mmhub_v2_0.c 21.6 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
/*
 * Copyright 2019 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 "amdgpu.h"
#include "mmhub_v2_0.h"

#include "mmhub/mmhub_2_0_0_offset.h"
#include "mmhub/mmhub_2_0_0_sh_mask.h"
#include "mmhub/mmhub_2_0_0_default.h"
#include "navi10_enum.h"

#include "soc15_common.h"

34 35 36 37 38
#define mmMM_ATC_L2_MISC_CG_Sienna_Cichlid                      0x064d
#define mmMM_ATC_L2_MISC_CG_Sienna_Cichlid_BASE_IDX             0
#define mmDAGB0_CNTL_MISC2_Sienna_Cichlid                       0x0070
#define mmDAGB0_CNTL_MISC2_Sienna_Cichlid_BASE_IDX              0

39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
static uint32_t mmhub_v2_0_get_invalidate_req(unsigned int vmid,
					      uint32_t flush_type)
{
	u32 req = 0;

	/* invalidate using legacy mode on vmid*/
	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type);
	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);

	return req;
}

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
static void
mmhub_v2_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
					     uint32_t status)
{
	dev_err(adev->dev,
		"MMVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
		status);
	dev_err(adev->dev, "\t Faulty UTCL2 client ID: 0x%lx\n",
		REG_GET_FIELD(status,
		MMVM_L2_PROTECTION_FAULT_STATUS, CID));
	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
		REG_GET_FIELD(status,
		MMVM_L2_PROTECTION_FAULT_STATUS, MORE_FAULTS));
	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
		REG_GET_FIELD(status,
		MMVM_L2_PROTECTION_FAULT_STATUS, WALKER_ERROR));
	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
		REG_GET_FIELD(status,
		MMVM_L2_PROTECTION_FAULT_STATUS, PERMISSION_FAULTS));
	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
		REG_GET_FIELD(status,
		MMVM_L2_PROTECTION_FAULT_STATUS, MAPPING_ERROR));
	dev_err(adev->dev, "\t RW: 0x%lx\n",
		REG_GET_FIELD(status,
		MMVM_L2_PROTECTION_FAULT_STATUS, RW));
}

86
static void mmhub_v2_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
87
				uint64_t page_table_base)
88
{
89
	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
90

91
	WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
92 93
			    hub->ctx_addr_distance * vmid,
			    lower_32_bits(page_table_base));
94

95
	WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
96 97
			    hub->ctx_addr_distance * vmid,
			    upper_32_bits(page_table_base));
98 99 100 101
}

static void mmhub_v2_0_init_gart_aperture_regs(struct amdgpu_device *adev)
{
102 103 104
	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);

	mmhub_v2_0_setup_vm_pt_regs(adev, 0, pt_base);
105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126

	WREG32_SOC15(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
		     (u32)(adev->gmc.gart_start >> 12));
	WREG32_SOC15(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
		     (u32)(adev->gmc.gart_start >> 44));

	WREG32_SOC15(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
		     (u32)(adev->gmc.gart_end >> 12));
	WREG32_SOC15(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
		     (u32)(adev->gmc.gart_end >> 44));
}

static void mmhub_v2_0_init_system_aperture_regs(struct amdgpu_device *adev)
{
	uint64_t value;
	uint32_t tmp;

	/* Disable AGP. */
	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_AGP_BASE, 0);
	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_AGP_TOP, 0);
	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_AGP_BOT, 0x00FFFFFF);

127 128 129 130 131 132 133 134 135 136 137 138
	if (!amdgpu_sriov_vf(adev)) {
		/*
		 * the new L1 policy will block SRIOV guest from writing
		 * these regs, and they will be programed at host.
		 * so skip programing these regs.
		 */
		/* Program the system aperture low logical page number. */
		WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_LOW_ADDR,
			     adev->gmc.vram_start >> 18);
		WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
			     adev->gmc.vram_end >> 18);
	}
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

	/* Set default page address. */
	value = adev->vram_scratch.gpu_addr - adev->gmc.vram_start +
		adev->vm_manager.vram_base_offset;
	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
		     (u32)(value >> 12));
	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
		     (u32)(value >> 44));

	/* Program "protection fault". */
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
		     (u32)(adev->dummy_page_addr >> 12));
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
		     (u32)((u64)adev->dummy_page_addr >> 44));

	tmp = RREG32_SOC15(MMHUB, 0, mmMMVM_L2_PROTECTION_FAULT_CNTL2);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL2,
			    ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_PROTECTION_FAULT_CNTL2, tmp);
}

static void mmhub_v2_0_init_tlb_regs(struct amdgpu_device *adev)
{
	uint32_t tmp;

	/* Setup TLB control */
	tmp = RREG32_SOC15(MMHUB, 0, mmMMMC_VM_MX_L1_TLB_CNTL);

	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
			    MTYPE, MTYPE_UC); /* UC, uncached */

	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_MX_L1_TLB_CNTL, tmp);
}

static void mmhub_v2_0_init_cache_regs(struct amdgpu_device *adev)
{
	uint32_t tmp;

184 185 186 187 188 189
	/* These registers are not accessible to VF-SRIOV.
	 * The PF will program them instead.
	 */
	if (amdgpu_sriov_vf(adev))
		return;

190 191 192 193
	/* Setup L2 cache */
	tmp = RREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 1);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
194 195
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL,
			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
196 197 198
	/* XXX for emulation, Refer to closed source code.*/
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE,
			    0);
199
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
200 201 202 203 204 205 206 207 208 209
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL, tmp);

	tmp = RREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL2);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL2, tmp);

	tmp = mmMMVM_L2_CNTL3_DEFAULT;
210 211 212 213 214 215 216 217 218
	if (adev->gmc.translate_further) {
		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 12);
		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
	} else {
		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 9);
		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
	}
219 220 221 222 223 224
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL3, tmp);

	tmp = mmMMVM_L2_CNTL4_DEFAULT;
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL4, tmp);
225 226 227 228

	tmp = mmMMVM_L2_CNTL5_DEFAULT;
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
	WREG32_SOC15(GC, 0, mmMMVM_L2_CNTL5, tmp);
229 230 231 232 233 234 235 236 237
}

static void mmhub_v2_0_enable_system_domain(struct amdgpu_device *adev)
{
	uint32_t tmp;

	tmp = RREG32_SOC15(MMHUB, 0, mmMMVM_CONTEXT0_CNTL);
	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
238 239
	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL,
			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
240 241 242 243 244
	WREG32_SOC15(MMHUB, 0, mmMMVM_CONTEXT0_CNTL, tmp);
}

static void mmhub_v2_0_disable_identity_aperture(struct amdgpu_device *adev)
{
245 246 247 248 249 250
	/* These registers are not accessible to VF-SRIOV.
	 * The PF will program them instead.
	 */
	if (amdgpu_sriov_vf(adev))
		return;

251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270
	WREG32_SOC15(MMHUB, 0,
		     mmMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
		     0xFFFFFFFF);
	WREG32_SOC15(MMHUB, 0,
		     mmMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
		     0x0000000F);

	WREG32_SOC15(MMHUB, 0,
		     mmMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, 0);
	WREG32_SOC15(MMHUB, 0,
		     mmMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, 0);

	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32,
		     0);
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32,
		     0);
}

static void mmhub_v2_0_setup_vmid_config(struct amdgpu_device *adev)
{
271
	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
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
	int i;
	uint32_t tmp;

	for (i = 0; i <= 14; i++) {
		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_CNTL, i);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
				    adev->vm_manager.num_level);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
				    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
				    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT,
				    1);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
				    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
				    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
				    READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
				    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
				    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
				    PAGE_TABLE_BLOCK_SIZE,
				    adev->vm_manager.block_size - 9);
		/* Send no-retry XNACK on fault to suppress VM fault storm. */
		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
300 301
				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
				    !amdgpu_noretry);
302 303 304 305 306 307 308 309 310 311 312 313
		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_CNTL,
				    i * hub->ctx_distance, tmp);
		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
				    i * hub->ctx_addr_distance, 0);
		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
				    i * hub->ctx_addr_distance, 0);
		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
				    i * hub->ctx_addr_distance,
				    lower_32_bits(adev->vm_manager.max_pfn - 1));
		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
				    i * hub->ctx_addr_distance,
				    upper_32_bits(adev->vm_manager.max_pfn - 1));
314 315 316 317 318
	}
}

static void mmhub_v2_0_program_invalidation(struct amdgpu_device *adev)
{
319
	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
320 321 322 323
	unsigned i;

	for (i = 0; i < 18; ++i) {
		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
324
				    i * hub->eng_addr_distance, 0xffffffff);
325
		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
326
				    i * hub->eng_addr_distance, 0x1f);
327 328 329
	}
}

330
static int mmhub_v2_0_gart_enable(struct amdgpu_device *adev)
331 332 333 334 335 336 337 338 339 340 341 342 343 344 345
{
	/* GART Enable. */
	mmhub_v2_0_init_gart_aperture_regs(adev);
	mmhub_v2_0_init_system_aperture_regs(adev);
	mmhub_v2_0_init_tlb_regs(adev);
	mmhub_v2_0_init_cache_regs(adev);

	mmhub_v2_0_enable_system_domain(adev);
	mmhub_v2_0_disable_identity_aperture(adev);
	mmhub_v2_0_setup_vmid_config(adev);
	mmhub_v2_0_program_invalidation(adev);

	return 0;
}

346
static void mmhub_v2_0_gart_disable(struct amdgpu_device *adev)
347
{
348
	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
349 350 351 352 353
	u32 tmp;
	u32 i;

	/* Disable all tables */
	for (i = 0; i < 16; i++)
354 355
		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT0_CNTL,
				    i * hub->ctx_distance, 0);
356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376

	/* Setup TLB control */
	tmp = RREG32_SOC15(MMHUB, 0, mmMMMC_VM_MX_L1_TLB_CNTL);
	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_MX_L1_TLB_CNTL, tmp);

	/* Setup L2 cache */
	tmp = RREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 0);
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL, tmp);
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL3, 0);
}

/**
 * mmhub_v2_0_set_fault_enable_default - update GART/VM fault handling
 *
 * @adev: amdgpu_device pointer
 * @value: true redirects VM faults to the default page
 */
377
static void mmhub_v2_0_set_fault_enable_default(struct amdgpu_device *adev, bool value)
378 379
{
	u32 tmp;
380 381 382 383 384 385 386

	/* These registers are not accessible to VF-SRIOV.
	 * The PF will program them instead.
	 */
	if (amdgpu_sriov_vf(adev))
		return;

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
	tmp = RREG32_SOC15(MMHUB, 0, mmMMVM_L2_PROTECTION_FAULT_CNTL);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
			    value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	if (!value) {
		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
				CRASH_ON_NO_RETRY_FAULT, 1);
		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
				CRASH_ON_RETRY_FAULT, 1);
	}
	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_PROTECTION_FAULT_CNTL, tmp);
}

420 421
static const struct amdgpu_vmhub_funcs mmhub_v2_0_vmhub_funcs = {
	.print_l2_protection_fault_status = mmhub_v2_0_print_l2_protection_fault_status,
422
	.get_invalidate_req = mmhub_v2_0_get_invalidate_req,
423 424
};

425
static void mmhub_v2_0_init(struct amdgpu_device *adev)
426
{
427
	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
428 429 430 431 432 433 434

	hub->ctx0_ptb_addr_lo32 =
		SOC15_REG_OFFSET(MMHUB, 0,
				 mmMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
	hub->ctx0_ptb_addr_hi32 =
		SOC15_REG_OFFSET(MMHUB, 0,
				 mmMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
435 436
	hub->vm_inv_eng0_sem =
		SOC15_REG_OFFSET(MMHUB, 0, mmMMVM_INVALIDATE_ENG0_SEM);
437 438 439 440 441 442 443 444 445 446 447
	hub->vm_inv_eng0_req =
		SOC15_REG_OFFSET(MMHUB, 0, mmMMVM_INVALIDATE_ENG0_REQ);
	hub->vm_inv_eng0_ack =
		SOC15_REG_OFFSET(MMHUB, 0, mmMMVM_INVALIDATE_ENG0_ACK);
	hub->vm_context0_cntl =
		SOC15_REG_OFFSET(MMHUB, 0, mmMMVM_CONTEXT0_CNTL);
	hub->vm_l2_pro_fault_status =
		SOC15_REG_OFFSET(MMHUB, 0, mmMMVM_L2_PROTECTION_FAULT_STATUS);
	hub->vm_l2_pro_fault_cntl =
		SOC15_REG_OFFSET(MMHUB, 0, mmMMVM_L2_PROTECTION_FAULT_CNTL);

448 449 450 451 452 453 454
	hub->ctx_distance = mmMMVM_CONTEXT1_CNTL - mmMMVM_CONTEXT0_CNTL;
	hub->ctx_addr_distance = mmMMVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
		mmMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
	hub->eng_distance = mmMMVM_INVALIDATE_ENG1_REQ -
		mmMMVM_INVALIDATE_ENG0_REQ;
	hub->eng_addr_distance = mmMMVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
		mmMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
455 456 457 458 459 460 461 462

	hub->vm_cntx_cntl_vm_fault = MMVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		MMVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		MMVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		MMVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		MMVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		MMVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
		MMVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
463 464

	hub->vmhub_funcs = &mmhub_v2_0_vmhub_funcs;
465 466 467 468 469 470 471
}

static void mmhub_v2_0_update_medium_grain_clock_gating(struct amdgpu_device *adev,
							bool enable)
{
	uint32_t def, data, def1, data1;

472 473
	switch (adev->asic_type) {
	case CHIP_SIENNA_CICHLID:
474
	case CHIP_NAVY_FLOUNDER:
475 476 477 478 479 480 481 482
		def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid);
		def1 = data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid);
		break;
	default:
		def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG);
		def1 = data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2);
		break;
	}
483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504

	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_MGCG)) {
		data |= MM_ATC_L2_MISC_CG__ENABLE_MASK;

		data1 &= ~(DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
		           DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
		           DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
		           DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
		           DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
		           DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);

	} else {
		data &= ~MM_ATC_L2_MISC_CG__ENABLE_MASK;

		data1 |= (DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
			  DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
			  DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
			  DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
			  DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
			  DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);
	}

505 506
	switch (adev->asic_type) {
	case CHIP_SIENNA_CICHLID:
507
	case CHIP_NAVY_FLOUNDER:
508 509 510 511 512 513 514 515 516 517 518 519
		if (def != data)
			WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid, data);
		if (def1 != data1)
			WREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid, data1);
		break;
	default:
		if (def != data)
			WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG, data);
		if (def1 != data1)
			WREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2, data1);
		break;
	}
520 521 522 523 524 525 526
}

static void mmhub_v2_0_update_medium_grain_light_sleep(struct amdgpu_device *adev,
						       bool enable)
{
	uint32_t def, data;

527 528
	switch (adev->asic_type) {
	case CHIP_SIENNA_CICHLID:
529
	case CHIP_NAVY_FLOUNDER:
530 531 532 533 534 535
		def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid);
		break;
	default:
		def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG);
		break;
	}
536 537 538 539 540 541

	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_LS))
		data |= MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
	else
		data &= ~MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;

542 543 544
	if (def != data) {
		switch (adev->asic_type) {
		case CHIP_SIENNA_CICHLID:
545
		case CHIP_NAVY_FLOUNDER:
546 547 548 549 550 551 552
			WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid, data);
			break;
		default:
			WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG, data);
			break;
		}
	}
553 554
}

555
static int mmhub_v2_0_set_clockgating(struct amdgpu_device *adev,
556 557 558 559 560 561 562
			       enum amd_clockgating_state state)
{
	if (amdgpu_sriov_vf(adev))
		return 0;

	switch (adev->asic_type) {
	case CHIP_NAVI10:
563
	case CHIP_NAVI14:
564
	case CHIP_NAVI12:
565
	case CHIP_SIENNA_CICHLID:
566
	case CHIP_NAVY_FLOUNDER:
567
		mmhub_v2_0_update_medium_grain_clock_gating(adev,
568
				state == AMD_CG_STATE_GATE);
569
		mmhub_v2_0_update_medium_grain_light_sleep(adev,
570
				state == AMD_CG_STATE_GATE);
571 572 573 574 575 576 577 578
		break;
	default:
		break;
	}

	return 0;
}

579
static void mmhub_v2_0_get_clockgating(struct amdgpu_device *adev, u32 *flags)
580 581 582 583 584 585
{
	int data, data1;

	if (amdgpu_sriov_vf(adev))
		*flags = 0;

586 587
	switch (adev->asic_type) {
	case CHIP_SIENNA_CICHLID:
588
	case CHIP_NAVY_FLOUNDER:
589 590 591 592 593 594 595 596
		data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid);
		data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid);
		break;
	default:
		data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG);
		data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2);
		break;
	}
597 598 599 600 601 602 603 604 605 606 607 608 609 610 611

	/* AMD_CG_SUPPORT_MC_MGCG */
	if ((data & MM_ATC_L2_MISC_CG__ENABLE_MASK) &&
	    !(data1 & (DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
		       DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
		       DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
		       DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
		       DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
		       DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK)))
		*flags |= AMD_CG_SUPPORT_MC_MGCG;

	/* AMD_CG_SUPPORT_MC_LS */
	if (data & MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK)
		*flags |= AMD_CG_SUPPORT_MC_LS;
}
612 613 614 615 616 617 618 619 620 621 622

const struct amdgpu_mmhub_funcs mmhub_v2_0_funcs = {
	.ras_late_init = amdgpu_mmhub_ras_late_init,
	.init = mmhub_v2_0_init,
	.gart_enable = mmhub_v2_0_gart_enable,
	.set_fault_enable_default = mmhub_v2_0_set_fault_enable_default,
	.gart_disable = mmhub_v2_0_gart_disable,
	.set_clockgating = mmhub_v2_0_set_clockgating,
	.get_clockgating = mmhub_v2_0_get_clockgating,
	.setup_vm_pt_regs = mmhub_v2_0_setup_vm_pt_regs,
};