
 */
bool amdgpu_boot_test_post_card(struct amdgpu_device *adev)
{
	if (amdgpu_card_posted(adev))
		return true;

	if (adev->bios) {
		DRM_INFO("GPU not posted. posting now...\n");
		if (adev->is_atom_bios)
			amdgpu_atom_asic_init(adev->mode_info.atom_context);
		return true;
	} else {
		dev_err(adev->dev, "Card not posted and no BIOS - ignoring\n");
		return false;
	}
}

/**
 * amdgpu_dummy_page_init - init dummy page used by the driver
 *
 * @adev: amdgpu_device pointer
 *
 * Allocate the dummy page used by the driver (all asics).
 * This dummy page is used by the driver as a filler for gart entries
 * when pages are taken out of the GART
 * Returns 0 on sucess, -ENOMEM on failure.
 */
int amdgpu_dummy_page_init(struct amdgpu_device *adev)
{
	if (adev->dummy_page.page)
		return 0;
	adev->dummy_page.page = alloc_page(GFP_DMA32 | GFP_KERNEL | __GFP_ZERO);
	if (adev->dummy_page.page == NULL)
		return -ENOMEM;
	adev->dummy_page.addr = pci_map_page(adev->pdev, adev->dummy_page.page,
					0, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
	if (pci_dma_mapping_error(adev->pdev, adev->dummy_page.addr)) {
		dev_err(&adev->pdev->dev, "Failed to DMA MAP the dummy page\n");
		__free_page(adev->dummy_page.page);
		adev->dummy_page.page = NULL;
		return -ENOMEM;
	}
	return 0;
}

/**
 * amdgpu_dummy_page_fini - free dummy page used by the driver
 *
 * @adev: amdgpu_device pointer
 *
 * Frees the dummy page used by the driver (all asics).
 */
void amdgpu_dummy_page_fini(struct amdgpu_device *adev)
{
	if (adev->dummy_page.page == NULL)
		return;
	pci_unmap_page(adev->pdev, adev->dummy_page.addr,
			PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
	__free_page(adev->dummy_page.page);
	adev->dummy_page.page = NULL;
}


/* ATOM accessor methods */
/*
 * ATOM is an interpreted byte code stored in tables in the vbios.  The
 * driver registers callbacks to access registers and the interpreter
 * in the driver parses the tables and executes then to program specific
 * actions (set display modes, asic init, etc.).  See amdgpu_atombios.c,
 * atombios.h, and atom.c
 */

/**
 * cail_pll_read - read PLL register
 *
 * @info: atom card_info pointer
 * @reg: PLL register offset
 *
 * Provides a PLL register accessor for the atom interpreter (r4xx+).
 * Returns the value of the PLL register.
 */
static uint32_t cail_pll_read(struct card_info *info, uint32_t reg)
{
	return 0;
}

/**
 * cail_pll_write - write PLL register
 *
 * @info: atom card_info pointer
 * @reg: PLL register offset
 * @val: value to write to the pll register
 *
 * Provides a PLL register accessor for the atom interpreter (r4xx+).
 */
static void cail_pll_write(struct card_info *info, uint32_t reg, uint32_t val)
{

}

/**
 * cail_mc_read - read MC (Memory Controller) register
 *
 * @info: atom card_info pointer
 * @reg: MC register offset
 *
 * Provides an MC register accessor for the atom interpreter (r4xx+).
 * Returns the value of the MC register.
 */
static uint32_t cail_mc_read(struct card_info *info, uint32_t reg)
{
	return 0;
}

/**
 * cail_mc_write - write MC (Memory Controller) register
 *
 * @info: atom card_info pointer
 * @reg: MC register offset
 * @val: value to write to the pll register
 *
 * Provides a MC register accessor for the atom interpreter (r4xx+).
 */
static void cail_mc_write(struct card_info *info, uint32_t reg, uint32_t val)
{

}

/**
 * cail_reg_write - write MMIO register
 *
 * @info: atom card_info pointer
 * @reg: MMIO register offset
 * @val: value to write to the pll register
 *
 * Provides a MMIO register accessor for the atom interpreter (r4xx+).
 */
static void cail_reg_write(struct card_info *info, uint32_t reg, uint32_t val)
{
	struct amdgpu_device *adev = info->dev->dev_private;

	WREG32(reg, val);
}

/**
 * cail_reg_read - read MMIO register
 *
 * @info: atom card_info pointer
 * @reg: MMIO register offset
 *
 * Provides an MMIO register accessor for the atom interpreter (r4xx+).
 * Returns the value of the MMIO register.
 */
static uint32_t cail_reg_read(struct card_info *info, uint32_t reg)
{
	struct amdgpu_device *adev = info->dev->dev_private;
	uint32_t r;

	r = RREG32(reg);
	return r;
}

/**
 * cail_ioreg_write - write IO register
 *
 * @info: atom card_info pointer
 * @reg: IO register offset
 * @val: value to write to the pll register
 *
 * Provides a IO register accessor for the atom interpreter (r4xx+).
 */
static void cail_ioreg_write(struct card_info *info, uint32_t reg, uint32_t val)
{
	struct amdgpu_device *adev = info->dev->dev_private;

	WREG32_IO(reg, val);
}

/**
 * cail_ioreg_read - read IO register
 *
 * @info: atom card_info pointer
 * @reg: IO register offset
 *
 * Provides an IO register accessor for the atom interpreter (r4xx+).
 * Returns the value of the IO register.
 */
static uint32_t cail_ioreg_read(struct card_info *info, uint32_t reg)
{
	struct amdgpu_device *adev = info->dev->dev_private;
	uint32_t r;

	r = RREG32_IO(reg);
	return r;
}

/**
 * amdgpu_atombios_fini - free the driver info and callbacks for atombios
 *
 * @adev: amdgpu_device pointer
 *
 * Frees the driver info and register access callbacks for the ATOM
 * interpreter (r4xx+).
 * Called at driver shutdown.
 */
static void amdgpu_atombios_fini(struct amdgpu_device *adev)
{
	if (adev->mode_info.atom_context)
		kfree(adev->mode_info.atom_context->scratch);
	kfree(adev->mode_info.atom_context);
	adev->mode_info.atom_context = NULL;
	kfree(adev->mode_info.atom_card_info);
	adev->mode_info.atom_card_info = NULL;
}

/**
 * amdgpu_atombios_init - init the driver info and callbacks for atombios
 *
 * @adev: amdgpu_device pointer
 *
 * Initializes the driver info and register access callbacks for the
 * ATOM interpreter (r4xx+).
 * Returns 0 on sucess, -ENOMEM on failure.
 * Called at driver startup.
 */
static int amdgpu_atombios_init(struct amdgpu_device *adev)
{
	struct card_info *atom_card_info =
	    kzalloc(sizeof(struct card_info), GFP_KERNEL);

	if (!atom_card_info)
		return -ENOMEM;

	adev->mode_info.atom_card_info = atom_card_info;
	atom_card_info->dev = adev->ddev;
	atom_card_info->reg_read = cail_reg_read;
	atom_card_info->reg_write = cail_reg_write;
	/* needed for iio ops */
	if (adev->rio_mem) {
		atom_card_info->ioreg_read = cail_ioreg_read;
		atom_card_info->ioreg_write = cail_ioreg_write;
	} else {
		DRM_ERROR("Unable to find PCI I/O BAR; using MMIO for ATOM IIO\n");
		atom_card_info->ioreg_read = cail_reg_read;
		atom_card_info->ioreg_write = cail_reg_write;
	}
	atom_card_info->mc_read = cail_mc_read;
	atom_card_info->mc_write = cail_mc_write;
	atom_card_info->pll_read = cail_pll_read;
	atom_card_info->pll_write = cail_pll_write;

	adev->mode_info.atom_context = amdgpu_atom_parse(atom_card_info, adev->bios);
	if (!adev->mode_info.atom_context) {
		amdgpu_atombios_fini(adev);
		return -ENOMEM;
	}

	mutex_init(&adev->mode_info.atom_context->mutex);
	amdgpu_atombios_scratch_regs_init(adev);
	amdgpu_atom_allocate_fb_scratch(adev->mode_info.atom_context);
	return 0;
}

/* if we get transitioned to only one device, take VGA back */
/**
 * amdgpu_vga_set_decode - enable/disable vga decode
 *
 * @cookie: amdgpu_device pointer
 * @state: enable/disable vga decode
 *
 * Enable/disable vga decode (all asics).
 * Returns VGA resource flags.
 */
static unsigned int amdgpu_vga_set_decode(void *cookie, bool state)
{
	struct amdgpu_device *adev = cookie;
	amdgpu_asic_set_vga_state(adev, state);
	if (state)
		return VGA_RSRC_LEGACY_IO | VGA_RSRC_LEGACY_MEM |
		       VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM;
	else
		return VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM;
}

/**
 * amdgpu_check_pot_argument - check that argument is a power of two
 *
 * @arg: value to check
 *
 * Validates that a certain argument is a power of two (all asics).
 * Returns true if argument is valid.
 */
static bool amdgpu_check_pot_argument(int arg)
{
	return (arg & (arg - 1)) == 0;
}

/**
 * amdgpu_check_arguments - validate module params
 *
 * @adev: amdgpu_device pointer
 *
 * Validates certain module parameters and updates
 * the associated values used by the driver (all asics).
 */
static void amdgpu_check_arguments(struct amdgpu_device *adev)
{
953 954 955 956 957 958 959 960 961
	if (amdgpu_sched_jobs < 4) {
		dev_warn(adev->dev, "sched jobs (%d) must be at least 4\n",
			 amdgpu_sched_jobs);
		amdgpu_sched_jobs = 4;
	} else if (!amdgpu_check_pot_argument(amdgpu_sched_jobs)){
		dev_warn(adev->dev, "sched jobs (%d) must be a power of 2\n",
			 amdgpu_sched_jobs);
		amdgpu_sched_jobs = roundup_pow_of_two(amdgpu_sched_jobs);
	}
A
Alex Deucher 已提交
962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984
	/* vramlimit must be a power of two */
	if (!amdgpu_check_pot_argument(amdgpu_vram_limit)) {
		dev_warn(adev->dev, "vram limit (%d) must be a power of 2\n",
				amdgpu_vram_limit);
		amdgpu_vram_limit = 0;
	}

	if (amdgpu_gart_size != -1) {
		/* gtt size must be power of two and greater or equal to 32M */
		if (amdgpu_gart_size < 32) {
			dev_warn(adev->dev, "gart size (%d) too small\n",
				 amdgpu_gart_size);
			amdgpu_gart_size = -1;
		} else if (!amdgpu_check_pot_argument(amdgpu_gart_size)) {
			dev_warn(adev->dev, "gart size (%d) must be a power of 2\n",
				 amdgpu_gart_size);
			amdgpu_gart_size = -1;
		}
	}

	if (!amdgpu_check_pot_argument(amdgpu_vm_size)) {
		dev_warn(adev->dev, "VM size (%d) must be a power of 2\n",
			 amdgpu_vm_size);
985
		amdgpu_vm_size = 8;
A
Alex Deucher 已提交
986 987 988 989 990
	}

	if (amdgpu_vm_size < 1) {
		dev_warn(adev->dev, "VM size (%d) too small, min is 1GB\n",
			 amdgpu_vm_size);
991
		amdgpu_vm_size = 8;
A
Alex Deucher 已提交
992 993 994 995 996 997 998 999
	}

	/*
	 * Max GPUVM size for Cayman, SI and CI are 40 bits.
	 */
	if (amdgpu_vm_size > 1024) {
		dev_warn(adev->dev, "VM size (%d) too large, max is 1TB\n",
			 amdgpu_vm_size);
1000
		amdgpu_vm_size = 8;
A
Alex Deucher 已提交
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
	}

	/* defines number of bits in page table versus page directory,
	 * a page is 4KB so we have 12 bits offset, minimum 9 bits in the
	 * page table and the remaining bits are in the page directory */
	if (amdgpu_vm_block_size == -1) {

		/* Total bits covered by PD + PTs */
		unsigned bits = ilog2(amdgpu_vm_size) + 18;

		/* Make sure the PD is 4K in size up to 8GB address space.
		   Above that split equal between PD and PTs */
		if (amdgpu_vm_size <= 8)
			amdgpu_vm_block_size = bits - 9;
		else
			amdgpu_vm_block_size = (bits + 3) / 2;

	} else if (amdgpu_vm_block_size < 9) {
		dev_warn(adev->dev, "VM page table size (%d) too small\n",
			 amdgpu_vm_block_size);
		amdgpu_vm_block_size = 9;
	}

	if (amdgpu_vm_block_size > 24 ||
	    (amdgpu_vm_size * 1024) < (1ull << amdgpu_vm_block_size)) {
		dev_warn(adev->dev, "VM page table size (%d) too large\n",
			 amdgpu_vm_block_size);
		amdgpu_vm_block_size = 9;
	}
}

/**
 * amdgpu_switcheroo_set_state - set switcheroo state
 *
 * @pdev: pci dev pointer
1036
 * @state: vga_switcheroo state
A
Alex Deucher 已提交
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097
 *
 * Callback for the switcheroo driver.  Suspends or resumes the
 * the asics before or after it is powered up using ACPI methods.
 */
static void amdgpu_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_state state)
{
	struct drm_device *dev = pci_get_drvdata(pdev);

	if (amdgpu_device_is_px(dev) && state == VGA_SWITCHEROO_OFF)
		return;

	if (state == VGA_SWITCHEROO_ON) {
		unsigned d3_delay = dev->pdev->d3_delay;

		printk(KERN_INFO "amdgpu: switched on\n");
		/* don't suspend or resume card normally */
		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;

		amdgpu_resume_kms(dev, true, true);

		dev->pdev->d3_delay = d3_delay;

		dev->switch_power_state = DRM_SWITCH_POWER_ON;
		drm_kms_helper_poll_enable(dev);
	} else {
		printk(KERN_INFO "amdgpu: switched off\n");
		drm_kms_helper_poll_disable(dev);
		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
		amdgpu_suspend_kms(dev, true, true);
		dev->switch_power_state = DRM_SWITCH_POWER_OFF;
	}
}

/**
 * amdgpu_switcheroo_can_switch - see if switcheroo state can change
 *
 * @pdev: pci dev pointer
 *
 * Callback for the switcheroo driver.  Check of the switcheroo
 * state can be changed.
 * Returns true if the state can be changed, false if not.
 */
static bool amdgpu_switcheroo_can_switch(struct pci_dev *pdev)
{
	struct drm_device *dev = pci_get_drvdata(pdev);

	/*
	* FIXME: open_count is protected by drm_global_mutex but that would lead to
	* locking inversion with the driver load path. And the access here is
	* completely racy anyway. So don't bother with locking for now.
	*/
	return dev->open_count == 0;
}

static const struct vga_switcheroo_client_ops amdgpu_switcheroo_ops = {
	.set_gpu_state = amdgpu_switcheroo_set_state,
	.reprobe = NULL,
	.can_switch = amdgpu_switcheroo_can_switch,
};

int amdgpu_set_clockgating_state(struct amdgpu_device *adev,
1098 1099
				  enum amd_ip_block_type block_type,
				  enum amd_clockgating_state state)
A
Alex Deucher 已提交
1100 1101 1102 1103 1104
{
	int i, r = 0;

	for (i = 0; i < adev->num_ip_blocks; i++) {
		if (adev->ip_blocks[i].type == block_type) {
1105
			r = adev->ip_blocks[i].funcs->set_clockgating_state((void *)adev,
A
Alex Deucher 已提交
1106 1107 1108 1109 1110 1111 1112 1113 1114
									    state);
			if (r)
				return r;
		}
	}
	return r;
}

int amdgpu_set_powergating_state(struct amdgpu_device *adev,
1115 1116
				  enum amd_ip_block_type block_type,
				  enum amd_powergating_state state)
A
Alex Deucher 已提交
1117 1118 1119 1120 1121
{
	int i, r = 0;

	for (i = 0; i < adev->num_ip_blocks; i++) {
		if (adev->ip_blocks[i].type == block_type) {
1122
			r = adev->ip_blocks[i].funcs->set_powergating_state((void *)adev,
A
Alex Deucher 已提交
1123 1124 1125 1126 1127 1128 1129 1130 1131 1132
									    state);
			if (r)
				return r;
		}
	}
	return r;
}

const struct amdgpu_ip_block_version * amdgpu_get_ip_block(
					struct amdgpu_device *adev,
1133
					enum amd_ip_block_type type)
A
Alex Deucher 已提交
1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147
{
	int i;

	for (i = 0; i < adev->num_ip_blocks; i++)
		if (adev->ip_blocks[i].type == type)
			return &adev->ip_blocks[i];

	return NULL;
}

/**
 * amdgpu_ip_block_version_cmp
 *
 * @adev: amdgpu_device pointer
1148
 * @type: enum amd_ip_block_type
A
Alex Deucher 已提交
1149 1150 1151 1152 1153 1154 1155
 * @major: major version
 * @minor: minor version
 *
 * return 0 if equal or greater
 * return 1 if smaller or the ip_block doesn't exist
 */
int amdgpu_ip_block_version_cmp(struct amdgpu_device *adev,
1156
				enum amd_ip_block_type type,
A
Alex Deucher 已提交
1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171
				u32 major, u32 minor)
{
	const struct amdgpu_ip_block_version *ip_block;
	ip_block = amdgpu_get_ip_block(adev, type);

	if (ip_block && ((ip_block->major > major) ||
			((ip_block->major == major) &&
			(ip_block->minor >= minor))))
		return 0;

	return 1;
}

static int amdgpu_early_init(struct amdgpu_device *adev)
{
1172
	int i, r;
A
Alex Deucher 已提交
1173 1174

	switch (adev->asic_type) {
1175 1176
	case CHIP_TOPAZ:
	case CHIP_TONGA:
1177
	case CHIP_FIJI:
1178
	case CHIP_CARRIZO:
1179 1180
	case CHIP_STONEY:
		if (adev->asic_type == CHIP_CARRIZO || adev->asic_type == CHIP_STONEY)
1181 1182 1183 1184 1185 1186 1187 1188
			adev->family = AMDGPU_FAMILY_CZ;
		else
			adev->family = AMDGPU_FAMILY_VI;

		r = vi_set_ip_blocks(adev);
		if (r)
			return r;
		break;
1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204
#ifdef CONFIG_DRM_AMDGPU_CIK
	case CHIP_BONAIRE:
	case CHIP_HAWAII:
	case CHIP_KAVERI:
	case CHIP_KABINI:
	case CHIP_MULLINS:
		if ((adev->asic_type == CHIP_BONAIRE) || (adev->asic_type == CHIP_HAWAII))
			adev->family = AMDGPU_FAMILY_CI;
		else
			adev->family = AMDGPU_FAMILY_KV;

		r = cik_set_ip_blocks(adev);
		if (r)
			return r;
		break;
#endif
A
Alex Deucher 已提交
1205 1206 1207 1208 1209
	default:
		/* FIXME: not supported yet */
		return -EINVAL;
	}

1210 1211 1212
	adev->ip_block_status = kcalloc(adev->num_ip_blocks,
					sizeof(struct amdgpu_ip_block_status), GFP_KERNEL);
	if (adev->ip_block_status == NULL)
1213
		return -ENOMEM;
A
Alex Deucher 已提交
1214 1215 1216 1217 1218 1219 1220 1221 1222

	if (adev->ip_blocks == NULL) {
		DRM_ERROR("No IP blocks found!\n");
		return r;
	}

	for (i = 0; i < adev->num_ip_blocks; i++) {
		if ((amdgpu_ip_block_mask & (1 << i)) == 0) {
			DRM_ERROR("disabled ip block: %d\n", i);
1223
			adev->ip_block_status[i].valid = false;
A
Alex Deucher 已提交
1224 1225
		} else {
			if (adev->ip_blocks[i].funcs->early_init) {
1226
				r = adev->ip_blocks[i].funcs->early_init((void *)adev);
1227
				if (r == -ENOENT) {
1228
					adev->ip_block_status[i].valid = false;
1229 1230
				} else if (r) {
					DRM_ERROR("early_init %d failed %d\n", i, r);
A
Alex Deucher 已提交
1231
					return r;
1232
				} else {
1233
					adev->ip_block_status[i].valid = true;
1234
				}
1235
			} else {
1236
				adev->ip_block_status[i].valid = true;
A
Alex Deucher 已提交
1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248
			}
		}
	}

	return 0;
}

static int amdgpu_init(struct amdgpu_device *adev)
{
	int i, r;

	for (i = 0; i < adev->num_ip_blocks; i++) {
1249
		if (!adev->ip_block_status[i].valid)
A
Alex Deucher 已提交
1250
			continue;
1251
		r = adev->ip_blocks[i].funcs->sw_init((void *)adev);
1252 1253
		if (r) {
			DRM_ERROR("sw_init %d failed %d\n", i, r);
A
Alex Deucher 已提交
1254
			return r;
1255
		}
1256
		adev->ip_block_status[i].sw = true;
A
Alex Deucher 已提交
1257
		/* need to do gmc hw init early so we can allocate gpu mem */
1258
		if (adev->ip_blocks[i].type == AMD_IP_BLOCK_TYPE_GMC) {
A
Alex Deucher 已提交
1259
			r = amdgpu_vram_scratch_init(adev);
1260 1261
			if (r) {
				DRM_ERROR("amdgpu_vram_scratch_init failed %d\n", r);
A
Alex Deucher 已提交
1262
				return r;
1263
			}
1264
			r = adev->ip_blocks[i].funcs->hw_init((void *)adev);
1265 1266
			if (r) {
				DRM_ERROR("hw_init %d failed %d\n", i, r);
A
Alex Deucher 已提交
1267
				return r;
1268
			}
A
Alex Deucher 已提交
1269
			r = amdgpu_wb_init(adev);
1270 1271
			if (r) {
				DRM_ERROR("amdgpu_wb_init failed %d\n", r);
A
Alex Deucher 已提交
1272
				return r;
1273
			}
1274
			adev->ip_block_status[i].hw = true;
A
Alex Deucher 已提交
1275 1276 1277 1278
		}
	}

	for (i = 0; i < adev->num_ip_blocks; i++) {
1279
		if (!adev->ip_block_status[i].sw)
A
Alex Deucher 已提交
1280 1281
			continue;
		/* gmc hw init is done early */
1282
		if (adev->ip_blocks[i].type == AMD_IP_BLOCK_TYPE_GMC)
A
Alex Deucher 已提交
1283
			continue;
1284
		r = adev->ip_blocks[i].funcs->hw_init((void *)adev);
1285 1286
		if (r) {
			DRM_ERROR("hw_init %d failed %d\n", i, r);
A
Alex Deucher 已提交
1287
			return r;
1288
		}
1289
		adev->ip_block_status[i].hw = true;
A
Alex Deucher 已提交
1290 1291 1292 1293 1294 1295 1296 1297 1298 1299
	}

	return 0;
}

static int amdgpu_late_init(struct amdgpu_device *adev)
{
	int i = 0, r;

	for (i = 0; i < adev->num_ip_blocks; i++) {
1300
		if (!adev->ip_block_status[i].valid)
A
Alex Deucher 已提交
1301 1302
			continue;
		/* enable clockgating to save power */
1303 1304
		r = adev->ip_blocks[i].funcs->set_clockgating_state((void *)adev,
								    AMD_CG_STATE_GATE);
1305 1306
		if (r) {
			DRM_ERROR("set_clockgating_state(gate) %d failed %d\n", i, r);
A
Alex Deucher 已提交
1307
			return r;
1308
		}
A
Alex Deucher 已提交
1309
		if (adev->ip_blocks[i].funcs->late_init) {
1310
			r = adev->ip_blocks[i].funcs->late_init((void *)adev);
1311 1312
			if (r) {
				DRM_ERROR("late_init %d failed %d\n", i, r);
A
Alex Deucher 已提交
1313
				return r;
1314
			}
A
Alex Deucher 已提交
1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325
		}
	}

	return 0;
}

static int amdgpu_fini(struct amdgpu_device *adev)
{
	int i, r;

	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
1326
		if (!adev->ip_block_status[i].hw)
A
Alex Deucher 已提交
1327
			continue;
1328
		if (adev->ip_blocks[i].type == AMD_IP_BLOCK_TYPE_GMC) {
A
Alex Deucher 已提交
1329 1330 1331 1332
			amdgpu_wb_fini(adev);
			amdgpu_vram_scratch_fini(adev);
		}
		/* ungate blocks before hw fini so that we can shutdown the blocks safely */
1333 1334
		r = adev->ip_blocks[i].funcs->set_clockgating_state((void *)adev,
								    AMD_CG_STATE_UNGATE);
1335 1336
		if (r) {
			DRM_ERROR("set_clockgating_state(ungate) %d failed %d\n", i, r);
A
Alex Deucher 已提交
1337
			return r;
1338
		}
1339
		r = adev->ip_blocks[i].funcs->hw_fini((void *)adev);
A
Alex Deucher 已提交
1340
		/* XXX handle errors */
1341 1342 1343
		if (r) {
			DRM_DEBUG("hw_fini %d failed %d\n", i, r);
		}
1344
		adev->ip_block_status[i].hw = false;
A
Alex Deucher 已提交
1345 1346 1347
	}

	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
1348
		if (!adev->ip_block_status[i].sw)
A
Alex Deucher 已提交
1349
			continue;
1350
		r = adev->ip_blocks[i].funcs->sw_fini((void *)adev);
A
Alex Deucher 已提交
1351
		/* XXX handle errors */
1352 1353 1354
		if (r) {
			DRM_DEBUG("sw_fini %d failed %d\n", i, r);
		}
1355 1356
		adev->ip_block_status[i].sw = false;
		adev->ip_block_status[i].valid = false;
A
Alex Deucher 已提交
1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
	}

	return 0;
}

static int amdgpu_suspend(struct amdgpu_device *adev)
{
	int i, r;

	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
1367
		if (!adev->ip_block_status[i].valid)
A
Alex Deucher 已提交
1368 1369
			continue;
		/* ungate blocks so that suspend can properly shut them down */
1370 1371
		r = adev->ip_blocks[i].funcs->set_clockgating_state((void *)adev,
								    AMD_CG_STATE_UNGATE);
1372 1373 1374
		if (r) {
			DRM_ERROR("set_clockgating_state(ungate) %d failed %d\n", i, r);
		}
A
Alex Deucher 已提交
1375 1376 1377
		/* XXX handle errors */
		r = adev->ip_blocks[i].funcs->suspend(adev);
		/* XXX handle errors */
1378 1379 1380
		if (r) {
			DRM_ERROR("suspend %d failed %d\n", i, r);
		}
A
Alex Deucher 已提交
1381 1382 1383 1384 1385 1386 1387 1388 1389 1390
	}

	return 0;
}

static int amdgpu_resume(struct amdgpu_device *adev)
{
	int i, r;

	for (i = 0; i < adev->num_ip_blocks; i++) {
1391
		if (!adev->ip_block_status[i].valid)
A
Alex Deucher 已提交
1392 1393
			continue;
		r = adev->ip_blocks[i].funcs->resume(adev);
1394 1395
		if (r) {
			DRM_ERROR("resume %d failed %d\n", i, r);
A
Alex Deucher 已提交
1396
			return r;
1397
		}
A
Alex Deucher 已提交
1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427
	}

	return 0;
}

/**
 * amdgpu_device_init - initialize the driver
 *
 * @adev: amdgpu_device pointer
 * @pdev: drm dev pointer
 * @pdev: pci dev pointer
 * @flags: driver flags
 *
 * Initializes the driver info and hw (all asics).
 * Returns 0 for success or an error on failure.
 * Called at driver startup.
 */
int amdgpu_device_init(struct amdgpu_device *adev,
		       struct drm_device *ddev,
		       struct pci_dev *pdev,
		       uint32_t flags)
{
	int r, i;
	bool runtime = false;

	adev->shutdown = false;
	adev->dev = &pdev->dev;
	adev->ddev = ddev;
	adev->pdev = pdev;
	adev->flags = flags;
1428
	adev->asic_type = flags & AMD_ASIC_MASK;
A
Alex Deucher 已提交
1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
	adev->is_atom_bios = false;
	adev->usec_timeout = AMDGPU_MAX_USEC_TIMEOUT;
	adev->mc.gtt_size = 512 * 1024 * 1024;
	adev->accel_working = false;
	adev->num_rings = 0;
	adev->mman.buffer_funcs = NULL;
	adev->mman.buffer_funcs_ring = NULL;
	adev->vm_manager.vm_pte_funcs = NULL;
	adev->vm_manager.vm_pte_funcs_ring = NULL;
	adev->gart.gart_funcs = NULL;
	adev->fence_context = fence_context_alloc(AMDGPU_MAX_RINGS);

	adev->smc_rreg = &amdgpu_invalid_rreg;
	adev->smc_wreg = &amdgpu_invalid_wreg;
	adev->pcie_rreg = &amdgpu_invalid_rreg;
	adev->pcie_wreg = &amdgpu_invalid_wreg;
	adev->uvd_ctx_rreg = &amdgpu_invalid_rreg;
	adev->uvd_ctx_wreg = &amdgpu_invalid_wreg;
	adev->didt_rreg = &amdgpu_invalid_rreg;
	adev->didt_wreg = &amdgpu_invalid_wreg;
	adev->audio_endpt_rreg = &amdgpu_block_invalid_rreg;
	adev->audio_endpt_wreg = &amdgpu_block_invalid_wreg;

1452 1453 1454
	DRM_INFO("initializing kernel modesetting (%s 0x%04X:0x%04X 0x%04X:0x%04X 0x%02X).\n",
		 amdgpu_asic_name[adev->asic_type], pdev->vendor, pdev->device,
		 pdev->subsystem_vendor, pdev->subsystem_device, pdev->revision);
A
Alex Deucher 已提交
1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528

	/* mutex initialization are all done here so we
	 * can recall function without having locking issues */
	mutex_init(&adev->ring_lock);
	atomic_set(&adev->irq.ih.lock, 0);
	mutex_init(&adev->gem.mutex);
	mutex_init(&adev->pm.mutex);
	mutex_init(&adev->gfx.gpu_clock_mutex);
	mutex_init(&adev->srbm_mutex);
	mutex_init(&adev->grbm_idx_mutex);
	mutex_init(&adev->mn_lock);
	hash_init(adev->mn_hash);

	amdgpu_check_arguments(adev);

	/* Registers mapping */
	/* TODO: block userspace mapping of io register */
	spin_lock_init(&adev->mmio_idx_lock);
	spin_lock_init(&adev->smc_idx_lock);
	spin_lock_init(&adev->pcie_idx_lock);
	spin_lock_init(&adev->uvd_ctx_idx_lock);
	spin_lock_init(&adev->didt_idx_lock);
	spin_lock_init(&adev->audio_endpt_idx_lock);

	adev->rmmio_base = pci_resource_start(adev->pdev, 5);
	adev->rmmio_size = pci_resource_len(adev->pdev, 5);
	adev->rmmio = ioremap(adev->rmmio_base, adev->rmmio_size);
	if (adev->rmmio == NULL) {
		return -ENOMEM;
	}
	DRM_INFO("register mmio base: 0x%08X\n", (uint32_t)adev->rmmio_base);
	DRM_INFO("register mmio size: %u\n", (unsigned)adev->rmmio_size);

	/* doorbell bar mapping */
	amdgpu_doorbell_init(adev);

	/* io port mapping */
	for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
		if (pci_resource_flags(adev->pdev, i) & IORESOURCE_IO) {
			adev->rio_mem_size = pci_resource_len(adev->pdev, i);
			adev->rio_mem = pci_iomap(adev->pdev, i, adev->rio_mem_size);
			break;
		}
	}
	if (adev->rio_mem == NULL)
		DRM_ERROR("Unable to find PCI I/O BAR\n");

	/* early init functions */
	r = amdgpu_early_init(adev);
	if (r)
		return r;

	/* if we have > 1 VGA cards, then disable the amdgpu VGA resources */
	/* this will fail for cards that aren't VGA class devices, just
	 * ignore it */
	vga_client_register(adev->pdev, adev, NULL, amdgpu_vga_set_decode);

	if (amdgpu_runtime_pm == 1)
		runtime = true;
	if (amdgpu_device_is_px(ddev))
		runtime = true;
	vga_switcheroo_register_client(adev->pdev, &amdgpu_switcheroo_ops, runtime);
	if (runtime)
		vga_switcheroo_init_domain_pm_ops(adev->dev, &adev->vga_pm_domain);

	/* Read BIOS */
	if (!amdgpu_get_bios(adev))
		return -EINVAL;
	/* Must be an ATOMBIOS */
	if (!adev->is_atom_bios) {
		dev_err(adev->dev, "Expecting atombios for GPU\n");
		return -EINVAL;
	}
	r = amdgpu_atombios_init(adev);
1529 1530
	if (r) {
		dev_err(adev->dev, "amdgpu_atombios_init failed\n");
A
Alex Deucher 已提交
1531
		return r;
1532
	}
A
Alex Deucher 已提交
1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545

	/* Post card if necessary */
	if (!amdgpu_card_posted(adev)) {
		if (!adev->bios) {
			dev_err(adev->dev, "Card not posted and no BIOS - ignoring\n");
			return -EINVAL;
		}
		DRM_INFO("GPU not posted. posting now...\n");
		amdgpu_atom_asic_init(adev->mode_info.atom_context);
	}

	/* Initialize clocks */
	r = amdgpu_atombios_get_clock_info(adev);
1546 1547
	if (r) {
		dev_err(adev->dev, "amdgpu_atombios_get_clock_info failed\n");
A
Alex Deucher 已提交
1548
		return r;
1549
	}
A
Alex Deucher 已提交
1550 1551 1552 1553 1554
	/* init i2c buses */
	amdgpu_atombios_i2c_init(adev);

	/* Fence driver */
	r = amdgpu_fence_driver_init(adev);
1555 1556
	if (r) {
		dev_err(adev->dev, "amdgpu_fence_driver_init failed\n");
A
Alex Deucher 已提交
1557
		return r;
1558
	}
A
Alex Deucher 已提交
1559 1560 1561 1562 1563 1564

	/* init the mode config */
	drm_mode_config_init(adev->ddev);

	r = amdgpu_init(adev);
	if (r) {
1565
		dev_err(adev->dev, "amdgpu_init failed\n");
A
Alex Deucher 已提交
1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579
		amdgpu_fini(adev);
		return r;
	}

	adev->accel_working = true;

	amdgpu_fbdev_init(adev);

	r = amdgpu_ib_pool_init(adev);
	if (r) {
		dev_err(adev->dev, "IB initialization failed (%d).\n", r);
		return r;
	}

1580
	r = amdgpu_ctx_init(adev, AMD_SCHED_PRIORITY_KERNEL, &adev->kernel_ctx);
1581 1582 1583
	if (r) {
		dev_err(adev->dev, "failed to create kernel context (%d).\n", r);
		return r;
1584
	}
A
Alex Deucher 已提交
1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621
	r = amdgpu_ib_ring_tests(adev);
	if (r)
		DRM_ERROR("ib ring test failed (%d).\n", r);

	r = amdgpu_gem_debugfs_init(adev);
	if (r) {
		DRM_ERROR("registering gem debugfs failed (%d).\n", r);
	}

	r = amdgpu_debugfs_regs_init(adev);
	if (r) {
		DRM_ERROR("registering register debugfs failed (%d).\n", r);
	}

	if ((amdgpu_testing & 1)) {
		if (adev->accel_working)
			amdgpu_test_moves(adev);
		else
			DRM_INFO("amdgpu: acceleration disabled, skipping move tests\n");
	}
	if ((amdgpu_testing & 2)) {
		if (adev->accel_working)
			amdgpu_test_syncing(adev);
		else
			DRM_INFO("amdgpu: acceleration disabled, skipping sync tests\n");
	}
	if (amdgpu_benchmarking) {
		if (adev->accel_working)
			amdgpu_benchmark(adev, amdgpu_benchmarking);
		else
			DRM_INFO("amdgpu: acceleration disabled, skipping benchmarks\n");
	}

	/* enable clockgating, etc. after ib tests, etc. since some blocks require
	 * explicit gating rather than handling it automatically.
	 */
	r = amdgpu_late_init(adev);
1622 1623
	if (r) {
		dev_err(adev->dev, "amdgpu_late_init failed\n");
A
Alex Deucher 已提交
1624
		return r;
1625
	}
A
Alex Deucher 已提交
1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647

	return 0;
}

static void amdgpu_debugfs_remove_files(struct amdgpu_device *adev);

/**
 * amdgpu_device_fini - tear down the driver
 *
 * @adev: amdgpu_device pointer
 *
 * Tear down the driver info (all asics).
 * Called at driver shutdown.
 */
void amdgpu_device_fini(struct amdgpu_device *adev)
{
	int r;

	DRM_INFO("amdgpu: finishing device.\n");
	adev->shutdown = true;
	/* evict vram memory */
	amdgpu_bo_evict_vram(adev);
1648
	amdgpu_ctx_fini(&adev->kernel_ctx);
A
Alex Deucher 已提交
1649 1650 1651 1652
	amdgpu_ib_pool_fini(adev);
	amdgpu_fence_driver_fini(adev);
	amdgpu_fbdev_fini(adev);
	r = amdgpu_fini(adev);
1653 1654
	kfree(adev->ip_block_status);
	adev->ip_block_status = NULL;
A
Alex Deucher 已提交
1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691
	adev->accel_working = false;
	/* free i2c buses */
	amdgpu_i2c_fini(adev);
	amdgpu_atombios_fini(adev);
	kfree(adev->bios);
	adev->bios = NULL;
	vga_switcheroo_unregister_client(adev->pdev);
	vga_client_register(adev->pdev, NULL, NULL, NULL);
	if (adev->rio_mem)
		pci_iounmap(adev->pdev, adev->rio_mem);
	adev->rio_mem = NULL;
	iounmap(adev->rmmio);
	adev->rmmio = NULL;
	amdgpu_doorbell_fini(adev);
	amdgpu_debugfs_regs_cleanup(adev);
	amdgpu_debugfs_remove_files(adev);
}


/*
 * Suspend & resume.
 */
/**
 * amdgpu_suspend_kms - initiate device suspend
 *
 * @pdev: drm dev pointer
 * @state: suspend state
 *
 * Puts the hw in the suspend state (all asics).
 * Returns 0 for success or an error on failure.
 * Called at driver suspend.
 */
int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
{
	struct amdgpu_device *adev;
	struct drm_crtc *crtc;
	struct drm_connector *connector;
1692
	int r;
A
Alex Deucher 已提交
1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705

	if (dev == NULL || dev->dev_private == NULL) {
		return -ENODEV;
	}

	adev = dev->dev_private;

	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
		return 0;

	drm_kms_helper_poll_disable(dev);

	/* turn off display hw */
1706
	drm_modeset_lock_all(dev);
A
Alex Deucher 已提交
1707 1708 1709
	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
		drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
	}
1710
	drm_modeset_unlock_all(dev);
A
Alex Deucher 已提交
1711

1712
	/* unpin the front buffers and cursors */
A
Alex Deucher 已提交
1713
	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
1714
		struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
A
Alex Deucher 已提交
1715 1716 1717
		struct amdgpu_framebuffer *rfb = to_amdgpu_framebuffer(crtc->primary->fb);
		struct amdgpu_bo *robj;

1718 1719 1720 1721 1722 1723 1724 1725 1726
		if (amdgpu_crtc->cursor_bo) {
			struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
			r = amdgpu_bo_reserve(aobj, false);
			if (r == 0) {
				amdgpu_bo_unpin(aobj);
				amdgpu_bo_unreserve(aobj);
			}
		}

A
Alex Deucher 已提交
1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742
		if (rfb == NULL || rfb->obj == NULL) {
			continue;
		}
		robj = gem_to_amdgpu_bo(rfb->obj);
		/* don't unpin kernel fb objects */
		if (!amdgpu_fbdev_robj_is_fb(adev, robj)) {
			r = amdgpu_bo_reserve(robj, false);
			if (r == 0) {
				amdgpu_bo_unpin(robj);
				amdgpu_bo_unreserve(robj);
			}
		}
	}
	/* evict vram memory */
	amdgpu_bo_evict_vram(adev);

1743
	amdgpu_fence_driver_suspend(adev);
A
Alex Deucher 已提交
1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777

	r = amdgpu_suspend(adev);

	/* evict remaining vram memory */
	amdgpu_bo_evict_vram(adev);

	pci_save_state(dev->pdev);
	if (suspend) {
		/* Shut down the device */
		pci_disable_device(dev->pdev);
		pci_set_power_state(dev->pdev, PCI_D3hot);
	}

	if (fbcon) {
		console_lock();
		amdgpu_fbdev_set_suspend(adev, 1);
		console_unlock();
	}
	return 0;
}

/**
 * amdgpu_resume_kms - initiate device resume
 *
 * @pdev: drm dev pointer
 *
 * Bring the hw back to operating state (all asics).
 * Returns 0 for success or an error on failure.
 * Called at driver resume.
 */
int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon)
{
	struct drm_connector *connector;
	struct amdgpu_device *adev = dev->dev_private;
1778
	struct drm_crtc *crtc;
A
Alex Deucher 已提交
1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801
	int r;

	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
		return 0;

	if (fbcon) {
		console_lock();
	}
	if (resume) {
		pci_set_power_state(dev->pdev, PCI_D0);
		pci_restore_state(dev->pdev);
		if (pci_enable_device(dev->pdev)) {
			if (fbcon)
				console_unlock();
			return -1;
		}
	}

	/* post card */
	amdgpu_atom_asic_init(adev->mode_info.atom_context);

	r = amdgpu_resume(adev);

1802 1803
	amdgpu_fence_driver_resume(adev);

A
Alex Deucher 已提交
1804 1805 1806 1807 1808 1809 1810 1811
	r = amdgpu_ib_ring_tests(adev);
	if (r)
		DRM_ERROR("ib ring test failed (%d).\n", r);

	r = amdgpu_late_init(adev);
	if (r)
		return r;

1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829
	/* pin cursors */
	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
		struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);

		if (amdgpu_crtc->cursor_bo) {
			struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
			r = amdgpu_bo_reserve(aobj, false);
			if (r == 0) {
				r = amdgpu_bo_pin(aobj,
						  AMDGPU_GEM_DOMAIN_VRAM,
						  &amdgpu_crtc->cursor_addr);
				if (r != 0)
					DRM_ERROR("Failed to pin cursor BO (%d)\n", r);
				amdgpu_bo_unreserve(aobj);
			}
		}
	}

A
Alex Deucher 已提交
1830 1831 1832 1833
	/* blat the mode back in */
	if (fbcon) {
		drm_helper_resume_force_mode(dev);
		/* turn on display hw */
1834
		drm_modeset_lock_all(dev);
A
Alex Deucher 已提交
1835 1836 1837
		list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
			drm_helper_connector_dpms(connector, DRM_MODE_DPMS_ON);
		}
1838
		drm_modeset_unlock_all(dev);
A
Alex Deucher 已提交
1839 1840 1841
	}

	drm_kms_helper_poll_enable(dev);
1842
	drm_helper_hpd_irq_event(dev);
A
Alex Deucher 已提交
1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869

	if (fbcon) {
		amdgpu_fbdev_set_suspend(adev, 0);
		console_unlock();
	}

	return 0;
}

/**
 * amdgpu_gpu_reset - reset the asic
 *
 * @adev: amdgpu device pointer
 *
 * Attempt the reset the GPU if it has hung (all asics).
 * Returns 0 for success or an error on failure.
 */
int amdgpu_gpu_reset(struct amdgpu_device *adev)
{
	unsigned ring_sizes[AMDGPU_MAX_RINGS];
	uint32_t *ring_data[AMDGPU_MAX_RINGS];

	bool saved = false;

	int i, r;
	int resched;

1870
	atomic_inc(&adev->gpu_reset_counter);
A
Alex Deucher 已提交
1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935

	/* block TTM */
	resched = ttm_bo_lock_delayed_workqueue(&adev->mman.bdev);

	r = amdgpu_suspend(adev);

	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
		struct amdgpu_ring *ring = adev->rings[i];
		if (!ring)
			continue;

		ring_sizes[i] = amdgpu_ring_backup(ring, &ring_data[i]);
		if (ring_sizes[i]) {
			saved = true;
			dev_info(adev->dev, "Saved %d dwords of commands "
				 "on ring %d.\n", ring_sizes[i], i);
		}
	}

retry:
	r = amdgpu_asic_reset(adev);
	if (!r) {
		dev_info(adev->dev, "GPU reset succeeded, trying to resume\n");
		r = amdgpu_resume(adev);
	}

	if (!r) {
		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
			struct amdgpu_ring *ring = adev->rings[i];
			if (!ring)
				continue;

			amdgpu_ring_restore(ring, ring_sizes[i], ring_data[i]);
			ring_sizes[i] = 0;
			ring_data[i] = NULL;
		}

		r = amdgpu_ib_ring_tests(adev);
		if (r) {
			dev_err(adev->dev, "ib ring test failed (%d).\n", r);
			if (saved) {
				saved = false;
				r = amdgpu_suspend(adev);
				goto retry;
			}
		}
	} else {
		amdgpu_fence_driver_force_completion(adev);
		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
			if (adev->rings[i])
				kfree(ring_data[i]);
		}
	}

	drm_helper_resume_force_mode(adev->ddev);

	ttm_bo_unlock_delayed_workqueue(&adev->mman.bdev, resched);
	if (r) {
		/* bad news, how to tell it to userspace ? */
		dev_info(adev->dev, "GPU reset failed\n");
	}

	return r;
}

1936 1937 1938
#define AMDGPU_DEFAULT_PCIE_GEN_MASK 0x30007  /* gen: chipset 1/2, asic 1/2/3 */
#define AMDGPU_DEFAULT_PCIE_MLW_MASK 0x2f0000 /* 1/2/4/8/16 lanes */

1939 1940 1941 1942 1943
void amdgpu_get_pcie_info(struct amdgpu_device *adev)
{
	u32 mask;
	int ret;

1944 1945
	if (amdgpu_pcie_gen_cap)
		adev->pm.pcie_gen_mask = amdgpu_pcie_gen_cap;
1946

1947 1948
	if (amdgpu_pcie_lane_cap)
		adev->pm.pcie_mlw_mask = amdgpu_pcie_lane_cap;
1949

1950 1951 1952 1953 1954 1955
	/* covers APUs as well */
	if (pci_is_root_bus(adev->pdev->bus)) {
		if (adev->pm.pcie_gen_mask == 0)
			adev->pm.pcie_gen_mask = AMDGPU_DEFAULT_PCIE_GEN_MASK;
		if (adev->pm.pcie_mlw_mask == 0)
			adev->pm.pcie_mlw_mask = AMDGPU_DEFAULT_PCIE_MLW_MASK;
1956
		return;
1957
	}
1958

1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026
	if (adev->pm.pcie_gen_mask == 0) {
		ret = drm_pcie_get_speed_cap_mask(adev->ddev, &mask);
		if (!ret) {
			adev->pm.pcie_gen_mask = (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
						  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 |
						  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3);

			if (mask & DRM_PCIE_SPEED_25)
				adev->pm.pcie_gen_mask |= CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1;
			if (mask & DRM_PCIE_SPEED_50)
				adev->pm.pcie_gen_mask |= CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2;
			if (mask & DRM_PCIE_SPEED_80)
				adev->pm.pcie_gen_mask |= CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3;
		} else {
			adev->pm.pcie_gen_mask = AMDGPU_DEFAULT_PCIE_GEN_MASK;
		}
	}
	if (adev->pm.pcie_mlw_mask == 0) {
		ret = drm_pcie_get_max_link_width(adev->ddev, &mask);
		if (!ret) {
			switch (mask) {
			case 32:
				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X32 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
				break;
			case 16:
				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
				break;
			case 12:
				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
				break;
			case 8:
				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
				break;
			case 4:
				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
				break;
			case 2:
				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
				break;
			case 1:
				adev->pm.pcie_mlw_mask = CAIL_PCIE_LINK_WIDTH_SUPPORT_X1;
				break;
			default:
				break;
			}
		} else {
			adev->pm.pcie_mlw_mask = AMDGPU_DEFAULT_PCIE_MLW_MASK;
2027 2028 2029
		}
	}
}
A
Alex Deucher 已提交
2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182

/*
 * Debugfs
 */
int amdgpu_debugfs_add_files(struct amdgpu_device *adev,
			     struct drm_info_list *files,
			     unsigned nfiles)
{
	unsigned i;

	for (i = 0; i < adev->debugfs_count; i++) {
		if (adev->debugfs[i].files == files) {
			/* Already registered */
			return 0;
		}
	}

	i = adev->debugfs_count + 1;
	if (i > AMDGPU_DEBUGFS_MAX_COMPONENTS) {
		DRM_ERROR("Reached maximum number of debugfs components.\n");
		DRM_ERROR("Report so we increase "
			  "AMDGPU_DEBUGFS_MAX_COMPONENTS.\n");
		return -EINVAL;
	}
	adev->debugfs[adev->debugfs_count].files = files;
	adev->debugfs[adev->debugfs_count].num_files = nfiles;
	adev->debugfs_count = i;
#if defined(CONFIG_DEBUG_FS)
	drm_debugfs_create_files(files, nfiles,
				 adev->ddev->control->debugfs_root,
				 adev->ddev->control);
	drm_debugfs_create_files(files, nfiles,
				 adev->ddev->primary->debugfs_root,
				 adev->ddev->primary);
#endif
	return 0;
}

static void amdgpu_debugfs_remove_files(struct amdgpu_device *adev)
{
#if defined(CONFIG_DEBUG_FS)
	unsigned i;

	for (i = 0; i < adev->debugfs_count; i++) {
		drm_debugfs_remove_files(adev->debugfs[i].files,
					 adev->debugfs[i].num_files,
					 adev->ddev->control);
		drm_debugfs_remove_files(adev->debugfs[i].files,
					 adev->debugfs[i].num_files,
					 adev->ddev->primary);
	}
#endif
}

#if defined(CONFIG_DEBUG_FS)

static ssize_t amdgpu_debugfs_regs_read(struct file *f, char __user *buf,
					size_t size, loff_t *pos)
{
	struct amdgpu_device *adev = f->f_inode->i_private;
	ssize_t result = 0;
	int r;

	if (size & 0x3 || *pos & 0x3)
		return -EINVAL;

	while (size) {
		uint32_t value;

		if (*pos > adev->rmmio_size)
			return result;

		value = RREG32(*pos >> 2);
		r = put_user(value, (uint32_t *)buf);
		if (r)
			return r;

		result += 4;
		buf += 4;
		*pos += 4;
		size -= 4;
	}

	return result;
}

static ssize_t amdgpu_debugfs_regs_write(struct file *f, const char __user *buf,
					 size_t size, loff_t *pos)
{
	struct amdgpu_device *adev = f->f_inode->i_private;
	ssize_t result = 0;
	int r;

	if (size & 0x3 || *pos & 0x3)
		return -EINVAL;

	while (size) {
		uint32_t value;

		if (*pos > adev->rmmio_size)
			return result;

		r = get_user(value, (uint32_t *)buf);
		if (r)
			return r;

		WREG32(*pos >> 2, value);

		result += 4;
		buf += 4;
		*pos += 4;
		size -= 4;
	}

	return result;
}

static const struct file_operations amdgpu_debugfs_regs_fops = {
	.owner = THIS_MODULE,
	.read = amdgpu_debugfs_regs_read,
	.write = amdgpu_debugfs_regs_write,
	.llseek = default_llseek
};

static int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
{
	struct drm_minor *minor = adev->ddev->primary;
	struct dentry *ent, *root = minor->debugfs_root;

	ent = debugfs_create_file("amdgpu_regs", S_IFREG | S_IRUGO, root,
				  adev, &amdgpu_debugfs_regs_fops);
	if (IS_ERR(ent))
		return PTR_ERR(ent);
	i_size_write(ent->d_inode, adev->rmmio_size);
	adev->debugfs_regs = ent;

	return 0;
}

static void amdgpu_debugfs_regs_cleanup(struct amdgpu_device *adev)
{
	debugfs_remove(adev->debugfs_regs);
	adev->debugfs_regs = NULL;
}

int amdgpu_debugfs_init(struct drm_minor *minor)
{
	return 0;
}

void amdgpu_debugfs_cleanup(struct drm_minor *minor)
{
}
2183 2184 2185 2186 2187 2188
#else
static int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
{
	return 0;
}
static void amdgpu_debugfs_regs_cleanup(struct amdgpu_device *adev) { }
A
Alex Deucher 已提交
2189
#endif