diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c | 183 |
1 files changed, 142 insertions, 41 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c index c0e3efcb09bf..fe0710b55c3a 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c @@ -36,7 +36,7 @@ #define MMHUB_NUM_INSTANCES 2 #define MMHUB_INSTANCE_REGISTER_OFFSET 0x3000 -u64 mmhub_v9_4_get_fb_location(struct amdgpu_device *adev) +static u64 mmhub_v9_4_get_fb_location(struct amdgpu_device *adev) { /* The base should be same b/t 2 mmhubs on Acrturus. Read one here. */ u64 base = RREG32_SOC15(MMHUB, 0, mmVMSHAREDVC0_MC_VM_FB_LOCATION_BASE); @@ -57,20 +57,16 @@ u64 mmhub_v9_4_get_fb_location(struct amdgpu_device *adev) static void mmhub_v9_4_setup_hubid_vm_pt_regs(struct amdgpu_device *adev, int hubid, uint32_t vmid, uint64_t value) { - /* two registers distance between mmVML2VC0_VM_CONTEXT0_* to - * mmVML2VC0_VM_CONTEXT1_* - */ - int dist = mmVML2VC0_VM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 - - mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, - dist * vmid + hubid * MMHUB_INSTANCE_REGISTER_OFFSET, + hub->ctx_addr_distance * vmid + hubid * MMHUB_INSTANCE_REGISTER_OFFSET, lower_32_bits(value)); WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32, - dist * vmid + hubid * MMHUB_INSTANCE_REGISTER_OFFSET, + hub->ctx_addr_distance * vmid + hubid * MMHUB_INSTANCE_REGISTER_OFFSET, upper_32_bits(value)); } @@ -101,7 +97,7 @@ static void mmhub_v9_4_init_gart_aperture_regs(struct amdgpu_device *adev, (u32)(adev->gmc.gart_end >> 44)); } -void mmhub_v9_4_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, +static void mmhub_v9_4_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { int i; @@ -112,7 +108,7 @@ void mmhub_v9_4_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, } static void mmhub_v9_4_init_system_aperture_regs(struct amdgpu_device *adev, - int hubid) + int hubid) { uint64_t value; uint32_t tmp; @@ -140,8 +136,7 @@ static void mmhub_v9_4_init_system_aperture_regs(struct amdgpu_device *adev, max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18); /* Set default page address. */ - value = adev->vram_scratch.gpu_addr - adev->gmc.vram_start + - adev->vm_manager.vram_base_offset; + value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr); WREG32_SOC15_OFFSET( MMHUB, 0, mmVMSHAREDPF0_MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, @@ -195,8 +190,6 @@ static void mmhub_v9_4_init_tlb_regs(struct amdgpu_device *adev, int hubid) tmp = REG_SET_FIELD(tmp, VMSHAREDVC0_MC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); tmp = REG_SET_FIELD(tmp, VMSHAREDVC0_MC_VM_MX_L1_TLB_CNTL, - ECO_BITS, 0); - tmp = REG_SET_FIELD(tmp, VMSHAREDVC0_MC_VM_MX_L1_TLB_CNTL, MTYPE, MTYPE_UC);/* XXX for emulation. */ tmp = REG_SET_FIELD(tmp, VMSHAREDVC0_MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); @@ -205,6 +198,36 @@ static void mmhub_v9_4_init_tlb_regs(struct amdgpu_device *adev, int hubid) hubid * MMHUB_INSTANCE_REGISTER_OFFSET, tmp); } +/* Set snoop bit for SDMA so that SDMA writes probe-invalidates RW lines */ +static void mmhub_v9_4_init_snoop_override_regs(struct amdgpu_device *adev, int hubid) +{ + uint32_t tmp; + int i; + uint32_t distance = mmDAGB1_WRCLI_GPU_SNOOP_OVERRIDE - + mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE; + uint32_t huboffset = hubid * MMHUB_INSTANCE_REGISTER_OFFSET; + + for (i = 0; i < 5 - (2 * hubid); i++) { + /* DAGB instances 0 to 4 are in hub0 and 5 to 7 are in hub1 */ + tmp = RREG32_SOC15_OFFSET(MMHUB, 0, + mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, + huboffset + i * distance); + tmp |= (1 << 15); /* SDMA client is BIT15 */ + WREG32_SOC15_OFFSET(MMHUB, 0, + mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, + huboffset + i * distance, tmp); + + tmp = RREG32_SOC15_OFFSET(MMHUB, 0, + mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, + huboffset + i * distance); + tmp |= (1 << 15); + WREG32_SOC15_OFFSET(MMHUB, 0, + mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, + huboffset + i * distance, tmp); + } + +} + static void mmhub_v9_4_init_cache_regs(struct amdgpu_device *adev, int hubid) { uint32_t tmp; @@ -301,17 +324,26 @@ static void mmhub_v9_4_disable_identity_aperture(struct amdgpu_device *adev, static void mmhub_v9_4_setup_vmid_config(struct amdgpu_device *adev, int hubid) { + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; + unsigned int num_level, block_size; uint32_t tmp; int i; + num_level = adev->vm_manager.num_level; + block_size = adev->vm_manager.block_size; + if (adev->gmc.translate_further) + num_level -= 1; + else + block_size -= 9; + for (i = 0; i <= 14; i++) { tmp = RREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT1_CNTL, - hubid * MMHUB_INSTANCE_REGISTER_OFFSET + i); + hubid * MMHUB_INSTANCE_REGISTER_OFFSET + i * hub->ctx_distance); tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1); tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH, - adev->vm_manager.num_level); + num_level); tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL, RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL, @@ -329,27 +361,31 @@ static void mmhub_v9_4_setup_vmid_config(struct amdgpu_device *adev, int hubid) EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL, PAGE_TABLE_BLOCK_SIZE, - adev->vm_manager.block_size - 9); + block_size); /* Send no-retry XNACK on fault to suppress VM fault storm. */ tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL, RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, - !amdgpu_noretry); + !adev->gmc.noretry); WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT1_CNTL, - hubid * MMHUB_INSTANCE_REGISTER_OFFSET + i, - tmp); + hubid * MMHUB_INSTANCE_REGISTER_OFFSET + + i * hub->ctx_distance, tmp); WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, - hubid * MMHUB_INSTANCE_REGISTER_OFFSET + i*2, 0); + hubid * MMHUB_INSTANCE_REGISTER_OFFSET + + i * hub->ctx_addr_distance, 0); WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, - hubid * MMHUB_INSTANCE_REGISTER_OFFSET + i*2, 0); + hubid * MMHUB_INSTANCE_REGISTER_OFFSET + + i * hub->ctx_addr_distance, 0); WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32, - hubid * MMHUB_INSTANCE_REGISTER_OFFSET + i*2, + hubid * MMHUB_INSTANCE_REGISTER_OFFSET + + i * hub->ctx_addr_distance, lower_32_bits(adev->vm_manager.max_pfn - 1)); WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32, - hubid * MMHUB_INSTANCE_REGISTER_OFFSET + i*2, + hubid * MMHUB_INSTANCE_REGISTER_OFFSET + + i * hub->ctx_addr_distance, upper_32_bits(adev->vm_manager.max_pfn - 1)); } } @@ -357,21 +393,24 @@ static void mmhub_v9_4_setup_vmid_config(struct amdgpu_device *adev, int hubid) static void mmhub_v9_4_program_invalidation(struct amdgpu_device *adev, int hubid) { + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned i; for (i = 0; i < 18; ++i) { WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_INVALIDATE_ENG0_ADDR_RANGE_LO32, - hubid * MMHUB_INSTANCE_REGISTER_OFFSET + 2 * i, + hubid * MMHUB_INSTANCE_REGISTER_OFFSET + + i * hub->eng_addr_distance, 0xffffffff); WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_INVALIDATE_ENG0_ADDR_RANGE_HI32, - hubid * MMHUB_INSTANCE_REGISTER_OFFSET + 2 * i, + hubid * MMHUB_INSTANCE_REGISTER_OFFSET + + i * hub->eng_addr_distance, 0x1f); } } -int mmhub_v9_4_gart_enable(struct amdgpu_device *adev) +static int mmhub_v9_4_gart_enable(struct amdgpu_device *adev) { int i; @@ -383,6 +422,7 @@ int mmhub_v9_4_gart_enable(struct amdgpu_device *adev) if (!amdgpu_sriov_vf(adev)) mmhub_v9_4_init_cache_regs(adev, i); + mmhub_v9_4_init_snoop_override_regs(adev, i); mmhub_v9_4_enable_system_domain(adev, i); if (!amdgpu_sriov_vf(adev)) mmhub_v9_4_disable_identity_aperture(adev, i); @@ -393,18 +433,19 @@ int mmhub_v9_4_gart_enable(struct amdgpu_device *adev) return 0; } -void mmhub_v9_4_gart_disable(struct amdgpu_device *adev) +static void mmhub_v9_4_gart_disable(struct amdgpu_device *adev) { + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; u32 tmp; u32 i, j; for (j = 0; j < MMHUB_NUM_INSTANCES; j++) { /* Disable all tables */ - for (i = 0; i < 16; i++) + for (i = 0; i < AMDGPU_NUM_VMID; i++) WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT0_CNTL, j * MMHUB_INSTANCE_REGISTER_OFFSET + - i, 0); + i * hub->ctx_distance, 0); /* Setup TLB control */ tmp = RREG32_SOC15_OFFSET(MMHUB, 0, @@ -432,12 +473,12 @@ void mmhub_v9_4_gart_disable(struct amdgpu_device *adev) } /** - * mmhub_v1_0_set_fault_enable_default - update GART/VM fault handling + * mmhub_v9_4_set_fault_enable_default - update GART/VM fault handling * * @adev: amdgpu_device pointer * @value: true redirects VM faults to the default page */ -void mmhub_v9_4_set_fault_enable_default(struct amdgpu_device *adev, bool value) +static void mmhub_v9_4_set_fault_enable_default(struct amdgpu_device *adev, bool value) { u32 tmp; int i; @@ -495,10 +536,10 @@ void mmhub_v9_4_set_fault_enable_default(struct amdgpu_device *adev, bool value) } } -void mmhub_v9_4_init(struct amdgpu_device *adev) +static void mmhub_v9_4_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub[MMHUB_NUM_INSTANCES] = - {&adev->vmhub[AMDGPU_MMHUB_0], &adev->vmhub[AMDGPU_MMHUB_1]}; + struct amdgpu_vmhub *hub[MMHUB_NUM_INSTANCES] = { + &adev->vmhub[AMDGPU_MMHUB0(0)], &adev->vmhub[AMDGPU_MMHUB1(0)]}; int i; for (i = 0; i < MMHUB_NUM_INSTANCES; i++) { @@ -534,6 +575,15 @@ void mmhub_v9_4_init(struct amdgpu_device *adev) SOC15_REG_OFFSET(MMHUB, 0, mmVML2PF0_VM_L2_PROTECTION_FAULT_CNTL) + i * MMHUB_INSTANCE_REGISTER_OFFSET; + + hub[i]->ctx_distance = mmVML2VC0_VM_CONTEXT1_CNTL - + mmVML2VC0_VM_CONTEXT0_CNTL; + hub[i]->ctx_addr_distance = mmVML2VC0_VM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 - + mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32; + hub[i]->eng_distance = mmVML2VC0_VM_INVALIDATE_ENG1_REQ - + mmVML2VC0_VM_INVALIDATE_ENG0_REQ; + hub[i]->eng_addr_distance = mmVML2VC0_VM_INVALIDATE_ENG1_ADDR_RANGE_LO32 - + mmVML2VC0_VM_INVALIDATE_ENG0_ADDR_RANGE_LO32; } } @@ -616,7 +666,7 @@ static void mmhub_v9_4_update_medium_grain_light_sleep(struct amdgpu_device *ade } } -int mmhub_v9_4_set_clockgating(struct amdgpu_device *adev, +static int mmhub_v9_4_set_clockgating(struct amdgpu_device *adev, enum amd_clockgating_state state) { if (amdgpu_sriov_vf(adev)) @@ -636,9 +686,9 @@ int mmhub_v9_4_set_clockgating(struct amdgpu_device *adev, return 0; } -void mmhub_v9_4_get_clockgating(struct amdgpu_device *adev, u32 *flags) +static void mmhub_v9_4_get_clockgating(struct amdgpu_device *adev, u64 *flags) { - int data, data1; + u32 data, data1; if (amdgpu_sriov_vf(adev)) *flags = 0; @@ -1549,7 +1599,7 @@ static int mmhub_v9_4_get_ras_error_count(struct amdgpu_device *adev, uint32_t sec_cnt, ded_cnt; for (i = 0; i < ARRAY_SIZE(mmhub_v9_4_ras_fields); i++) { - if(mmhub_v9_4_ras_fields[i].reg_offset != reg->reg_offset) + if (mmhub_v9_4_ras_fields[i].reg_offset != reg->reg_offset) continue; sec_cnt = (value & @@ -1610,8 +1660,59 @@ static void mmhub_v9_4_reset_ras_error_count(struct amdgpu_device *adev) } } -const struct amdgpu_mmhub_funcs mmhub_v9_4_funcs = { - .ras_late_init = amdgpu_mmhub_ras_late_init, +static const struct soc15_reg_entry mmhub_v9_4_err_status_regs[] = { + { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_ERR_STATUS), 0, 0, 0 }, + { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_ERR_STATUS), 0, 0, 0 }, + { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA2_ERR_STATUS), 0, 0, 0 }, + { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA3_ERR_STATUS), 0, 0, 0 }, + { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA4_ERR_STATUS), 0, 0, 0 }, + { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA5_ERR_STATUS), 0, 0, 0 }, + { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA6_ERR_STATUS), 0, 0, 0 }, + { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA7_ERR_STATUS), 0, 0, 0 }, +}; + +static void mmhub_v9_4_query_ras_error_status(struct amdgpu_device *adev) +{ + int i; + uint32_t reg_value; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__MMHUB)) + return; + + for (i = 0; i < ARRAY_SIZE(mmhub_v9_4_err_status_regs); i++) { + reg_value = + RREG32(SOC15_REG_ENTRY_OFFSET(mmhub_v9_4_err_status_regs[i])); + if (REG_GET_FIELD(reg_value, MMEA0_ERR_STATUS, SDP_RDRSP_STATUS) || + REG_GET_FIELD(reg_value, MMEA0_ERR_STATUS, SDP_WRRSP_STATUS) || + REG_GET_FIELD(reg_value, MMEA0_ERR_STATUS, SDP_RDRSP_DATAPARITY_ERROR)) { + /* SDP read/write error/parity error in FUE_IS_FATAL mode + * can cause system fatal error in arcturas. Harvest the error + * status before GPU reset */ + dev_warn(adev->dev, "MMHUB EA err detected at instance: %d, status: 0x%x!\n", + i, reg_value); + } + } +} + +const struct amdgpu_ras_block_hw_ops mmhub_v9_4_ras_hw_ops = { .query_ras_error_count = mmhub_v9_4_query_ras_error_count, .reset_ras_error_count = mmhub_v9_4_reset_ras_error_count, + .query_ras_error_status = mmhub_v9_4_query_ras_error_status, +}; + +struct amdgpu_mmhub_ras mmhub_v9_4_ras = { + .ras_block = { + .hw_ops = &mmhub_v9_4_ras_hw_ops, + }, +}; + +const struct amdgpu_mmhub_funcs mmhub_v9_4_funcs = { + .get_fb_location = mmhub_v9_4_get_fb_location, + .init = mmhub_v9_4_init, + .gart_enable = mmhub_v9_4_gart_enable, + .set_fault_enable_default = mmhub_v9_4_set_fault_enable_default, + .gart_disable = mmhub_v9_4_gart_disable, + .set_clockgating = mmhub_v9_4_set_clockgating, + .get_clockgating = mmhub_v9_4_get_clockgating, + .setup_vm_pt_regs = mmhub_v9_4_setup_vm_pt_regs, }; |