gmc_v7_0.c 37.9 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
/*
 * Copyright 2014 Advanced Micro Devices, Inc.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
 * OTHER DEALINGS IN THE SOFTWARE.
 *
 */
#include <linux/firmware.h>
#include "drmP.h"
#include "amdgpu.h"
#include "cikd.h"
#include "cik.h"
#include "gmc_v7_0.h"
#include "amdgpu_ucode.h"

#include "bif/bif_4_1_d.h"
#include "bif/bif_4_1_sh_mask.h"

#include "gmc/gmc_7_1_d.h"
#include "gmc/gmc_7_1_sh_mask.h"

#include "oss/oss_2_0_d.h"
#include "oss/oss_2_0_sh_mask.h"

static void gmc_v7_0_set_gart_funcs(struct amdgpu_device *adev);
static void gmc_v7_0_set_irq_funcs(struct amdgpu_device *adev);

43
MODULE_FIRMWARE("radeon/bonaire_mc.bin");
44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437
MODULE_FIRMWARE("radeon/hawaii_mc.bin");

/**
 * gmc8_mc_wait_for_idle - wait for MC idle callback.
 *
 * @adev: amdgpu_device pointer
 *
 * Wait for the MC (memory controller) to be idle.
 * (evergreen+).
 * Returns 0 if the MC is idle, -1 if not.
 */
int gmc_v7_0_mc_wait_for_idle(struct amdgpu_device *adev)
{
	unsigned i;
	u32 tmp;

	for (i = 0; i < adev->usec_timeout; i++) {
		/* read MC_STATUS */
		tmp = RREG32(mmSRBM_STATUS) & 0x1F00;
		if (!tmp)
			return 0;
		udelay(1);
	}
	return -1;
}

void gmc_v7_0_mc_stop(struct amdgpu_device *adev,
		      struct amdgpu_mode_mc_save *save)
{
	u32 blackout;

	if (adev->mode_info.num_crtc)
		amdgpu_display_stop_mc_access(adev, save);

	amdgpu_asic_wait_for_mc_idle(adev);

	blackout = RREG32(mmMC_SHARED_BLACKOUT_CNTL);
	if (REG_GET_FIELD(blackout, MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE) != 1) {
		/* Block CPU access */
		WREG32(mmBIF_FB_EN, 0);
		/* blackout the MC */
		blackout = REG_SET_FIELD(blackout,
					 MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE, 0);
		WREG32(mmMC_SHARED_BLACKOUT_CNTL, blackout | 1);
	}
	/* wait for the MC to settle */
	udelay(100);
}

void gmc_v7_0_mc_resume(struct amdgpu_device *adev,
			struct amdgpu_mode_mc_save *save)
{
	u32 tmp;

	/* unblackout the MC */
	tmp = RREG32(mmMC_SHARED_BLACKOUT_CNTL);
	tmp = REG_SET_FIELD(tmp, MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE, 0);
	WREG32(mmMC_SHARED_BLACKOUT_CNTL, tmp);
	/* allow CPU access */
	tmp = REG_SET_FIELD(0, BIF_FB_EN, FB_READ_EN, 1);
	tmp = REG_SET_FIELD(tmp, BIF_FB_EN, FB_WRITE_EN, 1);
	WREG32(mmBIF_FB_EN, tmp);

	if (adev->mode_info.num_crtc)
		amdgpu_display_resume_mc_access(adev, save);
}

/**
 * gmc_v7_0_init_microcode - load ucode images from disk
 *
 * @adev: amdgpu_device pointer
 *
 * Use the firmware interface to load the ucode images into
 * the driver (not loaded into hw).
 * Returns 0 on success, error on failure.
 */
static int gmc_v7_0_init_microcode(struct amdgpu_device *adev)
{
	const char *chip_name;
	char fw_name[30];
	int err;

	DRM_DEBUG("\n");

	switch (adev->asic_type) {
	case CHIP_BONAIRE:
		chip_name = "bonaire";
		break;
	case CHIP_HAWAII:
		chip_name = "hawaii";
		break;
	case CHIP_KAVERI:
	case CHIP_KABINI:
		return 0;
	default: BUG();
	}

	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
	err = request_firmware(&adev->mc.fw, fw_name, adev->dev);
	if (err)
		goto out;
	err = amdgpu_ucode_validate(adev->mc.fw);

out:
	if (err) {
		printk(KERN_ERR
		       "cik_mc: Failed to load firmware \"%s\"\n",
		       fw_name);
		release_firmware(adev->mc.fw);
		adev->mc.fw = NULL;
	}
	return err;
}

/**
 * gmc_v7_0_mc_load_microcode - load MC ucode into the hw
 *
 * @adev: amdgpu_device pointer
 *
 * Load the GDDR MC ucode into the hw (CIK).
 * Returns 0 on success, error on failure.
 */
static int gmc_v7_0_mc_load_microcode(struct amdgpu_device *adev)
{
	const struct mc_firmware_header_v1_0 *hdr;
	const __le32 *fw_data = NULL;
	const __le32 *io_mc_regs = NULL;
	u32 running, blackout = 0;
	int i, ucode_size, regs_size;

	if (!adev->mc.fw)
		return -EINVAL;

	hdr = (const struct mc_firmware_header_v1_0 *)adev->mc.fw->data;
	amdgpu_ucode_print_mc_hdr(&hdr->header);

	adev->mc.fw_version = le32_to_cpu(hdr->header.ucode_version);
	regs_size = le32_to_cpu(hdr->io_debug_size_bytes) / (4 * 2);
	io_mc_regs = (const __le32 *)
		(adev->mc.fw->data + le32_to_cpu(hdr->io_debug_array_offset_bytes));
	ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
	fw_data = (const __le32 *)
		(adev->mc.fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));

	running = REG_GET_FIELD(RREG32(mmMC_SEQ_SUP_CNTL), MC_SEQ_SUP_CNTL, RUN);

	if (running == 0) {
		if (running) {
			blackout = RREG32(mmMC_SHARED_BLACKOUT_CNTL);
			WREG32(mmMC_SHARED_BLACKOUT_CNTL, blackout | 1);
		}

		/* reset the engine and set to writable */
		WREG32(mmMC_SEQ_SUP_CNTL, 0x00000008);
		WREG32(mmMC_SEQ_SUP_CNTL, 0x00000010);

		/* load mc io regs */
		for (i = 0; i < regs_size; i++) {
			WREG32(mmMC_SEQ_IO_DEBUG_INDEX, le32_to_cpup(io_mc_regs++));
			WREG32(mmMC_SEQ_IO_DEBUG_DATA, le32_to_cpup(io_mc_regs++));
		}
		/* load the MC ucode */
		for (i = 0; i < ucode_size; i++)
			WREG32(mmMC_SEQ_SUP_PGM, le32_to_cpup(fw_data++));

		/* put the engine back into the active state */
		WREG32(mmMC_SEQ_SUP_CNTL, 0x00000008);
		WREG32(mmMC_SEQ_SUP_CNTL, 0x00000004);
		WREG32(mmMC_SEQ_SUP_CNTL, 0x00000001);

		/* wait for training to complete */
		for (i = 0; i < adev->usec_timeout; i++) {
			if (REG_GET_FIELD(RREG32(mmMC_SEQ_TRAIN_WAKEUP_CNTL),
					  MC_SEQ_TRAIN_WAKEUP_CNTL, TRAIN_DONE_D0))
				break;
			udelay(1);
		}
		for (i = 0; i < adev->usec_timeout; i++) {
			if (REG_GET_FIELD(RREG32(mmMC_SEQ_TRAIN_WAKEUP_CNTL),
					  MC_SEQ_TRAIN_WAKEUP_CNTL, TRAIN_DONE_D1))
				break;
			udelay(1);
		}

		if (running)
			WREG32(mmMC_SHARED_BLACKOUT_CNTL, blackout);
	}

	return 0;
}

static void gmc_v7_0_vram_gtt_location(struct amdgpu_device *adev,
				       struct amdgpu_mc *mc)
{
	if (mc->mc_vram_size > 0xFFC0000000ULL) {
		/* leave room for at least 1024M GTT */
		dev_warn(adev->dev, "limiting VRAM\n");
		mc->real_vram_size = 0xFFC0000000ULL;
		mc->mc_vram_size = 0xFFC0000000ULL;
	}
	amdgpu_vram_location(adev, &adev->mc, 0);
	adev->mc.gtt_base_align = 0;
	amdgpu_gtt_location(adev, mc);
}

/**
 * gmc_v7_0_mc_program - program the GPU memory controller
 *
 * @adev: amdgpu_device pointer
 *
 * Set the location of vram, gart, and AGP in the GPU's
 * physical address space (CIK).
 */
static void gmc_v7_0_mc_program(struct amdgpu_device *adev)
{
	struct amdgpu_mode_mc_save save;
	u32 tmp;
	int i, j;

	/* Initialize HDP */
	for (i = 0, j = 0; i < 32; i++, j += 0x6) {
		WREG32((0xb05 + j), 0x00000000);
		WREG32((0xb06 + j), 0x00000000);
		WREG32((0xb07 + j), 0x00000000);
		WREG32((0xb08 + j), 0x00000000);
		WREG32((0xb09 + j), 0x00000000);
	}
	WREG32(mmHDP_REG_COHERENCY_FLUSH_CNTL, 0);

	if (adev->mode_info.num_crtc)
		amdgpu_display_set_vga_render_state(adev, false);

	gmc_v7_0_mc_stop(adev, &save);
	if (amdgpu_asic_wait_for_mc_idle(adev)) {
		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
	}
	/* Update configuration */
	WREG32(mmMC_VM_SYSTEM_APERTURE_LOW_ADDR,
	       adev->mc.vram_start >> 12);
	WREG32(mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
	       adev->mc.vram_end >> 12);
	WREG32(mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
	       adev->vram_scratch.gpu_addr >> 12);
	tmp = ((adev->mc.vram_end >> 24) & 0xFFFF) << 16;
	tmp |= ((adev->mc.vram_start >> 24) & 0xFFFF);
	WREG32(mmMC_VM_FB_LOCATION, tmp);
	/* XXX double check these! */
	WREG32(mmHDP_NONSURFACE_BASE, (adev->mc.vram_start >> 8));
	WREG32(mmHDP_NONSURFACE_INFO, (2 << 7) | (1 << 30));
	WREG32(mmHDP_NONSURFACE_SIZE, 0x3FFFFFFF);
	WREG32(mmMC_VM_AGP_BASE, 0);
	WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF);
	WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF);
	if (amdgpu_asic_wait_for_mc_idle(adev)) {
		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
	}
	gmc_v7_0_mc_resume(adev, &save);

	WREG32(mmBIF_FB_EN, BIF_FB_EN__FB_READ_EN_MASK | BIF_FB_EN__FB_WRITE_EN_MASK);

	tmp = RREG32(mmHDP_MISC_CNTL);
	tmp = REG_SET_FIELD(tmp, HDP_MISC_CNTL, FLUSH_INVALIDATE_CACHE, 1);
	WREG32(mmHDP_MISC_CNTL, tmp);

	tmp = RREG32(mmHDP_HOST_PATH_CNTL);
	WREG32(mmHDP_HOST_PATH_CNTL, tmp);
}

/**
 * gmc_v7_0_mc_init - initialize the memory controller driver params
 *
 * @adev: amdgpu_device pointer
 *
 * Look up the amount of vram, vram width, and decide how to place
 * vram and gart within the GPU's physical address space (CIK).
 * Returns 0 for success.
 */
static int gmc_v7_0_mc_init(struct amdgpu_device *adev)
{
	u32 tmp;
	int chansize, numchan;

	/* Get VRAM informations */
	tmp = RREG32(mmMC_ARB_RAMCFG);
	if (REG_GET_FIELD(tmp, MC_ARB_RAMCFG, CHANSIZE)) {
		chansize = 64;
	} else {
		chansize = 32;
	}
	tmp = RREG32(mmMC_SHARED_CHMAP);
	switch (REG_GET_FIELD(tmp, MC_SHARED_CHMAP, NOOFCHAN)) {
	case 0:
	default:
		numchan = 1;
		break;
	case 1:
		numchan = 2;
		break;
	case 2:
		numchan = 4;
		break;
	case 3:
		numchan = 8;
		break;
	case 4:
		numchan = 3;
		break;
	case 5:
		numchan = 6;
		break;
	case 6:
		numchan = 10;
		break;
	case 7:
		numchan = 12;
		break;
	case 8:
		numchan = 16;
		break;
	}
	adev->mc.vram_width = numchan * chansize;
	/* Could aper size report 0 ? */
	adev->mc.aper_base = pci_resource_start(adev->pdev, 0);
	adev->mc.aper_size = pci_resource_len(adev->pdev, 0);
	/* size in MB on si */
	adev->mc.mc_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL;
	adev->mc.real_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL;
	adev->mc.visible_vram_size = adev->mc.aper_size;

	/* unless the user had overridden it, set the gart
	 * size equal to the 1024 or vram, whichever is larger.
	 */
	if (amdgpu_gart_size == -1)
		adev->mc.gtt_size = max((1024ULL << 20), adev->mc.mc_vram_size);
	else
		adev->mc.gtt_size = (uint64_t)amdgpu_gart_size << 20;

	gmc_v7_0_vram_gtt_location(adev, &adev->mc);

	return 0;
}

/*
 * GART
 * VMID 0 is the physical GPU addresses as used by the kernel.
 * VMIDs 1-15 are used for userspace clients and are handled
 * by the amdgpu vm/hsa code.
 */

/**
 * gmc_v7_0_gart_flush_gpu_tlb - gart tlb flush callback
 *
 * @adev: amdgpu_device pointer
 * @vmid: vm instance to flush
 *
 * Flush the TLB for the requested page table (CIK).
 */
static void gmc_v7_0_gart_flush_gpu_tlb(struct amdgpu_device *adev,
					uint32_t vmid)
{
	/* flush hdp cache */
	WREG32(mmHDP_MEM_COHERENCY_FLUSH_CNTL, 0);

	/* bits 0-15 are the VM contexts0-15 */
	WREG32(mmVM_INVALIDATE_REQUEST, 1 << vmid);
}

/**
 * gmc_v7_0_gart_set_pte_pde - update the page tables using MMIO
 *
 * @adev: amdgpu_device pointer
 * @cpu_pt_addr: cpu address of the page table
 * @gpu_page_idx: entry in the page table to update
 * @addr: dst addr to write into pte/pde
 * @flags: access flags
 *
 * Update the page tables using the CPU.
 */
static int gmc_v7_0_gart_set_pte_pde(struct amdgpu_device *adev,
				     void *cpu_pt_addr,
				     uint32_t gpu_page_idx,
				     uint64_t addr,
				     uint32_t flags)
{
	void __iomem *ptr = (void *)cpu_pt_addr;
	uint64_t value;

	value = addr & 0xFFFFFFFFFFFFF000ULL;
	value |= flags;
	writeq(value, ptr + (gpu_page_idx * 8));

	return 0;
}

438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464
/**
 * gmc_v8_0_set_fault_enable_default - update VM fault handling
 *
 * @adev: amdgpu_device pointer
 * @value: true redirects VM faults to the default page
 */
static void gmc_v7_0_set_fault_enable_default(struct amdgpu_device *adev,
					      bool value)
{
	u32 tmp;

	tmp = RREG32(mmVM_CONTEXT1_CNTL);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	WREG32(mmVM_CONTEXT1_CNTL, tmp);
}

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
/**
 * gmc_v7_0_gart_enable - gart enable
 *
 * @adev: amdgpu_device pointer
 *
 * This sets up the TLBs, programs the page tables for VMID0,
 * sets up the hw for VMIDs 1-15 which are allocated on
 * demand, and sets up the global locations for the LDS, GDS,
 * and GPUVM for FSA64 clients (CIK).
 * Returns 0 for success, errors for failure.
 */
static int gmc_v7_0_gart_enable(struct amdgpu_device *adev)
{
	int r, i;
	u32 tmp;

	if (adev->gart.robj == NULL) {
		dev_err(adev->dev, "No VRAM object for PCIE GART.\n");
		return -EINVAL;
	}
	r = amdgpu_gart_table_vram_pin(adev);
	if (r)
		return r;
	/* Setup TLB control */
	tmp = RREG32(mmMC_VM_MX_L1_TLB_CNTL);
	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_FRAGMENT_PROCESSING, 1);
	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_ADVANCED_DRIVER_MODEL, 1);
	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
	WREG32(mmMC_VM_MX_L1_TLB_CNTL, tmp);
	/* Setup L2 cache */
	tmp = RREG32(mmVM_L2_CNTL);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 1);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 1);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE, 1);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE, 1);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, EFFECTIVE_L2_QUEUE_SIZE, 7);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
504
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
505 506 507 508 509 510 511 512 513 514 515
	WREG32(mmVM_L2_CNTL, tmp);
	tmp = REG_SET_FIELD(0, VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
	WREG32(mmVM_L2_CNTL2, tmp);
	tmp = RREG32(mmVM_L2_CNTL3);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY, 1);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 4);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, L2_CACHE_BIGK_FRAGMENT_SIZE, 4);
	WREG32(mmVM_L2_CNTL3, tmp);
	/* setup context0 */
	WREG32(mmVM_CONTEXT0_PAGE_TABLE_START_ADDR, adev->mc.gtt_start >> 12);
516
	WREG32(mmVM_CONTEXT0_PAGE_TABLE_END_ADDR, (adev->mc.gtt_end >> 12) - 1);
517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
	WREG32(mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR, adev->gart.table_addr >> 12);
	WREG32(mmVM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR,
			(u32)(adev->dummy_page.addr >> 12));
	WREG32(mmVM_CONTEXT0_CNTL2, 0);
	tmp = RREG32(mmVM_CONTEXT0_CNTL);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
	WREG32(mmVM_CONTEXT0_CNTL, tmp);

	WREG32(0x575, 0);
	WREG32(0x576, 0);
	WREG32(0x577, 0);

	/* empty context1-15 */
	/* FIXME start with 4G, once using 2 level pt switch to full
	 * vm size space
	 */
	/* set vm size, must be a multiple of 4 */
	WREG32(mmVM_CONTEXT1_PAGE_TABLE_START_ADDR, 0);
537
	WREG32(mmVM_CONTEXT1_PAGE_TABLE_END_ADDR, adev->vm_manager.max_pfn - 1);
538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556
	for (i = 1; i < 16; i++) {
		if (i < 8)
			WREG32(mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + i,
			       adev->gart.table_addr >> 12);
		else
			WREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + i - 8,
			       adev->gart.table_addr >> 12);
	}

	/* enable context1-15 */
	WREG32(mmVM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR,
	       (u32)(adev->dummy_page.addr >> 12));
	WREG32(mmVM_CONTEXT1_CNTL2, 4);
	tmp = RREG32(mmVM_CONTEXT1_CNTL);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH, 1);
	tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_BLOCK_SIZE,
			    amdgpu_vm_block_size - 9);
	WREG32(mmVM_CONTEXT1_CNTL, tmp);
557 558 559 560
	if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS)
		gmc_v7_0_set_fault_enable_default(adev, false);
	else
		gmc_v7_0_set_fault_enable_default(adev, true);
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

	if (adev->asic_type == CHIP_KAVERI) {
		tmp = RREG32(mmCHUB_CONTROL);
		tmp &= ~BYPASS_VM;
		WREG32(mmCHUB_CONTROL, tmp);
	}

	gmc_v7_0_gart_flush_gpu_tlb(adev, 0);
	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
		 (unsigned)(adev->mc.gtt_size >> 20),
		 (unsigned long long)adev->gart.table_addr);
	adev->gart.ready = true;
	return 0;
}

static int gmc_v7_0_gart_init(struct amdgpu_device *adev)
{
	int r;

	if (adev->gart.robj) {
		WARN(1, "R600 PCIE GART already initialized\n");
		return 0;
	}
	/* Initialize common gart structure */
	r = amdgpu_gart_init(adev);
	if (r)
		return r;
	adev->gart.table_size = adev->gart.num_gpu_pages * 8;
	return amdgpu_gart_table_vram_alloc(adev);
}

/**
 * gmc_v7_0_gart_disable - gart disable
 *
 * @adev: amdgpu_device pointer
 *
 * This disables all VM page table (CIK).
 */
static void gmc_v7_0_gart_disable(struct amdgpu_device *adev)
{
	u32 tmp;

	/* Disable all tables */
	WREG32(mmVM_CONTEXT0_CNTL, 0);
	WREG32(mmVM_CONTEXT1_CNTL, 0);
	/* Setup TLB control */
	tmp = RREG32(mmMC_VM_MX_L1_TLB_CNTL);
	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_FRAGMENT_PROCESSING, 0);
	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_ADVANCED_DRIVER_MODEL, 0);
	WREG32(mmMC_VM_MX_L1_TLB_CNTL, tmp);
	/* Setup L2 cache */
	tmp = RREG32(mmVM_L2_CNTL);
	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 0);
	WREG32(mmVM_L2_CNTL, tmp);
	WREG32(mmVM_L2_CNTL2, 0);
	amdgpu_gart_table_vram_unpin(adev);
}

/**
 * gmc_v7_0_gart_fini - vm fini callback
 *
 * @adev: amdgpu_device pointer
 *
 * Tears down the driver GART/VM setup (CIK).
 */
static void gmc_v7_0_gart_fini(struct amdgpu_device *adev)
{
	amdgpu_gart_table_vram_free(adev);
	amdgpu_gart_fini(adev);
}

/*
 * vm
 * VMID 0 is the physical GPU addresses as used by the kernel.
 * VMIDs 1-15 are used for userspace clients and are handled
 * by the amdgpu vm/hsa code.
 */
/**
 * gmc_v7_0_vm_init - cik vm init callback
 *
 * @adev: amdgpu_device pointer
 *
 * Inits cik specific vm parameters (number of VMs, base of vram for
 * VMIDs 1-15) (CIK).
 * Returns 0 for success.
 */
static int gmc_v7_0_vm_init(struct amdgpu_device *adev)
{
	/*
	 * number of VMs
	 * VMID 0 is reserved for System
	 * amdgpu graphics/compute will use VMIDs 1-7
	 * amdkfd will use VMIDs 8-15
	 */
	adev->vm_manager.nvm = AMDGPU_NUM_OF_VMIDS;

	/* base offset of vram pages */
659
	if (adev->flags & AMD_IS_APU) {
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
		u64 tmp = RREG32(mmMC_VM_FB_OFFSET);
		tmp <<= 22;
		adev->vm_manager.vram_base_offset = tmp;
	} else
		adev->vm_manager.vram_base_offset = 0;

	return 0;
}

/**
 * gmc_v7_0_vm_fini - cik vm fini callback
 *
 * @adev: amdgpu_device pointer
 *
 * Tear down any asic specific VM setup (CIK).
 */
static void gmc_v7_0_vm_fini(struct amdgpu_device *adev)
{
}

/**
 * gmc_v7_0_vm_decode_fault - print human readable fault info
 *
 * @adev: amdgpu_device pointer
 * @status: VM_CONTEXT1_PROTECTION_FAULT_STATUS register value
 * @addr: VM_CONTEXT1_PROTECTION_FAULT_ADDR register value
 *
 * Print human readable fault information (CIK).
 */
static void gmc_v7_0_vm_decode_fault(struct amdgpu_device *adev,
				     u32 status, u32 addr, u32 mc_client)
{
	u32 mc_id;
	u32 vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID);
	u32 protections = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
					PROTECTIONS);
	char block[5] = { mc_client >> 24, (mc_client >> 16) & 0xff,
		(mc_client >> 8) & 0xff, mc_client & 0xff, 0 };

	mc_id = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
			      MEMORY_CLIENT_ID);

	printk("VM fault (0x%02x, vmid %d) at page %u, %s from '%s' (0x%08x) (%d)\n",
	       protections, vmid, addr,
	       REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
			     MEMORY_CLIENT_RW) ?
	       "write" : "read", block, mc_client, mc_id);
}


static const u32 mc_cg_registers[] = {
	mmMC_HUB_MISC_HUB_CG,
	mmMC_HUB_MISC_SIP_CG,
	mmMC_HUB_MISC_VM_CG,
	mmMC_XPB_CLK_GAT,
	mmATC_MISC_CG,
	mmMC_CITF_MISC_WR_CG,
	mmMC_CITF_MISC_RD_CG,
	mmMC_CITF_MISC_VM_CG,
	mmVM_L2_CG,
};

static const u32 mc_cg_ls_en[] = {
	MC_HUB_MISC_HUB_CG__MEM_LS_ENABLE_MASK,
	MC_HUB_MISC_SIP_CG__MEM_LS_ENABLE_MASK,
	MC_HUB_MISC_VM_CG__MEM_LS_ENABLE_MASK,
	MC_XPB_CLK_GAT__MEM_LS_ENABLE_MASK,
	ATC_MISC_CG__MEM_LS_ENABLE_MASK,
	MC_CITF_MISC_WR_CG__MEM_LS_ENABLE_MASK,
	MC_CITF_MISC_RD_CG__MEM_LS_ENABLE_MASK,
	MC_CITF_MISC_VM_CG__MEM_LS_ENABLE_MASK,
	VM_L2_CG__MEM_LS_ENABLE_MASK,
};

static const u32 mc_cg_en[] = {
	MC_HUB_MISC_HUB_CG__ENABLE_MASK,
	MC_HUB_MISC_SIP_CG__ENABLE_MASK,
	MC_HUB_MISC_VM_CG__ENABLE_MASK,
	MC_XPB_CLK_GAT__ENABLE_MASK,
	ATC_MISC_CG__ENABLE_MASK,
	MC_CITF_MISC_WR_CG__ENABLE_MASK,
	MC_CITF_MISC_RD_CG__ENABLE_MASK,
	MC_CITF_MISC_VM_CG__ENABLE_MASK,
	VM_L2_CG__ENABLE_MASK,
};

static void gmc_v7_0_enable_mc_ls(struct amdgpu_device *adev,
				  bool enable)
{
	int i;
	u32 orig, data;

	for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) {
		orig = data = RREG32(mc_cg_registers[i]);
		if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_MC_LS))
			data |= mc_cg_ls_en[i];
		else
			data &= ~mc_cg_ls_en[i];
		if (data != orig)
			WREG32(mc_cg_registers[i], data);
	}
}

static void gmc_v7_0_enable_mc_mgcg(struct amdgpu_device *adev,
				    bool enable)
{
	int i;
	u32 orig, data;

	for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) {
		orig = data = RREG32(mc_cg_registers[i]);
		if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_MC_MGCG))
			data |= mc_cg_en[i];
		else
			data &= ~mc_cg_en[i];
		if (data != orig)
			WREG32(mc_cg_registers[i], data);
	}
}

static void gmc_v7_0_enable_bif_mgls(struct amdgpu_device *adev,
				     bool enable)
{
	u32 orig, data;

	orig = data = RREG32_PCIE(ixPCIE_CNTL2);

	if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_BIF_LS)) {
		data = REG_SET_FIELD(data, PCIE_CNTL2, SLV_MEM_LS_EN, 1);
		data = REG_SET_FIELD(data, PCIE_CNTL2, MST_MEM_LS_EN, 1);
		data = REG_SET_FIELD(data, PCIE_CNTL2, REPLAY_MEM_LS_EN, 1);
		data = REG_SET_FIELD(data, PCIE_CNTL2, SLV_MEM_AGGRESSIVE_LS_EN, 1);
	} else {
		data = REG_SET_FIELD(data, PCIE_CNTL2, SLV_MEM_LS_EN, 0);
		data = REG_SET_FIELD(data, PCIE_CNTL2, MST_MEM_LS_EN, 0);
		data = REG_SET_FIELD(data, PCIE_CNTL2, REPLAY_MEM_LS_EN, 0);
		data = REG_SET_FIELD(data, PCIE_CNTL2, SLV_MEM_AGGRESSIVE_LS_EN, 0);
	}

	if (orig != data)
		WREG32_PCIE(ixPCIE_CNTL2, data);
}

static void gmc_v7_0_enable_hdp_mgcg(struct amdgpu_device *adev,
				     bool enable)
{
	u32 orig, data;

	orig = data = RREG32(mmHDP_HOST_PATH_CNTL);

	if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_HDP_MGCG))
		data = REG_SET_FIELD(data, HDP_HOST_PATH_CNTL, CLOCK_GATING_DIS, 0);
	else
		data = REG_SET_FIELD(data, HDP_HOST_PATH_CNTL, CLOCK_GATING_DIS, 1);

	if (orig != data)
		WREG32(mmHDP_HOST_PATH_CNTL, data);
}

static void gmc_v7_0_enable_hdp_ls(struct amdgpu_device *adev,
				   bool enable)
{
	u32 orig, data;

	orig = data = RREG32(mmHDP_MEM_POWER_LS);

	if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_HDP_LS))
		data = REG_SET_FIELD(data, HDP_MEM_POWER_LS, LS_ENABLE, 1);
	else
		data = REG_SET_FIELD(data, HDP_MEM_POWER_LS, LS_ENABLE, 0);

	if (orig != data)
		WREG32(mmHDP_MEM_POWER_LS, data);
}

835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856
static int gmc_v7_0_convert_vram_type(int mc_seq_vram_type)
{
	switch (mc_seq_vram_type) {
	case MC_SEQ_MISC0__MT__GDDR1:
		return AMDGPU_VRAM_TYPE_GDDR1;
	case MC_SEQ_MISC0__MT__DDR2:
		return AMDGPU_VRAM_TYPE_DDR2;
	case MC_SEQ_MISC0__MT__GDDR3:
		return AMDGPU_VRAM_TYPE_GDDR3;
	case MC_SEQ_MISC0__MT__GDDR4:
		return AMDGPU_VRAM_TYPE_GDDR4;
	case MC_SEQ_MISC0__MT__GDDR5:
		return AMDGPU_VRAM_TYPE_GDDR5;
	case MC_SEQ_MISC0__MT__HBM:
		return AMDGPU_VRAM_TYPE_HBM;
	case MC_SEQ_MISC0__MT__DDR3:
		return AMDGPU_VRAM_TYPE_DDR3;
	default:
		return AMDGPU_VRAM_TYPE_UNKNOWN;
	}
}

857
static int gmc_v7_0_early_init(void *handle)
858
{
859 860
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;

861 862 863
	gmc_v7_0_set_gart_funcs(adev);
	gmc_v7_0_set_irq_funcs(adev);

864
	if (adev->flags & AMD_IS_APU) {
865
		adev->mc.vram_type = AMDGPU_VRAM_TYPE_UNKNOWN;
866 867
	} else {
		u32 tmp = RREG32(mmMC_SEQ_MISC0);
868 869
		tmp &= MC_SEQ_MISC0__MT__MASK;
		adev->mc.vram_type = gmc_v7_0_convert_vram_type(tmp);
870 871 872 873 874
	}

	return 0;
}

875 876 877 878 879 880 881
static int gmc_v7_0_late_init(void *handle)
{
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;

	return amdgpu_irq_get(adev, &adev->mc.vm_fault, 0);
}

882
static int gmc_v7_0_sw_init(void *handle)
883 884 885
{
	int r;
	int dma_bits;
886
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961

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

	r = amdgpu_irq_add_id(adev, 146, &adev->mc.vm_fault);
	if (r)
		return r;

	r = amdgpu_irq_add_id(adev, 147, &adev->mc.vm_fault);
	if (r)
		return r;

	/* Adjust VM size here.
	 * Currently set to 4GB ((1 << 20) 4k pages).
	 * Max GPUVM size for cayman and SI is 40 bits.
	 */
	adev->vm_manager.max_pfn = amdgpu_vm_size << 18;

	/* Set the internal MC address mask
	 * This is the max address of the GPU's
	 * internal address space.
	 */
	adev->mc.mc_mask = 0xffffffffffULL; /* 40 bit MC */

	/* set DMA mask + need_dma32 flags.
	 * PCIE - can handle 40-bits.
	 * IGP - can handle 40-bits
	 * PCI - dma32 for legacy pci gart, 40 bits on newer asics
	 */
	adev->need_dma32 = false;
	dma_bits = adev->need_dma32 ? 32 : 40;
	r = pci_set_dma_mask(adev->pdev, DMA_BIT_MASK(dma_bits));
	if (r) {
		adev->need_dma32 = true;
		dma_bits = 32;
		printk(KERN_WARNING "amdgpu: No suitable DMA available.\n");
	}
	r = pci_set_consistent_dma_mask(adev->pdev, DMA_BIT_MASK(dma_bits));
	if (r) {
		pci_set_consistent_dma_mask(adev->pdev, DMA_BIT_MASK(32));
		printk(KERN_WARNING "amdgpu: No coherent DMA available.\n");
	}

	r = gmc_v7_0_init_microcode(adev);
	if (r) {
		DRM_ERROR("Failed to load mc firmware!\n");
		return r;
	}

	r = gmc_v7_0_mc_init(adev);
	if (r)
		return r;

	/* Memory manager */
	r = amdgpu_bo_init(adev);
	if (r)
		return r;

	r = gmc_v7_0_gart_init(adev);
	if (r)
		return r;

	if (!adev->vm_manager.enabled) {
		r = gmc_v7_0_vm_init(adev);
		if (r) {
			dev_err(adev->dev, "vm manager initialization failed (%d).\n", r);
			return r;
		}
		adev->vm_manager.enabled = true;
	}

	return r;
}

962
static int gmc_v7_0_sw_fini(void *handle)
963 964
{
	int i;
965
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
966 967 968

	if (adev->vm_manager.enabled) {
		for (i = 0; i < AMDGPU_NUM_VM; ++i)
969
			fence_put(adev->vm_manager.active[i]);
970 971 972 973 974 975 976 977 978 979
		gmc_v7_0_vm_fini(adev);
		adev->vm_manager.enabled = false;
	}
	gmc_v7_0_gart_fini(adev);
	amdgpu_gem_fini(adev);
	amdgpu_bo_fini(adev);

	return 0;
}

980
static int gmc_v7_0_hw_init(void *handle)
981 982
{
	int r;
983
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
984 985 986

	gmc_v7_0_mc_program(adev);

987
	if (!(adev->flags & AMD_IS_APU)) {
988 989 990 991 992 993 994 995 996 997 998 999 1000 1001
		r = gmc_v7_0_mc_load_microcode(adev);
		if (r) {
			DRM_ERROR("Failed to load MC firmware!\n");
			return r;
		}
	}

	r = gmc_v7_0_gart_enable(adev);
	if (r)
		return r;

	return r;
}

1002
static int gmc_v7_0_hw_fini(void *handle)
1003
{
1004 1005
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;

1006
	amdgpu_irq_put(adev, &adev->mc.vm_fault, 0);
1007 1008 1009 1010 1011
	gmc_v7_0_gart_disable(adev);

	return 0;
}

1012
static int gmc_v7_0_suspend(void *handle)
1013 1014
{
	int i;
1015
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
1016 1017 1018

	if (adev->vm_manager.enabled) {
		for (i = 0; i < AMDGPU_NUM_VM; ++i)
1019
			fence_put(adev->vm_manager.active[i]);
1020 1021 1022 1023 1024 1025 1026 1027
		gmc_v7_0_vm_fini(adev);
		adev->vm_manager.enabled = false;
	}
	gmc_v7_0_hw_fini(adev);

	return 0;
}

1028
static int gmc_v7_0_resume(void *handle)
1029 1030
{
	int r;
1031
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048

	r = gmc_v7_0_hw_init(adev);
	if (r)
		return r;

	if (!adev->vm_manager.enabled) {
		r = gmc_v7_0_vm_init(adev);
		if (r) {
			dev_err(adev->dev, "vm manager initialization failed (%d).\n", r);
			return r;
		}
		adev->vm_manager.enabled = true;
	}

	return r;
}

1049
static bool gmc_v7_0_is_idle(void *handle)
1050
{
1051
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
1052 1053 1054 1055 1056 1057 1058 1059 1060
	u32 tmp = RREG32(mmSRBM_STATUS);

	if (tmp & (SRBM_STATUS__MCB_BUSY_MASK | SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK |
		   SRBM_STATUS__MCC_BUSY_MASK | SRBM_STATUS__MCD_BUSY_MASK | SRBM_STATUS__VMC_BUSY_MASK))
		return false;

	return true;
}

1061
static int gmc_v7_0_wait_for_idle(void *handle)
1062 1063 1064
{
	unsigned i;
	u32 tmp;
1065
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081

	for (i = 0; i < adev->usec_timeout; i++) {
		/* read MC_STATUS */
		tmp = RREG32(mmSRBM_STATUS) & (SRBM_STATUS__MCB_BUSY_MASK |
					       SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK |
					       SRBM_STATUS__MCC_BUSY_MASK |
					       SRBM_STATUS__MCD_BUSY_MASK |
					       SRBM_STATUS__VMC_BUSY_MASK);
		if (!tmp)
			return 0;
		udelay(1);
	}
	return -ETIMEDOUT;

}

1082
static void gmc_v7_0_print_status(void *handle)
1083 1084
{
	int i, j;
1085
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189

	dev_info(adev->dev, "GMC 8.x registers\n");
	dev_info(adev->dev, "  SRBM_STATUS=0x%08X\n",
		RREG32(mmSRBM_STATUS));
	dev_info(adev->dev, "  SRBM_STATUS2=0x%08X\n",
		RREG32(mmSRBM_STATUS2));

	dev_info(adev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
		 RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_ADDR));
	dev_info(adev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
		 RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS));
	dev_info(adev->dev, "  MC_VM_MX_L1_TLB_CNTL=0x%08X\n",
		 RREG32(mmMC_VM_MX_L1_TLB_CNTL));
	dev_info(adev->dev, "  VM_L2_CNTL=0x%08X\n",
		 RREG32(mmVM_L2_CNTL));
	dev_info(adev->dev, "  VM_L2_CNTL2=0x%08X\n",
		 RREG32(mmVM_L2_CNTL2));
	dev_info(adev->dev, "  VM_L2_CNTL3=0x%08X\n",
		 RREG32(mmVM_L2_CNTL3));
	dev_info(adev->dev, "  VM_CONTEXT0_PAGE_TABLE_START_ADDR=0x%08X\n",
		 RREG32(mmVM_CONTEXT0_PAGE_TABLE_START_ADDR));
	dev_info(adev->dev, "  VM_CONTEXT0_PAGE_TABLE_END_ADDR=0x%08X\n",
		 RREG32(mmVM_CONTEXT0_PAGE_TABLE_END_ADDR));
	dev_info(adev->dev, "  VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR=0x%08X\n",
		 RREG32(mmVM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR));
	dev_info(adev->dev, "  VM_CONTEXT0_CNTL2=0x%08X\n",
		 RREG32(mmVM_CONTEXT0_CNTL2));
	dev_info(adev->dev, "  VM_CONTEXT0_CNTL=0x%08X\n",
		 RREG32(mmVM_CONTEXT0_CNTL));
	dev_info(adev->dev, "  0x15D4=0x%08X\n",
		 RREG32(0x575));
	dev_info(adev->dev, "  0x15D8=0x%08X\n",
		 RREG32(0x576));
	dev_info(adev->dev, "  0x15DC=0x%08X\n",
		 RREG32(0x577));
	dev_info(adev->dev, "  VM_CONTEXT1_PAGE_TABLE_START_ADDR=0x%08X\n",
		 RREG32(mmVM_CONTEXT1_PAGE_TABLE_START_ADDR));
	dev_info(adev->dev, "  VM_CONTEXT1_PAGE_TABLE_END_ADDR=0x%08X\n",
		 RREG32(mmVM_CONTEXT1_PAGE_TABLE_END_ADDR));
	dev_info(adev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR=0x%08X\n",
		 RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR));
	dev_info(adev->dev, "  VM_CONTEXT1_CNTL2=0x%08X\n",
		 RREG32(mmVM_CONTEXT1_CNTL2));
	dev_info(adev->dev, "  VM_CONTEXT1_CNTL=0x%08X\n",
		 RREG32(mmVM_CONTEXT1_CNTL));
	for (i = 0; i < 16; i++) {
		if (i < 8)
			dev_info(adev->dev, "  VM_CONTEXT%d_PAGE_TABLE_BASE_ADDR=0x%08X\n",
				 i, RREG32(mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + i));
		else
			dev_info(adev->dev, "  VM_CONTEXT%d_PAGE_TABLE_BASE_ADDR=0x%08X\n",
				 i, RREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + i - 8));
	}
	dev_info(adev->dev, "  MC_VM_SYSTEM_APERTURE_LOW_ADDR=0x%08X\n",
		 RREG32(mmMC_VM_SYSTEM_APERTURE_LOW_ADDR));
	dev_info(adev->dev, "  MC_VM_SYSTEM_APERTURE_HIGH_ADDR=0x%08X\n",
		 RREG32(mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR));
	dev_info(adev->dev, "  MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR=0x%08X\n",
		 RREG32(mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR));
	dev_info(adev->dev, "  MC_VM_FB_LOCATION=0x%08X\n",
		 RREG32(mmMC_VM_FB_LOCATION));
	dev_info(adev->dev, "  MC_VM_AGP_BASE=0x%08X\n",
		 RREG32(mmMC_VM_AGP_BASE));
	dev_info(adev->dev, "  MC_VM_AGP_TOP=0x%08X\n",
		 RREG32(mmMC_VM_AGP_TOP));
	dev_info(adev->dev, "  MC_VM_AGP_BOT=0x%08X\n",
		 RREG32(mmMC_VM_AGP_BOT));

	if (adev->asic_type == CHIP_KAVERI) {
		dev_info(adev->dev, "  CHUB_CONTROL=0x%08X\n",
			 RREG32(mmCHUB_CONTROL));
	}

	dev_info(adev->dev, "  HDP_REG_COHERENCY_FLUSH_CNTL=0x%08X\n",
		 RREG32(mmHDP_REG_COHERENCY_FLUSH_CNTL));
	dev_info(adev->dev, "  HDP_NONSURFACE_BASE=0x%08X\n",
		 RREG32(mmHDP_NONSURFACE_BASE));
	dev_info(adev->dev, "  HDP_NONSURFACE_INFO=0x%08X\n",
		 RREG32(mmHDP_NONSURFACE_INFO));
	dev_info(adev->dev, "  HDP_NONSURFACE_SIZE=0x%08X\n",
		 RREG32(mmHDP_NONSURFACE_SIZE));
	dev_info(adev->dev, "  HDP_MISC_CNTL=0x%08X\n",
		 RREG32(mmHDP_MISC_CNTL));
	dev_info(adev->dev, "  HDP_HOST_PATH_CNTL=0x%08X\n",
		 RREG32(mmHDP_HOST_PATH_CNTL));

	for (i = 0, j = 0; i < 32; i++, j += 0x6) {
		dev_info(adev->dev, "  %d:\n", i);
		dev_info(adev->dev, "  0x%04X=0x%08X\n",
			 0xb05 + j, RREG32(0xb05 + j));
		dev_info(adev->dev, "  0x%04X=0x%08X\n",
			 0xb06 + j, RREG32(0xb06 + j));
		dev_info(adev->dev, "  0x%04X=0x%08X\n",
			 0xb07 + j, RREG32(0xb07 + j));
		dev_info(adev->dev, "  0x%04X=0x%08X\n",
			 0xb08 + j, RREG32(0xb08 + j));
		dev_info(adev->dev, "  0x%04X=0x%08X\n",
			 0xb09 + j, RREG32(0xb09 + j));
	}

	dev_info(adev->dev, "  BIF_FB_EN=0x%08X\n",
		 RREG32(mmBIF_FB_EN));
}

1190
static int gmc_v7_0_soft_reset(void *handle)
1191
{
1192
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
	struct amdgpu_mode_mc_save save;
	u32 srbm_soft_reset = 0;
	u32 tmp = RREG32(mmSRBM_STATUS);

	if (tmp & SRBM_STATUS__VMC_BUSY_MASK)
		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset,
						SRBM_SOFT_RESET, SOFT_RESET_VMC, 1);

	if (tmp & (SRBM_STATUS__MCB_BUSY_MASK | SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK |
		   SRBM_STATUS__MCC_BUSY_MASK | SRBM_STATUS__MCD_BUSY_MASK)) {
1203
		if (!(adev->flags & AMD_IS_APU))
1204 1205 1206 1207 1208
			srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset,
							SRBM_SOFT_RESET, SOFT_RESET_MC, 1);
	}

	if (srbm_soft_reset) {
1209
		gmc_v7_0_print_status((void *)adev);
1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234

		gmc_v7_0_mc_stop(adev, &save);
		if (gmc_v7_0_wait_for_idle(adev)) {
			dev_warn(adev->dev, "Wait for GMC idle timed out !\n");
		}


		tmp = RREG32(mmSRBM_SOFT_RESET);
		tmp |= srbm_soft_reset;
		dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
		WREG32(mmSRBM_SOFT_RESET, tmp);
		tmp = RREG32(mmSRBM_SOFT_RESET);

		udelay(50);

		tmp &= ~srbm_soft_reset;
		WREG32(mmSRBM_SOFT_RESET, tmp);
		tmp = RREG32(mmSRBM_SOFT_RESET);

		/* Wait a little for things to settle down */
		udelay(50);

		gmc_v7_0_mc_resume(adev, &save);
		udelay(50);

1235
		gmc_v7_0_print_status((void *)adev);
1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
	}

	return 0;
}

static int gmc_v7_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
					     struct amdgpu_irq_src *src,
					     unsigned type,
					     enum amdgpu_interrupt_state state)
{
	u32 tmp;
	u32 bits = (VM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		    VM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		    VM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		    VM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		    VM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		    VM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK);

	switch (state) {
	case AMDGPU_IRQ_STATE_DISABLE:
		/* system context */
		tmp = RREG32(mmVM_CONTEXT0_CNTL);
		tmp &= ~bits;
		WREG32(mmVM_CONTEXT0_CNTL, tmp);
		/* VMs */
		tmp = RREG32(mmVM_CONTEXT1_CNTL);
		tmp &= ~bits;
		WREG32(mmVM_CONTEXT1_CNTL, tmp);
		break;
	case AMDGPU_IRQ_STATE_ENABLE:
		/* system context */
		tmp = RREG32(mmVM_CONTEXT0_CNTL);
		tmp |= bits;
		WREG32(mmVM_CONTEXT0_CNTL, tmp);
		/* VMs */
		tmp = RREG32(mmVM_CONTEXT1_CNTL);
		tmp |= bits;
		WREG32(mmVM_CONTEXT1_CNTL, tmp);
		break;
	default:
		break;
	}

	return 0;
}

static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev,
				      struct amdgpu_irq_src *source,
				      struct amdgpu_iv_entry *entry)
{
	u32 addr, status, mc_client;

	addr = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_ADDR);
	status = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS);
	mc_client = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_MCCLIENT);
1291 1292 1293 1294 1295 1296
	/* reset addr and status */
	WREG32_P(mmVM_CONTEXT1_CNTL2, 1, ~1);

	if (!addr && !status)
		return 0;

1297 1298 1299
	if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_FIRST)
		gmc_v7_0_set_fault_enable_default(adev, false);

1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310
	dev_err(adev->dev, "GPU fault detected: %d 0x%08x\n",
		entry->src_id, entry->src_data);
	dev_err(adev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
		addr);
	dev_err(adev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
		status);
	gmc_v7_0_vm_decode_fault(adev, status, addr, mc_client);

	return 0;
}

1311 1312
static int gmc_v7_0_set_clockgating_state(void *handle,
					  enum amd_clockgating_state state)
1313 1314
{
	bool gate = false;
1315
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
1316

1317
	if (state == AMD_CG_STATE_GATE)
1318 1319
		gate = true;

1320
	if (!(adev->flags & AMD_IS_APU)) {
1321 1322 1323 1324 1325 1326 1327 1328 1329 1330
		gmc_v7_0_enable_mc_mgcg(adev, gate);
		gmc_v7_0_enable_mc_ls(adev, gate);
	}
	gmc_v7_0_enable_bif_mgls(adev, gate);
	gmc_v7_0_enable_hdp_mgcg(adev, gate);
	gmc_v7_0_enable_hdp_ls(adev, gate);

	return 0;
}

1331 1332
static int gmc_v7_0_set_powergating_state(void *handle,
					  enum amd_powergating_state state)
1333 1334 1335 1336
{
	return 0;
}

1337
const struct amd_ip_funcs gmc_v7_0_ip_funcs = {
1338
	.early_init = gmc_v7_0_early_init,
1339
	.late_init = gmc_v7_0_late_init,
1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
	.sw_init = gmc_v7_0_sw_init,
	.sw_fini = gmc_v7_0_sw_fini,
	.hw_init = gmc_v7_0_hw_init,
	.hw_fini = gmc_v7_0_hw_fini,
	.suspend = gmc_v7_0_suspend,
	.resume = gmc_v7_0_resume,
	.is_idle = gmc_v7_0_is_idle,
	.wait_for_idle = gmc_v7_0_wait_for_idle,
	.soft_reset = gmc_v7_0_soft_reset,
	.print_status = gmc_v7_0_print_status,
	.set_clockgating_state = gmc_v7_0_set_clockgating_state,
	.set_powergating_state = gmc_v7_0_set_powergating_state,
};

static const struct amdgpu_gart_funcs gmc_v7_0_gart_funcs = {
	.flush_gpu_tlb = gmc_v7_0_gart_flush_gpu_tlb,
	.set_pte_pde = gmc_v7_0_gart_set_pte_pde,
};

static const struct amdgpu_irq_src_funcs gmc_v7_0_irq_funcs = {
	.set = gmc_v7_0_vm_fault_interrupt_state,
	.process = gmc_v7_0_process_interrupt,
};

static void gmc_v7_0_set_gart_funcs(struct amdgpu_device *adev)
{
	if (adev->gart.gart_funcs == NULL)
		adev->gart.gart_funcs = &gmc_v7_0_gart_funcs;
}

static void gmc_v7_0_set_irq_funcs(struct amdgpu_device *adev)
{
	adev->mc.vm_fault.num_types = 1;
	adev->mc.vm_fault.funcs = &gmc_v7_0_irq_funcs;
}