Contributors: 9
Author Tokens Token Proportion Commits Commit Proportion
Hawking Zhang 1736 90.61% 1 9.09%
Alex Deucher 65 3.39% 1 9.09%
Yong Zhao 43 2.24% 2 18.18%
Rohit Khaire 19 0.99% 1 9.09%
Felix Kuhling 15 0.78% 2 18.18%
changzhu 13 0.68% 1 9.09%
Jack Xiao 13 0.68% 1 9.09%
Tiecheng Zhou 11 0.57% 1 9.09%
Le Ma 1 0.05% 1 9.09%
Total 1916 11 


/*
 * 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 "gfxhub_v2_0.h"

#include "gc/gc_10_1_0_offset.h"
#include "gc/gc_10_1_0_sh_mask.h"
#include "gc/gc_10_1_0_default.h"
#include "navi10_enum.h"

#include "soc15_common.h"

u64 gfxhub_v2_0_get_fb_location(struct amdgpu_device *adev)
{
	u64 base = RREG32_SOC15(GC, 0, mmGCMC_VM_FB_LOCATION_BASE);

	base &= GCMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
	base <<= 24;

	return base;
}

u64 gfxhub_v2_0_get_mc_fb_offset(struct amdgpu_device *adev)
{
	return (u64)RREG32_SOC15(GC, 0, mmGCMC_VM_FB_OFFSET) << 24;
}

void gfxhub_v2_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
				uint64_t page_table_base)
{
	/* two registers distance between mmGCVM_CONTEXT0_* to mmGCVM_CONTEXT1_* */
	int offset = mmGCVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32
			- mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;

	WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
				offset * vmid, lower_32_bits(page_table_base));

	WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
				offset * vmid, upper_32_bits(page_table_base));
}

static void gfxhub_v2_0_init_gart_aperture_regs(struct amdgpu_device *adev)
{
	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);

	gfxhub_v2_0_setup_vm_pt_regs(adev, 0, pt_base);

	WREG32_SOC15(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
		     (u32)(adev->gmc.gart_start >> 12));
	WREG32_SOC15(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
		     (u32)(adev->gmc.gart_start >> 44));

	WREG32_SOC15(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
		     (u32)(adev->gmc.gart_end >> 12));
	WREG32_SOC15(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
		     (u32)(adev->gmc.gart_end >> 44));
}

static void gfxhub_v2_0_init_system_aperture_regs(struct amdgpu_device *adev)
{
	uint64_t value;

	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.
		 */
		/* Disable AGP. */
		WREG32_SOC15(GC, 0, mmGCMC_VM_AGP_BASE, 0);
		WREG32_SOC15(GC, 0, mmGCMC_VM_AGP_TOP, 0);
		WREG32_SOC15(GC, 0, mmGCMC_VM_AGP_BOT, 0x00FFFFFF);

		/* Program the system aperture low logical page number. */
		WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_LOW_ADDR,
			     adev->gmc.vram_start >> 18);
		WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
			     adev->gmc.vram_end >> 18);

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

	/* Program "protection fault". */
	WREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
		     (u32)(adev->dummy_page_addr >> 12));
	WREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
		     (u32)((u64)adev->dummy_page_addr >> 44));

	WREG32_FIELD15(GC, 0, GCVM_L2_PROTECTION_FAULT_CNTL2,
		       ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
}


static void gfxhub_v2_0_init_tlb_regs(struct amdgpu_device *adev)
{
	uint32_t tmp;

	/* Setup TLB control */
	tmp = RREG32_SOC15(GC, 0, mmGCMC_VM_MX_L1_TLB_CNTL);

	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
			    MTYPE, MTYPE_UC); /* UC, uncached */

	WREG32_SOC15(GC, 0, mmGCMC_VM_MX_L1_TLB_CNTL, tmp);
}

static void gfxhub_v2_0_init_cache_regs(struct amdgpu_device *adev)
{
	uint32_t tmp;

	/* These regs are not accessible for VF, PF will program these in SRIOV */
	if (amdgpu_sriov_vf(adev))
		return;

	/* Setup L2 cache */
	tmp = RREG32_SOC15(GC, 0, mmGCVM_L2_CNTL);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_CACHE, 1);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
	/* XXX for emulation, Refer to closed source code.*/
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
			    L2_PDE0_CACHE_TAG_GENERATION_MODE, 0);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
	WREG32_SOC15(GC, 0, mmGCVM_L2_CNTL, tmp);

	tmp = RREG32_SOC15(GC, 0, mmGCVM_L2_CNTL2);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
	WREG32_SOC15(GC, 0, mmGCVM_L2_CNTL2, tmp);

	tmp = mmGCVM_L2_CNTL3_DEFAULT;
	if (adev->gmc.translate_further) {
		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 12);
		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
	} else {
		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 9);
		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
	}
	WREG32_SOC15(GC, 0, mmGCVM_L2_CNTL3, tmp);

	tmp = mmGCVM_L2_CNTL4_DEFAULT;
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
	WREG32_SOC15(GC, 0, mmGCVM_L2_CNTL4, tmp);
}

static void gfxhub_v2_0_enable_system_domain(struct amdgpu_device *adev)
{
	uint32_t tmp;

	tmp = RREG32_SOC15(GC, 0, mmGCVM_CONTEXT0_CNTL);
	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL,
			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
	WREG32_SOC15(GC, 0, mmGCVM_CONTEXT0_CNTL, tmp);
}

static void gfxhub_v2_0_disable_identity_aperture(struct amdgpu_device *adev)
{
	WREG32_SOC15(GC, 0, mmGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
		     0xFFFFFFFF);
	WREG32_SOC15(GC, 0, mmGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
		     0x0000000F);

	WREG32_SOC15(GC, 0, mmGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32,
		     0);
	WREG32_SOC15(GC, 0, mmGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32,
		     0);

	WREG32_SOC15(GC, 0, mmGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0);
	WREG32_SOC15(GC, 0, mmGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0);

}

static void gfxhub_v2_0_setup_vmid_config(struct amdgpu_device *adev)
{
	int i;
	uint32_t tmp;

	for (i = 0; i <= 14; i++) {
		tmp = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_CNTL, i);
		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
				    adev->vm_manager.num_level);
		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
				VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
				READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
				EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
		tmp = REG_SET_FIELD(tmp, GCVM_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, GCVM_CONTEXT1_CNTL,
				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
				    !amdgpu_noretry);
		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_CNTL, i, tmp);
		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, i*2, 0);
		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, i*2, 0);
		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,  i*2,
			lower_32_bits(adev->vm_manager.max_pfn - 1));
		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32, i*2,
			upper_32_bits(adev->vm_manager.max_pfn - 1));
	}
}

static void gfxhub_v2_0_program_invalidation(struct amdgpu_device *adev)
{
	unsigned i;

	for (i = 0 ; i < 18; ++i) {
		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
				    2 * i, 0xffffffff);
		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
				    2 * i, 0x1f);
	}
}

int gfxhub_v2_0_gart_enable(struct amdgpu_device *adev)
{
	/* GART Enable. */
	gfxhub_v2_0_init_gart_aperture_regs(adev);
	gfxhub_v2_0_init_system_aperture_regs(adev);
	gfxhub_v2_0_init_tlb_regs(adev);
	gfxhub_v2_0_init_cache_regs(adev);

	gfxhub_v2_0_enable_system_domain(adev);
	gfxhub_v2_0_disable_identity_aperture(adev);
	gfxhub_v2_0_setup_vmid_config(adev);
	gfxhub_v2_0_program_invalidation(adev);

	return 0;
}

void gfxhub_v2_0_gart_disable(struct amdgpu_device *adev)
{
	u32 tmp;
	u32 i;

	/* Disable all tables */
	for (i = 0; i < 16; i++)
		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_CNTL, i, 0);

	/* Setup TLB control */
	tmp = RREG32_SOC15(GC, 0, mmGCMC_VM_MX_L1_TLB_CNTL);
	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
	WREG32_SOC15(GC, 0, mmGCMC_VM_MX_L1_TLB_CNTL, tmp);

	if (!amdgpu_sriov_vf(adev)) {
		/* Setup L2 cache */
		WREG32_FIELD15(GC, 0, GCVM_L2_CNTL, ENABLE_L2_CACHE, 0);
		WREG32_SOC15(GC, 0, mmGCVM_L2_CNTL3, 0);
	}
}

/**
 * gfxhub_v2_0_set_fault_enable_default - update GART/VM fault handling
 *
 * @adev: amdgpu_device pointer
 * @value: true redirects VM faults to the default page
 */
void gfxhub_v2_0_set_fault_enable_default(struct amdgpu_device *adev,
					  bool value)
{
	u32 tmp;
	tmp = RREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_CNTL);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
			    value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
	if (!value) {
		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
				CRASH_ON_NO_RETRY_FAULT, 1);
		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
				CRASH_ON_RETRY_FAULT, 1);
	}
	WREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_CNTL, tmp);
}

void gfxhub_v2_0_init(struct amdgpu_device *adev)
{
	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];

	hub->ctx0_ptb_addr_lo32 =
		SOC15_REG_OFFSET(GC, 0,
				 mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
	hub->ctx0_ptb_addr_hi32 =
		SOC15_REG_OFFSET(GC, 0,
				 mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
	hub->vm_inv_eng0_sem =
		SOC15_REG_OFFSET(GC, 0, mmGCVM_INVALIDATE_ENG0_SEM);
	hub->vm_inv_eng0_req =
		SOC15_REG_OFFSET(GC, 0, mmGCVM_INVALIDATE_ENG0_REQ);
	hub->vm_inv_eng0_ack =
		SOC15_REG_OFFSET(GC, 0, mmGCVM_INVALIDATE_ENG0_ACK);
	hub->vm_context0_cntl =
		SOC15_REG_OFFSET(GC, 0, mmGCVM_CONTEXT0_CNTL);
	hub->vm_l2_pro_fault_status =
		SOC15_REG_OFFSET(GC, 0, mmGCVM_L2_PROTECTION_FAULT_STATUS);
	hub->vm_l2_pro_fault_cntl =
		SOC15_REG_OFFSET(GC, 0, mmGCVM_L2_PROTECTION_FAULT_CNTL);
}