diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu')
123 files changed, 3525 insertions, 1681 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index 5b21674b07fb..aacc810cabb3 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/drivers/gpu/drm/amd/amdgpu/Makefile @@ -65,7 +65,8 @@ amdgpu-y += amdgpu_device.o amdgpu_doorbell_mgr.o amdgpu_kms.o \ amdgpu_umc.o smu_v11_0_i2c.o amdgpu_fru_eeprom.o amdgpu_rap.o \ amdgpu_fw_attestation.o amdgpu_securedisplay.o \ amdgpu_eeprom.o amdgpu_mca.o amdgpu_psp_ta.o amdgpu_lsdma.o \ - amdgpu_ring_mux.o amdgpu_xcp.o amdgpu_seq64.o amdgpu_aca.o amdgpu_dev_coredump.o + amdgpu_ring_mux.o amdgpu_xcp.o amdgpu_seq64.o amdgpu_aca.o amdgpu_dev_coredump.o \ + amdgpu_cper.o amdgpu-$(CONFIG_PROC_FS) += amdgpu_fdinfo.o diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 69895fccb474..2b1990ea9639 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -109,6 +109,7 @@ #include "amdgpu_mca.h" #include "amdgpu_aca.h" #include "amdgpu_ras.h" +#include "amdgpu_cper.h" #include "amdgpu_xcp.h" #include "amdgpu_seq64.h" #include "amdgpu_reg_state.h" @@ -415,6 +416,7 @@ bool amdgpu_get_bios(struct amdgpu_device *adev); bool amdgpu_read_bios(struct amdgpu_device *adev); bool amdgpu_soc15_read_bios_from_rom(struct amdgpu_device *adev, u8 *bios, u32 length_bytes); +void amdgpu_bios_release(struct amdgpu_device *adev); /* * Clocks */ @@ -1090,6 +1092,9 @@ struct amdgpu_device { /* ACA */ struct amdgpu_aca aca; + /* CPER */ + struct amdgpu_cper cper; + struct amdgpu_ip_block ip_blocks[AMDGPU_MAX_IP_NUM]; uint32_t harvest_ip_mask; int num_ip_blocks; @@ -1149,6 +1154,7 @@ struct amdgpu_device { struct ratelimit_state throttling_logging_rs; uint32_t ras_hw_enabled; uint32_t ras_enabled; + bool ras_default_ecc_enabled; bool no_hw_access; struct pci_saved_state *pci_state; @@ -1192,6 +1198,11 @@ struct amdgpu_device { struct mutex enforce_isolation_mutex; struct amdgpu_init_level *init_lvl; + + /* This flag is used to determine how VRAM allocations are handled for APUs + * in KFD: VRAM or GTT. + */ + bool apu_prefer_gtt; }; static inline uint32_t amdgpu_ip_version(const struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c index 9d6345146495..d11593cd1922 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c @@ -30,16 +30,6 @@ typedef int bank_handler_t(struct aca_handle *handle, struct aca_bank *bank, enum aca_smu_type type, void *data); -struct aca_banks { - int nr_banks; - struct list_head list; -}; - -struct aca_hwip { - int hwid; - int mcatype; -}; - static struct aca_hwip aca_hwid_mcatypes[ACA_HWIP_TYPE_COUNT] = { ACA_BANK_HWID(SMU, 0x01, 0x01), ACA_BANK_HWID(PCS_XGMI, 0x50, 0x00), @@ -111,7 +101,7 @@ static struct aca_regs_dump { {"STATUS", ACA_REG_IDX_STATUS}, {"ADDR", ACA_REG_IDX_ADDR}, {"MISC", ACA_REG_IDX_MISC0}, - {"CONFIG", ACA_REG_IDX_CONFG}, + {"CONFIG", ACA_REG_IDX_CONFIG}, {"IPID", ACA_REG_IDX_IPID}, {"SYND", ACA_REG_IDX_SYND}, {"DESTAT", ACA_REG_IDX_DESTAT}, @@ -168,7 +158,7 @@ static int aca_smu_get_valid_aca_banks(struct amdgpu_device *adev, enum aca_smu_ if (ret) return ret; - bank.type = type; + bank.smu_err_type = type; aca_smu_bank_dump(adev, i, count, &bank, qctx); @@ -394,6 +384,36 @@ static bool aca_bank_should_update(struct amdgpu_device *adev, enum aca_smu_type return ret; } +static void aca_banks_generate_cper(struct amdgpu_device *adev, + enum aca_smu_type type, + struct aca_banks *banks, + int count) +{ + struct aca_bank_node *node; + struct aca_bank *bank; + + if (!banks || !count) { + dev_warn(adev->dev, "fail to generate cper records\n"); + return; + } + + /* UEs must be encoded into separate CPER entries */ + if (type == ACA_SMU_TYPE_UE) { + list_for_each_entry(node, &banks->list, node) { + bank = &node->bank; + if (amdgpu_cper_generate_ue_record(adev, bank)) + dev_warn(adev->dev, "fail to generate ue cper records\n"); + } + } else { + /* + * SMU_TYPE_CE banks are combined into 1 CPER entries, + * they could be CEs or DEs or both + */ + if (amdgpu_cper_generate_ce_records(adev, banks, count)) + dev_warn(adev->dev, "fail to generate ce cper records\n"); + } +} + static int aca_banks_update(struct amdgpu_device *adev, enum aca_smu_type type, bank_handler_t handler, struct ras_query_context *qctx, void *data) { @@ -431,6 +451,8 @@ static int aca_banks_update(struct amdgpu_device *adev, enum aca_smu_type type, if (ret) goto err_release_banks; + aca_banks_generate_cper(adev, type, &banks, count); + err_release_banks: aca_banks_release(&banks); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h index f3289d289913..b84a3489b116 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h @@ -81,7 +81,7 @@ enum aca_reg_idx { ACA_REG_IDX_STATUS = 1, ACA_REG_IDX_ADDR = 2, ACA_REG_IDX_MISC0 = 3, - ACA_REG_IDX_CONFG = 4, + ACA_REG_IDX_CONFIG = 4, ACA_REG_IDX_IPID = 5, ACA_REG_IDX_SYND = 6, ACA_REG_IDX_DESTAT = 8, @@ -108,13 +108,20 @@ enum aca_error_type { }; enum aca_smu_type { + ACA_SMU_TYPE_INVALID = -1, ACA_SMU_TYPE_UE = 0, ACA_SMU_TYPE_CE, ACA_SMU_TYPE_COUNT, }; +struct aca_hwip { + int hwid; + int mcatype; +}; + struct aca_bank { - enum aca_smu_type type; + enum aca_error_type aca_err_type; + enum aca_smu_type smu_err_type; u64 regs[ACA_MAX_REGS_COUNT]; }; @@ -123,6 +130,11 @@ struct aca_bank_node { struct list_head node; }; +struct aca_banks { + int nr_banks; + struct list_head list; +}; + struct aca_bank_info { int die_id; int socket_id; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 2c1b38c5cfc6..0312231b703e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -459,7 +459,7 @@ void amdgpu_amdkfd_get_local_mem_info(struct amdgpu_device *adev, else mem_info->local_mem_size_private = KFD_XCP_MEMORY_SIZE(adev, xcp->id); - } else if (adev->flags & AMD_IS_APU) { + } else if (adev->apu_prefer_gtt) { mem_info->local_mem_size_public = (ttm_tt_pages_limit() << PAGE_SHIFT); mem_info->local_mem_size_private = 0; } else { @@ -818,7 +818,7 @@ u64 amdgpu_amdkfd_xcp_memory_size(struct amdgpu_device *adev, int xcp_id) } do_div(tmp, adev->xcp_mgr->num_xcp_per_mem_partition); return ALIGN_DOWN(tmp, PAGE_SIZE); - } else if (adev->flags & AMD_IS_APU) { + } else if (adev->apu_prefer_gtt) { return (ttm_tt_pages_limit() << PAGE_SHIFT); } else { return adev->gmc.real_vram_size; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index 8af67f18500a..236b73e283e8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -47,6 +47,7 @@ enum TLB_FLUSH_TYPE { }; struct amdgpu_device; +struct kfd_process_device; struct amdgpu_reset_context; enum kfd_mem_attachment_type { @@ -299,14 +300,10 @@ bool amdgpu_amdkfd_compute_active(struct amdgpu_device *adev, uint32_t node_id); (&((struct amdgpu_fpriv *) \ ((struct drm_file *)(drm_priv))->driver_priv)->vm) -int amdgpu_amdkfd_gpuvm_set_vm_pasid(struct amdgpu_device *adev, - struct amdgpu_vm *avm, u32 pasid); int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev, struct amdgpu_vm *avm, void **process_info, struct dma_fence **ef); -void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev, - void *drm_priv); uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv); size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev, uint8_t xcp_id); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 1e998f972c30..ea3f7ee18923 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -197,7 +197,7 @@ int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, return -EINVAL; vram_size = KFD_XCP_MEMORY_SIZE(adev, xcp_id); - if (adev->flags & AMD_IS_APU) { + if (adev->apu_prefer_gtt) { system_mem_needed = size; ttm_mem_needed = size; } @@ -234,7 +234,7 @@ int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, if (adev && xcp_id >= 0) { adev->kfd.vram_used[xcp_id] += vram_needed; adev->kfd.vram_used_aligned[xcp_id] += - (adev->flags & AMD_IS_APU) ? + adev->apu_prefer_gtt ? vram_needed : ALIGN(vram_needed, VRAM_AVAILABLITY_ALIGN); } @@ -262,7 +262,7 @@ void amdgpu_amdkfd_unreserve_mem_limit(struct amdgpu_device *adev, if (adev) { adev->kfd.vram_used[xcp_id] -= size; - if (adev->flags & AMD_IS_APU) { + if (adev->apu_prefer_gtt) { adev->kfd.vram_used_aligned[xcp_id] -= size; kfd_mem_limit.system_mem_used -= size; kfd_mem_limit.ttm_mem_used -= size; @@ -890,7 +890,7 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, * if peer device has large BAR. In contrast, access over xGMI is * allowed for both small and large BAR configurations of peer device */ - if ((adev != bo_adev && !(adev->flags & AMD_IS_APU)) && + if ((adev != bo_adev && !adev->apu_prefer_gtt) && ((mem->domain == AMDGPU_GEM_DOMAIN_VRAM) || (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL) || (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP))) { @@ -1529,27 +1529,6 @@ static void amdgpu_amdkfd_gpuvm_unpin_bo(struct amdgpu_bo *bo) amdgpu_bo_unreserve(bo); } -int amdgpu_amdkfd_gpuvm_set_vm_pasid(struct amdgpu_device *adev, - struct amdgpu_vm *avm, u32 pasid) - -{ - int ret; - - /* Free the original amdgpu allocated pasid, - * will be replaced with kfd allocated pasid. - */ - if (avm->pasid) { - amdgpu_pasid_free(avm->pasid); - amdgpu_vm_set_pasid(adev, avm, 0); - } - - ret = amdgpu_vm_set_pasid(adev, avm, pasid); - if (ret) - return ret; - - return 0; -} - int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev, struct amdgpu_vm *avm, void **process_info, @@ -1607,27 +1586,6 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev, } } -void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev, - void *drm_priv) -{ - struct amdgpu_vm *avm; - - if (WARN_ON(!adev || !drm_priv)) - return; - - avm = drm_priv_to_vm(drm_priv); - - pr_debug("Releasing process vm %p\n", avm); - - /* The original pasid of amdgpu vm has already been - * released during making a amdgpu vm to a compute vm - * The current pasid is managed by kfd and will be - * released on kfd process destroy. Set amdgpu pasid - * to 0 to avoid duplicate release. - */ - amdgpu_vm_release_compute(adev, avm); -} - uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv) { struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv); @@ -1688,7 +1646,7 @@ size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev, - reserved_for_pt - reserved_for_ras; - if (adev->flags & AMD_IS_APU) { + if (adev->apu_prefer_gtt) { system_mem_available = no_system_mem_limit ? kfd_mem_limit.max_system_mem_limit : kfd_mem_limit.max_system_mem_limit - @@ -1736,7 +1694,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( if (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) { domain = alloc_domain = AMDGPU_GEM_DOMAIN_VRAM; - if (adev->flags & AMD_IS_APU) { + if (adev->apu_prefer_gtt) { domain = AMDGPU_GEM_DOMAIN_GTT; alloc_domain = AMDGPU_GEM_DOMAIN_GTT; alloc_flags = 0; @@ -1987,7 +1945,7 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( if (size) { if (!is_imported && (mem->bo->preferred_domains == AMDGPU_GEM_DOMAIN_VRAM || - ((adev->flags & AMD_IS_APU) && + (adev->apu_prefer_gtt && mem->bo->preferred_domains == AMDGPU_GEM_DOMAIN_GTT))) *size = bo_size; else @@ -2414,7 +2372,7 @@ static int import_obj_create(struct amdgpu_device *adev, (*mem)->bo = bo; (*mem)->va = va; (*mem)->domain = (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) && - !(adev->flags & AMD_IS_APU) ? + !adev->apu_prefer_gtt ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT; (*mem)->mapped_to_gpu_memory = 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c index 093141ad6ed0..e476e45b996a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c @@ -36,13 +36,6 @@ #include "atombios_encoders.h" #include "bif/bif_4_1_d.h" -static void amdgpu_atombios_lookup_i2c_gpio_quirks(struct amdgpu_device *adev, - ATOM_GPIO_I2C_ASSIGMENT *gpio, - u8 index) -{ - -} - static struct amdgpu_i2c_bus_rec amdgpu_atombios_get_bus_rec_for_i2c_gpio(ATOM_GPIO_I2C_ASSIGMENT *gpio) { struct amdgpu_i2c_bus_rec i2c; @@ -108,9 +101,6 @@ struct amdgpu_i2c_bus_rec amdgpu_atombios_lookup_i2c_gpio(struct amdgpu_device * gpio = &i2c_info->asGPIO_Info[0]; for (i = 0; i < num_indices; i++) { - - amdgpu_atombios_lookup_i2c_gpio_quirks(adev, gpio, i); - if (gpio->sucI2cId.ucAccess == id) { i2c = amdgpu_atombios_get_bus_rec_for_i2c_gpio(gpio); break; @@ -142,8 +132,6 @@ void amdgpu_atombios_i2c_init(struct amdgpu_device *adev) gpio = &i2c_info->asGPIO_Info[0]; for (i = 0; i < num_indices; i++) { - amdgpu_atombios_lookup_i2c_gpio_quirks(adev, gpio, i); - i2c = amdgpu_atombios_get_bus_rec_for_i2c_gpio(gpio); if (i2c.valid) { @@ -156,6 +144,38 @@ void amdgpu_atombios_i2c_init(struct amdgpu_device *adev) } } +void amdgpu_atombios_oem_i2c_init(struct amdgpu_device *adev, u8 i2c_id) +{ + struct atom_context *ctx = adev->mode_info.atom_context; + ATOM_GPIO_I2C_ASSIGMENT *gpio; + struct amdgpu_i2c_bus_rec i2c; + int index = GetIndexIntoMasterTable(DATA, GPIO_I2C_Info); + struct _ATOM_GPIO_I2C_INFO *i2c_info; + uint16_t data_offset, size; + int i, num_indices; + char stmp[32]; + + if (amdgpu_atom_parse_data_header(ctx, index, &size, NULL, NULL, &data_offset)) { + i2c_info = (struct _ATOM_GPIO_I2C_INFO *)(ctx->bios + data_offset); + + num_indices = (size - sizeof(ATOM_COMMON_TABLE_HEADER)) / + sizeof(ATOM_GPIO_I2C_ASSIGMENT); + + gpio = &i2c_info->asGPIO_Info[0]; + for (i = 0; i < num_indices; i++) { + i2c = amdgpu_atombios_get_bus_rec_for_i2c_gpio(gpio); + + if (i2c.valid && i2c.i2c_id == i2c_id) { + sprintf(stmp, "OEM 0x%x", i2c.i2c_id); + adev->i2c_bus[i] = amdgpu_i2c_create(adev_to_drm(adev), &i2c, stmp); + break; + } + gpio = (ATOM_GPIO_I2C_ASSIGMENT *) + ((u8 *)gpio + sizeof(ATOM_GPIO_I2C_ASSIGMENT)); + } + } +} + struct amdgpu_gpio_rec amdgpu_atombios_lookup_gpio(struct amdgpu_device *adev, u8 id) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h index 0e16432d9a72..867bc5c5ce67 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h @@ -136,6 +136,7 @@ amdgpu_atombios_lookup_gpio(struct amdgpu_device *adev, struct amdgpu_i2c_bus_rec amdgpu_atombios_lookup_i2c_gpio(struct amdgpu_device *adev, uint8_t id); void amdgpu_atombios_i2c_init(struct amdgpu_device *adev); +void amdgpu_atombios_oem_i2c_init(struct amdgpu_device *adev, u8 i2c_id); bool amdgpu_atombios_has_dce_engine_info(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c index f873dd3cae16..eb015bdda8a7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c @@ -549,9 +549,10 @@ bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev) u16 data_offset, size; union umc_info *umc_info; u8 frev, crev; - bool ecc_default_enabled = false; + bool mem_ecc_enabled = false; u8 umc_config; u32 umc_config1; + adev->ras_default_ecc_enabled = false; index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1, umc_info); @@ -563,20 +564,22 @@ bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev) switch (crev) { case 1: umc_config = le32_to_cpu(umc_info->v31.umc_config); - ecc_default_enabled = + mem_ecc_enabled = (umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false; break; case 2: umc_config = le32_to_cpu(umc_info->v32.umc_config); - ecc_default_enabled = + mem_ecc_enabled = (umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false; break; case 3: umc_config = le32_to_cpu(umc_info->v33.umc_config); umc_config1 = le32_to_cpu(umc_info->v33.umc_config1); - ecc_default_enabled = + mem_ecc_enabled = ((umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) || (umc_config1 & UMC_CONFIG1__ENABLE_ECC_CAPABLE)) ? true : false; + adev->ras_default_ecc_enabled = + (umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false; break; default: /* unsupported crev */ @@ -585,9 +588,12 @@ bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev) } else if (frev == 4) { switch (crev) { case 0: + umc_config = le32_to_cpu(umc_info->v40.umc_config); umc_config1 = le32_to_cpu(umc_info->v40.umc_config1); - ecc_default_enabled = + mem_ecc_enabled = (umc_config1 & UMC_CONFIG1__ENABLE_ECC_CAPABLE) ? true : false; + adev->ras_default_ecc_enabled = + (umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false; break; default: /* unsupported crev */ @@ -599,7 +605,7 @@ bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev) } } - return ecc_default_enabled; + return mem_ecc_enabled; } /* diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c index 423fd2eebe1e..75fcc521c171 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c @@ -84,6 +84,13 @@ static bool check_atom_bios(struct amdgpu_device *adev, size_t size) return false; } +void amdgpu_bios_release(struct amdgpu_device *adev) +{ + kfree(adev->bios); + adev->bios = NULL; + adev->bios_size = 0; +} + /* If you boot an IGP board with a discrete card as the primary, * the IGP rom is not accessible via the rom bar as the IGP rom is * part of the system bios. On boot, the system bios puts a @@ -121,7 +128,7 @@ static bool amdgpu_read_bios_from_vram(struct amdgpu_device *adev) iounmap(bios); if (!check_atom_bios(adev, size)) { - kfree(adev->bios); + amdgpu_bios_release(adev); return false; } @@ -149,7 +156,7 @@ bool amdgpu_read_bios(struct amdgpu_device *adev) pci_unmap_rom(adev->pdev, bios); if (!check_atom_bios(adev, size)) { - kfree(adev->bios); + amdgpu_bios_release(adev); return false; } @@ -189,7 +196,7 @@ static bool amdgpu_read_bios_from_rom(struct amdgpu_device *adev) amdgpu_asic_read_bios_from_rom(adev, adev->bios, len); if (!check_atom_bios(adev, len)) { - kfree(adev->bios); + amdgpu_bios_release(adev); return false; } @@ -225,7 +232,8 @@ static bool amdgpu_read_platform_bios(struct amdgpu_device *adev) return true; free_bios: - kfree(adev->bios); + amdgpu_bios_release(adev); + return false; } @@ -327,7 +335,7 @@ static bool amdgpu_atrm_get_bios(struct amdgpu_device *adev) } if (!check_atom_bios(adev, size)) { - kfree(adev->bios); + amdgpu_bios_release(adev); return false; } adev->bios_size = size; @@ -392,7 +400,7 @@ static bool amdgpu_acpi_vfct_bios(struct amdgpu_device *adev) GFP_KERNEL); if (!check_atom_bios(adev, vhdr->ImageLength)) { - kfree(adev->bios); + amdgpu_bios_release(adev); return false; } adev->bios_size = vhdr->ImageLength; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c index 344e0a9ee08a..5e375e9c4f5d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c @@ -674,7 +674,7 @@ static int amdgpu_connector_lvds_get_modes(struct drm_connector *connector) } static enum drm_mode_status amdgpu_connector_lvds_mode_valid(struct drm_connector *connector, - struct drm_display_mode *mode) + const struct drm_display_mode *mode) { struct drm_encoder *encoder = amdgpu_connector_best_single_encoder(connector); @@ -839,7 +839,7 @@ static int amdgpu_connector_vga_get_modes(struct drm_connector *connector) } static enum drm_mode_status amdgpu_connector_vga_mode_valid(struct drm_connector *connector, - struct drm_display_mode *mode) + const struct drm_display_mode *mode) { struct drm_device *dev = connector->dev; struct amdgpu_device *adev = drm_to_adev(dev); @@ -1196,7 +1196,7 @@ static void amdgpu_connector_dvi_force(struct drm_connector *connector) } static enum drm_mode_status amdgpu_connector_dvi_mode_valid(struct drm_connector *connector, - struct drm_display_mode *mode) + const struct drm_display_mode *mode) { struct drm_device *dev = connector->dev; struct amdgpu_device *adev = drm_to_adev(dev); @@ -1464,7 +1464,7 @@ out: } static enum drm_mode_status amdgpu_connector_dp_mode_valid(struct drm_connector *connector, - struct drm_display_mode *mode) + const struct drm_display_mode *mode) { struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector); struct amdgpu_connector_atom_dig *amdgpu_dig_connector = amdgpu_connector->con_priv; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c new file mode 100644 index 000000000000..20c474a32852 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c @@ -0,0 +1,564 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright 2025 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/list.h> +#include "amdgpu.h" + +static const guid_t MCE = CPER_NOTIFY_MCE; +static const guid_t CMC = CPER_NOTIFY_CMC; +static const guid_t BOOT = BOOT_TYPE; + +static const guid_t CRASHDUMP = AMD_CRASHDUMP; +static const guid_t RUNTIME = AMD_GPU_NONSTANDARD_ERROR; + +static void __inc_entry_length(struct cper_hdr *hdr, uint32_t size) +{ + hdr->record_length += size; +} + +static void amdgpu_cper_get_timestamp(struct cper_timestamp *timestamp) +{ + struct tm tm; + time64_t now = ktime_get_real_seconds(); + + time64_to_tm(now, 0, &tm); + timestamp->seconds = tm.tm_sec; + timestamp->minutes = tm.tm_min; + timestamp->hours = tm.tm_hour; + timestamp->flag = 0; + timestamp->day = tm.tm_mday; + timestamp->month = 1 + tm.tm_mon; + timestamp->year = (1900 + tm.tm_year) % 100; + timestamp->century = (1900 + tm.tm_year) / 100; +} + +void amdgpu_cper_entry_fill_hdr(struct amdgpu_device *adev, + struct cper_hdr *hdr, + enum amdgpu_cper_type type, + enum cper_error_severity sev) +{ + hdr->signature[0] = 'C'; + hdr->signature[1] = 'P'; + hdr->signature[2] = 'E'; + hdr->signature[3] = 'R'; + hdr->revision = CPER_HDR_REV_1; + hdr->signature_end = 0xFFFFFFFF; + hdr->error_severity = sev; + + hdr->valid_bits.platform_id = 1; + hdr->valid_bits.partition_id = 1; + hdr->valid_bits.timestamp = 1; + + amdgpu_cper_get_timestamp(&hdr->timestamp); + + snprintf(hdr->record_id, 8, "%d", atomic_inc_return(&adev->cper.unique_id)); + snprintf(hdr->platform_id, 16, "0x%04X:0x%04X", + adev->pdev->vendor, adev->pdev->device); + /* pmfw version should be part of creator_id according to CPER spec */ + snprintf(hdr->creator_id, 16, "%s", CPER_CREATOR_ID_AMDGPU); + + switch (type) { + case AMDGPU_CPER_TYPE_BOOT: + hdr->notify_type = BOOT; + break; + case AMDGPU_CPER_TYPE_FATAL: + case AMDGPU_CPER_TYPE_BP_THRESHOLD: + hdr->notify_type = MCE; + break; + case AMDGPU_CPER_TYPE_RUNTIME: + if (sev == CPER_SEV_NON_FATAL_CORRECTED) + hdr->notify_type = CMC; + else + hdr->notify_type = MCE; + break; + default: + dev_err(adev->dev, "Unknown CPER Type\n"); + break; + } + + __inc_entry_length(hdr, HDR_LEN); +} + +static int amdgpu_cper_entry_fill_section_desc(struct amdgpu_device *adev, + struct cper_sec_desc *section_desc, + bool bp_threshold, + bool poison, + enum cper_error_severity sev, + guid_t sec_type, + uint32_t section_length, + uint32_t section_offset) +{ + section_desc->revision_minor = CPER_SEC_MINOR_REV_1; + section_desc->revision_major = CPER_SEC_MAJOR_REV_22; + section_desc->sec_offset = section_offset; + section_desc->sec_length = section_length; + section_desc->valid_bits.fru_id = 1; + section_desc->valid_bits.fru_text = 1; + section_desc->flag_bits.primary = 1; + section_desc->severity = sev; + section_desc->sec_type = sec_type; + + if (adev->smuio.funcs && + adev->smuio.funcs->get_socket_id) + snprintf(section_desc->fru_text, 20, "OAM%d", + adev->smuio.funcs->get_socket_id(adev)); + /* TODO: fru_id is 16 bytes in CPER spec, but driver defines it as 20 bytes */ + snprintf(section_desc->fru_id, 16, "%llx", adev->unique_id); + + if (bp_threshold) + section_desc->flag_bits.exceed_err_threshold = 1; + if (poison) + section_desc->flag_bits.latent_err = 1; + + return 0; +} + +int amdgpu_cper_entry_fill_fatal_section(struct amdgpu_device *adev, + struct cper_hdr *hdr, + uint32_t idx, + struct cper_sec_crashdump_reg_data reg_data) +{ + struct cper_sec_desc *section_desc; + struct cper_sec_crashdump_fatal *section; + + section_desc = (struct cper_sec_desc *)((uint8_t *)hdr + SEC_DESC_OFFSET(idx)); + section = (struct cper_sec_crashdump_fatal *)((uint8_t *)hdr + + FATAL_SEC_OFFSET(hdr->sec_cnt, idx)); + + amdgpu_cper_entry_fill_section_desc(adev, section_desc, false, false, + CPER_SEV_FATAL, CRASHDUMP, FATAL_SEC_LEN, + FATAL_SEC_OFFSET(hdr->sec_cnt, idx)); + + section->body.reg_ctx_type = CPER_CTX_TYPE_CRASH; + section->body.reg_arr_size = sizeof(reg_data); + section->body.data = reg_data; + + __inc_entry_length(hdr, SEC_DESC_LEN + FATAL_SEC_LEN); + + return 0; +} + +int amdgpu_cper_entry_fill_runtime_section(struct amdgpu_device *adev, + struct cper_hdr *hdr, + uint32_t idx, + enum cper_error_severity sev, + uint32_t *reg_dump, + uint32_t reg_count) +{ + struct cper_sec_desc *section_desc; + struct cper_sec_nonstd_err *section; + bool poison; + + poison = (sev == CPER_SEV_NON_FATAL_CORRECTED) ? false : true; + section_desc = (struct cper_sec_desc *)((uint8_t *)hdr + SEC_DESC_OFFSET(idx)); + section = (struct cper_sec_nonstd_err *)((uint8_t *)hdr + + NONSTD_SEC_OFFSET(hdr->sec_cnt, idx)); + + amdgpu_cper_entry_fill_section_desc(adev, section_desc, false, poison, + sev, RUNTIME, NONSTD_SEC_LEN, + NONSTD_SEC_OFFSET(hdr->sec_cnt, idx)); + + reg_count = umin(reg_count, CPER_ACA_REG_COUNT); + + section->hdr.valid_bits.err_info_cnt = 1; + section->hdr.valid_bits.err_context_cnt = 1; + + section->info.error_type = RUNTIME; + section->info.ms_chk_bits.err_type_valid = 1; + section->ctx.reg_ctx_type = CPER_CTX_TYPE_CRASH; + section->ctx.reg_arr_size = sizeof(section->ctx.reg_dump); + + memcpy(section->ctx.reg_dump, reg_dump, reg_count * sizeof(uint32_t)); + + __inc_entry_length(hdr, SEC_DESC_LEN + NONSTD_SEC_LEN); + + return 0; +} + +int amdgpu_cper_entry_fill_bad_page_threshold_section(struct amdgpu_device *adev, + struct cper_hdr *hdr, + uint32_t idx) +{ + struct cper_sec_desc *section_desc; + struct cper_sec_nonstd_err *section; + + section_desc = (struct cper_sec_desc *)((uint8_t *)hdr + SEC_DESC_OFFSET(idx)); + section = (struct cper_sec_nonstd_err *)((uint8_t *)hdr + + NONSTD_SEC_OFFSET(hdr->sec_cnt, idx)); + + amdgpu_cper_entry_fill_section_desc(adev, section_desc, true, false, + CPER_SEV_NUM, RUNTIME, NONSTD_SEC_LEN, + NONSTD_SEC_OFFSET(hdr->sec_cnt, idx)); + + section->hdr.valid_bits.err_info_cnt = 1; + section->hdr.valid_bits.err_context_cnt = 1; + + section->info.error_type = RUNTIME; + section->info.ms_chk_bits.err_type_valid = 1; + section->ctx.reg_ctx_type = CPER_CTX_TYPE_CRASH; + section->ctx.reg_arr_size = sizeof(section->ctx.reg_dump); + + /* Hardcoded Reg dump for bad page threshold CPER */ + section->ctx.reg_dump[CPER_ACA_REG_CTL_LO] = 0x1; + section->ctx.reg_dump[CPER_ACA_REG_CTL_HI] = 0x0; + section->ctx.reg_dump[CPER_ACA_REG_STATUS_LO] = 0x137; + section->ctx.reg_dump[CPER_ACA_REG_STATUS_HI] = 0xB0000000; + section->ctx.reg_dump[CPER_ACA_REG_ADDR_LO] = 0x0; + section->ctx.reg_dump[CPER_ACA_REG_ADDR_HI] = 0x0; + section->ctx.reg_dump[CPER_ACA_REG_MISC0_LO] = 0x0; + section->ctx.reg_dump[CPER_ACA_REG_MISC0_HI] = 0x0; + section->ctx.reg_dump[CPER_ACA_REG_CONFIG_LO] = 0x2; + section->ctx.reg_dump[CPER_ACA_REG_CONFIG_HI] = 0x1ff; + section->ctx.reg_dump[CPER_ACA_REG_IPID_LO] = 0x0; + section->ctx.reg_dump[CPER_ACA_REG_IPID_HI] = 0x96; + section->ctx.reg_dump[CPER_ACA_REG_SYND_LO] = 0x0; + section->ctx.reg_dump[CPER_ACA_REG_SYND_HI] = 0x0; + + __inc_entry_length(hdr, SEC_DESC_LEN + NONSTD_SEC_LEN); + + return 0; +} + +struct cper_hdr *amdgpu_cper_alloc_entry(struct amdgpu_device *adev, + enum amdgpu_cper_type type, + uint16_t section_count) +{ + struct cper_hdr *hdr; + uint32_t size = 0; + + size += HDR_LEN; + size += (SEC_DESC_LEN * section_count); + + switch (type) { + case AMDGPU_CPER_TYPE_RUNTIME: + case AMDGPU_CPER_TYPE_BP_THRESHOLD: + size += (NONSTD_SEC_LEN * section_count); + break; + case AMDGPU_CPER_TYPE_FATAL: + size += (FATAL_SEC_LEN * section_count); + break; + case AMDGPU_CPER_TYPE_BOOT: + size += (BOOT_SEC_LEN * section_count); + break; + default: + dev_err(adev->dev, "Unknown CPER Type!\n"); + return NULL; + } + + hdr = kzalloc(size, GFP_KERNEL); + if (!hdr) + return NULL; + + /* Save this early */ + hdr->sec_cnt = section_count; + + return hdr; +} + +int amdgpu_cper_generate_ue_record(struct amdgpu_device *adev, + struct aca_bank *bank) +{ + struct cper_hdr *fatal = NULL; + struct cper_sec_crashdump_reg_data reg_data = { 0 }; + struct amdgpu_ring *ring = &adev->cper.ring_buf; + int ret; + + fatal = amdgpu_cper_alloc_entry(adev, AMDGPU_CPER_TYPE_FATAL, 1); + if (!fatal) { + dev_err(adev->dev, "fail to alloc cper entry for ue record\n"); + return -ENOMEM; + } + + reg_data.status_lo = lower_32_bits(bank->regs[ACA_REG_IDX_STATUS]); + reg_data.status_hi = upper_32_bits(bank->regs[ACA_REG_IDX_STATUS]); + reg_data.addr_lo = lower_32_bits(bank->regs[ACA_REG_IDX_ADDR]); + reg_data.addr_hi = upper_32_bits(bank->regs[ACA_REG_IDX_ADDR]); + reg_data.ipid_lo = lower_32_bits(bank->regs[ACA_REG_IDX_IPID]); + reg_data.ipid_hi = upper_32_bits(bank->regs[ACA_REG_IDX_IPID]); + reg_data.synd_lo = lower_32_bits(bank->regs[ACA_REG_IDX_SYND]); + reg_data.synd_hi = upper_32_bits(bank->regs[ACA_REG_IDX_SYND]); + + amdgpu_cper_entry_fill_hdr(adev, fatal, AMDGPU_CPER_TYPE_FATAL, CPER_SEV_FATAL); + ret = amdgpu_cper_entry_fill_fatal_section(adev, fatal, 0, reg_data); + if (ret) + return ret; + + amdgpu_cper_ring_write(ring, fatal, fatal->record_length); + + return 0; +} + +int amdgpu_cper_generate_bp_threshold_record(struct amdgpu_device *adev) +{ + struct cper_hdr *bp_threshold = NULL; + struct amdgpu_ring *ring = &adev->cper.ring_buf; + int ret; + + bp_threshold = amdgpu_cper_alloc_entry(adev, AMDGPU_CPER_TYPE_BP_THRESHOLD, 1); + if (!bp_threshold) { + dev_err(adev->dev, "fail to alloc cper entry for bad page threshold record\n"); + return -ENOMEM; + } + + amdgpu_cper_entry_fill_hdr(adev, bp_threshold, AMDGPU_CPER_TYPE_BP_THRESHOLD, CPER_SEV_NUM); + ret = amdgpu_cper_entry_fill_bad_page_threshold_section(adev, bp_threshold, 0); + if (ret) + return ret; + + amdgpu_cper_ring_write(ring, bp_threshold, bp_threshold->record_length); + + return 0; +} + +static enum cper_error_severity amdgpu_aca_err_type_to_cper_sev(struct amdgpu_device *adev, + enum aca_error_type aca_err_type) +{ + switch (aca_err_type) { + case ACA_ERROR_TYPE_UE: + return CPER_SEV_FATAL; + case ACA_ERROR_TYPE_CE: + return CPER_SEV_NON_FATAL_CORRECTED; + case ACA_ERROR_TYPE_DEFERRED: + return CPER_SEV_NON_FATAL_UNCORRECTED; + default: + dev_err(adev->dev, "Unknown ACA error type!\n"); + return CPER_SEV_FATAL; + } +} + +int amdgpu_cper_generate_ce_records(struct amdgpu_device *adev, + struct aca_banks *banks, + uint16_t bank_count) +{ + struct cper_hdr *corrected = NULL; + enum cper_error_severity sev = CPER_SEV_NON_FATAL_CORRECTED; + struct amdgpu_ring *ring = &adev->cper.ring_buf; + uint32_t reg_data[CPER_ACA_REG_COUNT] = { 0 }; + struct aca_bank_node *node; + struct aca_bank *bank; + uint32_t i = 0; + int ret; + + corrected = amdgpu_cper_alloc_entry(adev, AMDGPU_CPER_TYPE_RUNTIME, bank_count); + if (!corrected) { + dev_err(adev->dev, "fail to allocate cper entry for ce records\n"); + return -ENOMEM; + } + + /* Raise severity if any DE is detected in the ACA bank list */ + list_for_each_entry(node, &banks->list, node) { + bank = &node->bank; + if (bank->aca_err_type == ACA_ERROR_TYPE_DEFERRED) { + sev = CPER_SEV_NON_FATAL_UNCORRECTED; + break; + } + } + + amdgpu_cper_entry_fill_hdr(adev, corrected, AMDGPU_CPER_TYPE_RUNTIME, sev); + + /* Combine CE and UE in cper record */ + list_for_each_entry(node, &banks->list, node) { + bank = &node->bank; + reg_data[CPER_ACA_REG_CTL_LO] = lower_32_bits(bank->regs[ACA_REG_IDX_CTL]); + reg_data[CPER_ACA_REG_CTL_HI] = upper_32_bits(bank->regs[ACA_REG_IDX_CTL]); + reg_data[CPER_ACA_REG_STATUS_LO] = lower_32_bits(bank->regs[ACA_REG_IDX_STATUS]); + reg_data[CPER_ACA_REG_STATUS_HI] = upper_32_bits(bank->regs[ACA_REG_IDX_STATUS]); + reg_data[CPER_ACA_REG_ADDR_LO] = lower_32_bits(bank->regs[ACA_REG_IDX_ADDR]); + reg_data[CPER_ACA_REG_ADDR_HI] = upper_32_bits(bank->regs[ACA_REG_IDX_ADDR]); + reg_data[CPER_ACA_REG_MISC0_LO] = lower_32_bits(bank->regs[ACA_REG_IDX_MISC0]); + reg_data[CPER_ACA_REG_MISC0_HI] = upper_32_bits(bank->regs[ACA_REG_IDX_MISC0]); + reg_data[CPER_ACA_REG_CONFIG_LO] = lower_32_bits(bank->regs[ACA_REG_IDX_CONFIG]); + reg_data[CPER_ACA_REG_CONFIG_HI] = upper_32_bits(bank->regs[ACA_REG_IDX_CONFIG]); + reg_data[CPER_ACA_REG_IPID_LO] = lower_32_bits(bank->regs[ACA_REG_IDX_IPID]); + reg_data[CPER_ACA_REG_IPID_HI] = upper_32_bits(bank->regs[ACA_REG_IDX_IPID]); + reg_data[CPER_ACA_REG_SYND_LO] = lower_32_bits(bank->regs[ACA_REG_IDX_SYND]); + reg_data[CPER_ACA_REG_SYND_HI] = upper_32_bits(bank->regs[ACA_REG_IDX_SYND]); + + ret = amdgpu_cper_entry_fill_runtime_section(adev, corrected, i++, + amdgpu_aca_err_type_to_cper_sev(adev, bank->aca_err_type), + reg_data, CPER_ACA_REG_COUNT); + if (ret) + return ret; + } + + amdgpu_cper_ring_write(ring, corrected, corrected->record_length); + + return 0; +} + +static bool amdgpu_cper_is_hdr(struct amdgpu_ring *ring, u64 pos) +{ + struct cper_hdr *chdr; + + chdr = (struct cper_hdr *)&(ring->ring[pos]); + return strcmp(chdr->signature, "CPER") ? false : true; +} + +static u32 amdgpu_cper_ring_get_ent_sz(struct amdgpu_ring *ring, u64 pos) +{ + struct cper_hdr *chdr; + u64 p; + u32 chunk, rec_len = 0; + + chdr = (struct cper_hdr *)&(ring->ring[pos]); + chunk = ring->ring_size - (pos << 2); + + if (!strcmp(chdr->signature, "CPER")) { + rec_len = chdr->record_length; + goto calc; + } + + /* ring buffer is not full, no cper data after ring->wptr */ + if (ring->count_dw) + goto calc; + + for (p = pos + 1; p <= ring->buf_mask; p++) { + chdr = (struct cper_hdr *)&(ring->ring[p]); + if (!strcmp(chdr->signature, "CPER")) { + rec_len = (p - pos) << 2; + goto calc; + } + } + +calc: + if (!rec_len) + return chunk; + else + return umin(rec_len, chunk); +} + +void amdgpu_cper_ring_write(struct amdgpu_ring *ring, + void *src, int count) +{ + u64 pos, wptr_old, rptr = *ring->rptr_cpu_addr & ring->ptr_mask; + u32 chunk, ent_sz; + u8 *s = (u8 *)src; + + if (count >= ring->ring_size - 4) { + dev_err(ring->adev->dev, + "CPER data size(%d) is larger than ring size(%d)\n", + count, ring->ring_size - 4); + + return; + } + + wptr_old = ring->wptr; + + mutex_lock(&ring->adev->cper.ring_lock); + while (count) { + ent_sz = amdgpu_cper_ring_get_ent_sz(ring, ring->wptr); + chunk = umin(ent_sz, count); + + memcpy(&ring->ring[ring->wptr], s, chunk); + + ring->wptr += (chunk >> 2); + ring->wptr &= ring->ptr_mask; + count -= chunk; + s += chunk; + } + + /* the buffer is overflow, adjust rptr */ + if (((wptr_old < rptr) && (rptr <= ring->wptr)) || + ((ring->wptr < wptr_old) && (wptr_old < rptr)) || + ((rptr <= ring->wptr) && (ring->wptr < wptr_old))) { + pos = (ring->wptr + 1) & ring->ptr_mask; + + do { + ent_sz = amdgpu_cper_ring_get_ent_sz(ring, pos); + + rptr += (ent_sz >> 2); + rptr &= ring->ptr_mask; + *ring->rptr_cpu_addr = rptr; + + pos = rptr; + } while (!amdgpu_cper_is_hdr(ring, rptr)); + } + mutex_unlock(&ring->adev->cper.ring_lock); + + if (ring->count_dw >= (count >> 2)) + ring->count_dw -= (count >> 2); + else + ring->count_dw = 0; +} + +static u64 amdgpu_cper_ring_get_rptr(struct amdgpu_ring *ring) +{ + return *(ring->rptr_cpu_addr); +} + +static u64 amdgpu_cper_ring_get_wptr(struct amdgpu_ring *ring) +{ + return ring->wptr; +} + +static const struct amdgpu_ring_funcs cper_ring_funcs = { + .type = AMDGPU_RING_TYPE_CPER, + .align_mask = 0xff, + .support_64bit_ptrs = false, + .get_rptr = amdgpu_cper_ring_get_rptr, + .get_wptr = amdgpu_cper_ring_get_wptr, +}; + +static int amdgpu_cper_ring_init(struct amdgpu_device *adev) +{ + struct amdgpu_ring *ring = &(adev->cper.ring_buf); + + mutex_init(&adev->cper.ring_lock); + + ring->adev = NULL; + ring->ring_obj = NULL; + ring->use_doorbell = false; + ring->no_scheduler = true; + ring->funcs = &cper_ring_funcs; + + sprintf(ring->name, "cper"); + return amdgpu_ring_init(adev, ring, CPER_MAX_RING_SIZE, NULL, 0, + AMDGPU_RING_PRIO_DEFAULT, NULL); +} + +int amdgpu_cper_init(struct amdgpu_device *adev) +{ + if (!amdgpu_aca_is_enabled(adev)) + return 0; + + mutex_init(&adev->cper.cper_lock); + + adev->cper.enabled = true; + adev->cper.max_count = CPER_MAX_ALLOWED_COUNT; + + return amdgpu_cper_ring_init(adev); +} + +int amdgpu_cper_fini(struct amdgpu_device *adev) +{ + if (!amdgpu_aca_is_enabled(adev)) + return 0; + + adev->cper.enabled = false; + + amdgpu_ring_fini(&(adev->cper.ring_buf)); + adev->cper.count = 0; + adev->cper.wptr = 0; + + return 0; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h new file mode 100644 index 000000000000..bcb97d245673 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h @@ -0,0 +1,105 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright 2025 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. + * + */ + +#ifndef __AMDGPU_CPER_H__ +#define __AMDGPU_CPER_H__ + +#include "amd_cper.h" +#include "amdgpu_aca.h" + +#define CPER_MAX_ALLOWED_COUNT 0x1000 +#define CPER_MAX_RING_SIZE 0X100000 +#define HDR_LEN (sizeof(struct cper_hdr)) +#define SEC_DESC_LEN (sizeof(struct cper_sec_desc)) + +#define BOOT_SEC_LEN (sizeof(struct cper_sec_crashdump_boot)) +#define FATAL_SEC_LEN (sizeof(struct cper_sec_crashdump_fatal)) +#define NONSTD_SEC_LEN (sizeof(struct cper_sec_nonstd_err)) + +#define SEC_DESC_OFFSET(idx) (HDR_LEN + (SEC_DESC_LEN * idx)) + +#define BOOT_SEC_OFFSET(count, idx) (HDR_LEN + (SEC_DESC_LEN * count) + (BOOT_SEC_LEN * idx)) +#define FATAL_SEC_OFFSET(count, idx) (HDR_LEN + (SEC_DESC_LEN * count) + (FATAL_SEC_LEN * idx)) +#define NONSTD_SEC_OFFSET(count, idx) (HDR_LEN + (SEC_DESC_LEN * count) + (NONSTD_SEC_LEN * idx)) + +enum amdgpu_cper_type { + AMDGPU_CPER_TYPE_RUNTIME, + AMDGPU_CPER_TYPE_FATAL, + AMDGPU_CPER_TYPE_BOOT, + AMDGPU_CPER_TYPE_BP_THRESHOLD, +}; + +struct amdgpu_cper { + bool enabled; + + atomic_t unique_id; + struct mutex cper_lock; + + /* Lifetime CPERs generated */ + uint32_t count; + uint32_t max_count; + + uint32_t wptr; + + void *ring[CPER_MAX_ALLOWED_COUNT]; + struct amdgpu_ring ring_buf; + struct mutex ring_lock; +}; + +void amdgpu_cper_entry_fill_hdr(struct amdgpu_device *adev, + struct cper_hdr *hdr, + enum amdgpu_cper_type type, + enum cper_error_severity sev); +int amdgpu_cper_entry_fill_fatal_section(struct amdgpu_device *adev, + struct cper_hdr *hdr, + uint32_t idx, + struct cper_sec_crashdump_reg_data reg_data); +int amdgpu_cper_entry_fill_runtime_section(struct amdgpu_device *adev, + struct cper_hdr *hdr, + uint32_t idx, + enum cper_error_severity sev, + uint32_t *reg_dump, + uint32_t reg_count); +int amdgpu_cper_entry_fill_bad_page_threshold_section(struct amdgpu_device *adev, + struct cper_hdr *hdr, + uint32_t section_idx); + +struct cper_hdr *amdgpu_cper_alloc_entry(struct amdgpu_device *adev, + enum amdgpu_cper_type type, + uint16_t section_count); +/* UE must be encoded into separated cper entries, 1 UE 1 cper */ +int amdgpu_cper_generate_ue_record(struct amdgpu_device *adev, + struct aca_bank *bank); +/* CEs and DEs are combined into 1 cper entry */ +int amdgpu_cper_generate_ce_records(struct amdgpu_device *adev, + struct aca_banks *banks, + uint16_t bank_count); +/* Bad page threshold is encoded into separated cper entry */ +int amdgpu_cper_generate_bp_threshold_record(struct amdgpu_device *adev); +void amdgpu_cper_ring_write(struct amdgpu_ring *ring, + void *src, int count); +int amdgpu_cper_init(struct amdgpu_device *adev); +int amdgpu_cper_fini(struct amdgpu_device *adev); + +#endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index d100bb7a137c..17e5967bfa60 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -102,6 +102,9 @@ MODULE_FIRMWARE("amdgpu/navi12_gpu_info.bin"); #define AMDGPU_PCIE_INDEX_HI_FALLBACK (0x44 >> 2) #define AMDGPU_PCIE_DATA_FALLBACK (0x3C >> 2) +#define AMDGPU_VBIOS_SKIP (1U << 0) +#define AMDGPU_VBIOS_OPTIONAL (1U << 1) + static const struct drm_driver amdgpu_kms_driver; const char *amdgpu_asic_name[] = { @@ -1389,6 +1392,17 @@ static void amdgpu_block_invalid_wreg(struct amdgpu_device *adev, BUG(); } +static uint32_t amdgpu_device_get_vbios_flags(struct amdgpu_device *adev) +{ + if (hweight32(adev->aid_mask) && (adev->flags & AMD_IS_APU)) + return AMDGPU_VBIOS_SKIP; + + if (hweight32(adev->aid_mask) && amdgpu_passthrough(adev)) + return AMDGPU_VBIOS_OPTIONAL; + + return 0; +} + /** * amdgpu_device_asic_init - Wrapper for atom asic_init * @@ -1398,18 +1412,28 @@ static void amdgpu_block_invalid_wreg(struct amdgpu_device *adev, */ static int amdgpu_device_asic_init(struct amdgpu_device *adev) { + uint32_t flags; + bool optional; int ret; amdgpu_asic_pre_asic_init(adev); + flags = amdgpu_device_get_vbios_flags(adev); + optional = !!(flags & (AMDGPU_VBIOS_OPTIONAL | AMDGPU_VBIOS_SKIP)); if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0) || amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0)) { amdgpu_psp_wait_for_bootloader(adev); + if (optional && !adev->bios) + return 0; + ret = amdgpu_atomfirmware_asic_init(adev, true); return ret; } else { + if (optional && !adev->bios) + return 0; + return amdgpu_atom_asic_init(adev->mode_info.atom_context); } @@ -1698,14 +1722,6 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev) return 0; } -static bool amdgpu_device_read_bios(struct amdgpu_device *adev) -{ - if (hweight32(adev->aid_mask) && (adev->flags & AMD_IS_APU)) - return false; - - return true; -} - /* * GPU helpers function. */ @@ -1720,12 +1736,15 @@ static bool amdgpu_device_read_bios(struct amdgpu_device *adev) */ bool amdgpu_device_need_post(struct amdgpu_device *adev) { - uint32_t reg; + uint32_t reg, flags; if (amdgpu_sriov_vf(adev)) return false; - if (!amdgpu_device_read_bios(adev)) + flags = amdgpu_device_get_vbios_flags(adev); + if (flags & AMDGPU_VBIOS_SKIP) + return false; + if ((flags & AMDGPU_VBIOS_OPTIONAL) && !adev->bios) return false; if (amdgpu_passthrough(adev)) { @@ -2233,7 +2252,8 @@ void amdgpu_device_ip_get_clockgating_state(struct amdgpu_device *adev, if (!adev->ip_blocks[i].status.valid) continue; if (adev->ip_blocks[i].version->funcs->get_clockgating_state) - adev->ip_blocks[i].version->funcs->get_clockgating_state((void *)adev, flags); + adev->ip_blocks[i].version->funcs->get_clockgating_state( + &adev->ip_blocks[i], flags); } } @@ -2578,8 +2598,9 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) { struct amdgpu_ip_block *ip_block; struct pci_dev *parent; + bool total, skip_bios; + uint32_t bios_flags; int i, r; - bool total; amdgpu_device_enable_virtual_display(adev); @@ -2692,16 +2713,31 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) if (r) return r; + bios_flags = amdgpu_device_get_vbios_flags(adev); + skip_bios = !!(bios_flags & AMDGPU_VBIOS_SKIP); /* Read BIOS */ - if (amdgpu_device_read_bios(adev)) { - if (!amdgpu_get_bios(adev)) + if (!skip_bios) { + bool optional = + !!(bios_flags & AMDGPU_VBIOS_OPTIONAL); + if (!amdgpu_get_bios(adev) && !optional) return -EINVAL; - r = amdgpu_atombios_init(adev); - if (r) { - dev_err(adev->dev, "amdgpu_atombios_init failed\n"); - amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL, 0, 0); - return r; + if (optional && !adev->bios) + dev_info( + adev->dev, + "VBIOS image optional, proceeding without VBIOS image"); + + if (adev->bios) { + r = amdgpu_atombios_init(adev); + if (r) { + dev_err(adev->dev, + "amdgpu_atombios_init failed\n"); + amdgpu_vf_error_put( + adev, + AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL, + 0, 0); + return r; + } } } @@ -2823,6 +2859,12 @@ static int amdgpu_device_fw_loading(struct amdgpu_device *adev) static int amdgpu_device_init_schedulers(struct amdgpu_device *adev) { + struct drm_sched_init_args args = { + .ops = &amdgpu_sched_ops, + .num_rqs = DRM_SCHED_PRIORITY_COUNT, + .timeout_wq = adev->reset_domain->wq, + .dev = adev->dev, + }; long timeout; int r, i; @@ -2848,12 +2890,12 @@ static int amdgpu_device_init_schedulers(struct amdgpu_device *adev) break; } - r = drm_sched_init(&ring->sched, &amdgpu_sched_ops, NULL, - DRM_SCHED_PRIORITY_COUNT, - ring->num_hw_submission, 0, - timeout, adev->reset_domain->wq, - ring->sched_score, ring->name, - adev->dev); + args.timeout = timeout; + args.credit_limit = ring->num_hw_submission; + args.score = ring->sched_score; + args.name = ring->name; + + r = drm_sched_init(&ring->sched, &args); if (r) { DRM_ERROR("Failed to create scheduler on ring %s.\n", ring->name); @@ -3056,6 +3098,8 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) amdgpu_fru_get_product_info(adev); + r = amdgpu_cper_init(adev); + init_failed: return r; @@ -3416,6 +3460,8 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev) { int i, r; + amdgpu_cper_fini(adev); + if (amdgpu_sriov_vf(adev) && adev->virt.ras_init_done) amdgpu_virt_release_ras_err_handler_data(adev); @@ -4216,7 +4262,6 @@ int amdgpu_device_init(struct amdgpu_device *adev, mutex_init(&adev->grbm_idx_mutex); mutex_init(&adev->mn_lock); mutex_init(&adev->virt.vf_errors.lock); - mutex_init(&adev->virt.rlcg_reg_lock); hash_init(adev->mn_hash); mutex_init(&adev->psp.mutex); mutex_init(&adev->notifier_lock); @@ -4242,6 +4287,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, spin_lock_init(&adev->se_cac_idx_lock); spin_lock_init(&adev->audio_endpt_idx_lock); spin_lock_init(&adev->mm_stats.lock); + spin_lock_init(&adev->virt.rlcg_reg_lock); spin_lock_init(&adev->wb.lock); INIT_LIST_HEAD(&adev->reset_list); @@ -4465,8 +4511,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, goto failed; } /* init i2c buses */ - if (!amdgpu_device_has_dc_support(adev)) - amdgpu_atombios_i2c_init(adev); + amdgpu_i2c_init(adev); } } @@ -4735,14 +4780,13 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev) amdgpu_reset_fini(adev); /* free i2c buses */ - if (!amdgpu_device_has_dc_support(adev)) - amdgpu_i2c_fini(adev); - - if (amdgpu_emu_mode != 1) - amdgpu_atombios_fini(adev); + amdgpu_i2c_fini(adev); - kfree(adev->bios); - adev->bios = NULL; + if (adev->bios) { + if (amdgpu_emu_mode != 1) + amdgpu_atombios_fini(adev); + amdgpu_bios_release(adev); + } kfree(adev->fru_info); adev->fru_info = NULL; @@ -5368,7 +5412,8 @@ int amdgpu_device_mode1_reset(struct amdgpu_device *adev) u32 i; int ret = 0; - amdgpu_atombios_scratch_regs_engine_hung(adev, true); + if (adev->bios) + amdgpu_atombios_scratch_regs_engine_hung(adev, true); dev_info(adev->dev, "GPU mode1 reset\n"); @@ -5410,7 +5455,8 @@ int amdgpu_device_mode1_reset(struct amdgpu_device *adev) goto mode1_reset_failed; } - amdgpu_atombios_scratch_regs_engine_hung(adev, false); + if (adev->bios) + amdgpu_atombios_scratch_regs_engine_hung(adev, false); return 0; @@ -6116,6 +6162,10 @@ end_reset: dev_info(adev->dev, "GPU reset end with ret = %d\n", r); atomic_set(&adev->reset_domain->reset_res, r); + + if (!r) + drm_dev_wedged_event(adev_to_drm(adev), DRM_WEDGE_RECOVERY_NONE); + return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c index 949d74eff294..a4258127083d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c @@ -587,16 +587,19 @@ void amdgpu_discovery_fini(struct amdgpu_device *adev) adev->mman.discovery_bin = NULL; } -static int amdgpu_discovery_validate_ip(const struct ip_v4 *ip) +static int amdgpu_discovery_validate_ip(struct amdgpu_device *adev, + uint8_t instance, uint16_t hw_id) { - if (ip->instance_number >= HWIP_MAX_INSTANCE) { - DRM_ERROR("Unexpected instance_number (%d) from ip discovery blob\n", - ip->instance_number); + if (instance >= HWIP_MAX_INSTANCE) { + dev_err(adev->dev, + "Unexpected instance_number (%d) from ip discovery blob\n", + instance); return -EINVAL; } - if (le16_to_cpu(ip->hw_id) >= HW_ID_MAX) { - DRM_ERROR("Unexpected hw_id (%d) from ip discovery blob\n", - le16_to_cpu(ip->hw_id)); + if (hw_id >= HW_ID_MAX) { + dev_err(adev->dev, + "Unexpected hw_id (%d) from ip discovery blob\n", + hw_id); return -EINVAL; } @@ -609,8 +612,10 @@ static void amdgpu_discovery_read_harvest_bit_per_ip(struct amdgpu_device *adev, struct binary_header *bhdr; struct ip_discovery_header *ihdr; struct die_header *dhdr; - struct ip_v4 *ip; + struct ip *ip; uint16_t die_offset, ip_offset, num_dies, num_ips; + uint16_t hw_id; + uint8_t inst; int i, j; bhdr = (struct binary_header *)adev->mman.discovery_bin; @@ -626,16 +631,18 @@ static void amdgpu_discovery_read_harvest_bit_per_ip(struct amdgpu_device *adev, ip_offset = die_offset + sizeof(*dhdr); for (j = 0; j < num_ips; j++) { - ip = (struct ip_v4 *)(adev->mman.discovery_bin + ip_offset); - - if (amdgpu_discovery_validate_ip(ip)) + ip = (struct ip *)(adev->mman.discovery_bin + + ip_offset); + inst = ip->number_instance; + hw_id = le16_to_cpu(ip->hw_id); + if (amdgpu_discovery_validate_ip(adev, inst, hw_id)) goto next_ip; - if (le16_to_cpu(ip->variant) == 1) { - switch (le16_to_cpu(ip->hw_id)) { + if (ip->harvest == 1) { + switch (hw_id) { case VCN_HWID: (*vcn_harvest_count)++; - if (ip->instance_number == 0) { + if (inst == 0) { adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0; adev->vcn.inst_mask &= ~AMDGPU_VCN_HARVEST_VCN0; @@ -657,10 +664,8 @@ static void amdgpu_discovery_read_harvest_bit_per_ip(struct amdgpu_device *adev, } } next_ip: - if (ihdr->base_addr_64_bit) - ip_offset += struct_size(ip, base_address_64, ip->num_base_address); - else - ip_offset += struct_size(ip, base_address, ip->num_base_address); + ip_offset += struct_size(ip, base_address, + ip->num_base_address); } } } @@ -1019,6 +1024,8 @@ static int amdgpu_discovery_sysfs_ips(struct amdgpu_device *adev, bool reg_base_64) { int ii, jj, kk, res; + uint16_t hw_id; + uint8_t inst; DRM_DEBUG("num_ips:%d", num_ips); @@ -1034,8 +1041,10 @@ static int amdgpu_discovery_sysfs_ips(struct amdgpu_device *adev, struct ip_hw_instance *ip_hw_instance; ip = (struct ip_v4 *)(adev->mman.discovery_bin + ip_offset); - if (amdgpu_discovery_validate_ip(ip) || - le16_to_cpu(ip->hw_id) != ii) + inst = ip->instance_number; + hw_id = le16_to_cpu(ip->hw_id); + if (amdgpu_discovery_validate_ip(adev, inst, hw_id) || + hw_id != ii) goto next_ip; DRM_DEBUG("match:%d @ ip_offset:%zu", ii, ip_offset); @@ -1282,6 +1291,8 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) uint16_t ip_offset; uint16_t num_dies; uint16_t num_ips; + uint16_t hw_id; + uint8_t inst; int hw_ip; int i, j, k; int r; @@ -1321,7 +1332,9 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) for (j = 0; j < num_ips; j++) { ip = (struct ip_v4 *)(adev->mman.discovery_bin + ip_offset); - if (amdgpu_discovery_validate_ip(ip)) + inst = ip->instance_number; + hw_id = le16_to_cpu(ip->hw_id); + if (amdgpu_discovery_validate_ip(adev, inst, hw_id)) goto next_ip; num_base_address = ip->num_base_address; @@ -1460,17 +1473,24 @@ next_ip: static void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) { + struct ip_discovery_header *ihdr; + struct binary_header *bhdr; int vcn_harvest_count = 0; int umc_harvest_count = 0; + uint16_t offset, ihdr_ver; + bhdr = (struct binary_header *)adev->mman.discovery_bin; + offset = le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset); + ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin + + offset); + ihdr_ver = le16_to_cpu(ihdr->version); /* * Harvest table does not fit Navi1x and legacy GPUs, * so read harvest bit per IP data structure to set * harvest configuration. */ if (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(10, 2, 0) && - amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 3) && - amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 4)) { + ihdr_ver <= 2) { if ((adev->pdev->device == 0x731E && (adev->pdev->revision == 0xC6 || adev->pdev->revision == 0xC7)) || @@ -1864,6 +1884,7 @@ static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): amdgpu_device_ip_block_add(adev, &soc21_common_ip_block); break; case IP_VERSION(12, 0, 0): @@ -1919,6 +1940,7 @@ static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): amdgpu_device_ip_block_add(adev, &gmc_v11_0_ip_block); break; case IP_VERSION(12, 0, 0): @@ -1998,6 +2020,7 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 12): case IP_VERSION(11, 0, 13): case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); break; case IP_VERSION(11, 0, 8): @@ -2029,6 +2052,7 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev) break; case IP_VERSION(14, 0, 2): case IP_VERSION(14, 0, 3): + case IP_VERSION(14, 0, 5): amdgpu_device_ip_block_add(adev, &psp_v14_0_ip_block); break; default: @@ -2061,6 +2085,7 @@ static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 12): case IP_VERSION(11, 0, 13): case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); break; case IP_VERSION(12, 0, 0): @@ -2079,6 +2104,7 @@ static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(13, 0, 10): case IP_VERSION(13, 0, 11): case IP_VERSION(13, 0, 14): + case IP_VERSION(13, 0, 12): amdgpu_device_ip_block_add(adev, &smu_v13_0_ip_block); break; case IP_VERSION(14, 0, 0): @@ -2086,6 +2112,7 @@ static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(14, 0, 2): case IP_VERSION(14, 0, 3): case IP_VERSION(14, 0, 4): + case IP_VERSION(14, 0, 5): amdgpu_device_ip_block_add(adev, &smu_v14_0_ip_block); break; default: @@ -2137,6 +2164,7 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(3, 2, 1): case IP_VERSION(3, 5, 0): case IP_VERSION(3, 5, 1): + case IP_VERSION(3, 6, 0): case IP_VERSION(4, 1, 0): /* TODO: Fix IP version. DC code expects version 4.0.1 */ if (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(4, 1, 0)) @@ -2215,6 +2243,7 @@ static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): amdgpu_device_ip_block_add(adev, &gfx_v11_0_ip_block); break; case IP_VERSION(12, 0, 0): @@ -2270,6 +2299,7 @@ static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(6, 1, 0): case IP_VERSION(6, 1, 1): case IP_VERSION(6, 1, 2): + case IP_VERSION(6, 1, 3): amdgpu_device_ip_block_add(adev, &sdma_v6_0_ip_block); break; case IP_VERSION(7, 0, 0): @@ -2393,6 +2423,7 @@ static int amdgpu_discovery_set_mes_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): amdgpu_device_ip_block_add(adev, &mes_v11_0_ip_block); adev->enable_mes = true; adev->enable_mes_kiq = true; @@ -2708,6 +2739,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): adev->family = AMDGPU_FAMILY_GC_11_5_0; break; case IP_VERSION(12, 0, 0): @@ -2733,6 +2765,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): adev->flags |= AMD_IS_APU; break; default: @@ -2766,11 +2799,13 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg; break; case IP_VERSION(7, 9, 0): + case IP_VERSION(7, 9, 1): adev->nbio.funcs = &nbio_v7_9_funcs; adev->nbio.hdp_flush_reg = &nbio_v7_9_hdp_flush_reg; break; case IP_VERSION(7, 11, 0): case IP_VERSION(7, 11, 1): + case IP_VERSION(7, 11, 2): case IP_VERSION(7, 11, 3): adev->nbio.funcs = &nbio_v7_11_funcs; adev->nbio.hdp_flush_reg = &nbio_v7_11_hdp_flush_reg; @@ -2898,6 +2933,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 10): case IP_VERSION(11, 0, 11): case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): case IP_VERSION(13, 0, 1): case IP_VERSION(13, 0, 9): case IP_VERSION(13, 0, 10): @@ -2907,6 +2943,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->smuio.funcs = &smuio_v13_0_funcs; break; case IP_VERSION(13, 0, 3): + case IP_VERSION(13, 0, 11): adev->smuio.funcs = &smuio_v13_0_3_funcs; if (adev->smuio.funcs->get_pkg_type(adev) == AMDGPU_PKG_TYPE_APU) { adev->flags |= AMD_IS_APU; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 817116e53d44..1819166cb4cf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -119,9 +119,12 @@ * - 3.57.0 - Compute tunneling on GFX10+ * - 3.58.0 - Add GFX12 DCC support * - 3.59.0 - Cleared VRAM + * - 3.60.0 - Add AMDGPU_TILING_GFX12_DCC_WRITE_COMPRESS_DISABLE (Vulkan requirement) + * - 3.61.0 - Contains fix for RV/PCO compute queues + * - 3.62.0 - Add AMDGPU_IDS_FLAGS_MODE_PF, AMDGPU_IDS_FLAGS_MODE_VF & AMDGPU_IDS_FLAGS_MODE_PT */ #define KMS_DRIVER_MAJOR 3 -#define KMS_DRIVER_MINOR 59 +#define KMS_DRIVER_MINOR 62 #define KMS_DRIVER_PATCHLEVEL 0 /* @@ -281,6 +284,7 @@ module_param_named(gartsize, amdgpu_gart_size, uint, 0600); * DOC: gttsize (int) * Restrict the size of GTT domain (for userspace use) in MiB for testing. * The default is -1 (Use value specified by TTM). + * This parameter is deprecated and will be removed in the future. */ MODULE_PARM_DESC(gttsize, "Size of the GTT userspace domain in megabytes (-1 = auto)"); module_param_named(gttsize, amdgpu_gtt_size, int, 0600); @@ -961,7 +965,7 @@ module_param_named_unsafe(reset_method, amdgpu_reset_method, int, 0644); * result in the GPU entering bad status when the number of total * faulty pages by ECC exceeds the threshold value. */ -MODULE_PARM_DESC(bad_page_threshold, "Bad page threshold(-1 = ignore threshold (default value), 0 = disable bad page retirement, -2 = driver sets threshold)"); +MODULE_PARM_DESC(bad_page_threshold, "Bad page threshold(-1 = ignore threshold (default value), 0 = disable bad page retirement, -2 = threshold determined by a formula, 0 < threshold < max records, user-defined threshold)"); module_param_named(bad_page_threshold, amdgpu_bad_page_threshold, int, 0444); MODULE_PARM_DESC(num_kcq, "number of kernel compute queue user want to setup (8 if set to greater than 8 or less than 0, only affect gfx 8+)"); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c index 09c9194d5bd5..89109eb2ce16 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c @@ -63,10 +63,10 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev, u32 *fru_addr) switch (adev->asic_type) { case CHIP_VEGA20: /* D161 and D163 are the VG20 server SKUs */ - if (strnstr(atom_ctx->vbios_pn, "D161", - sizeof(atom_ctx->vbios_pn)) || - strnstr(atom_ctx->vbios_pn, "D163", - sizeof(atom_ctx->vbios_pn))) { + if (atom_ctx && (strnstr(atom_ctx->vbios_pn, "D161", + sizeof(atom_ctx->vbios_pn)) || + strnstr(atom_ctx->vbios_pn, "D163", + sizeof(atom_ctx->vbios_pn)))) { if (fru_addr) *fru_addr = FRU_EEPROM_MADDR_6; return true; @@ -78,8 +78,8 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev, u32 *fru_addr) return false; } case IP_VERSION(11, 0, 7): - if (strnstr(atom_ctx->vbios_pn, "D603", - sizeof(atom_ctx->vbios_pn))) { + if (atom_ctx && strnstr(atom_ctx->vbios_pn, "D603", + sizeof(atom_ctx->vbios_pn))) { if (strnstr(atom_ctx->vbios_pn, "D603GLXE", sizeof(atom_ctx->vbios_pn))) { return false; @@ -94,8 +94,8 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev, u32 *fru_addr) } case IP_VERSION(13, 0, 2): /* All Aldebaran SKUs have an FRU */ - if (!strnstr(atom_ctx->vbios_pn, "D673", - sizeof(atom_ctx->vbios_pn))) + if (atom_ctx && !strnstr(atom_ctx->vbios_pn, "D673", + sizeof(atom_ctx->vbios_pn))) if (fru_addr) *fru_addr = FRU_EEPROM_MADDR_6; return true; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c index 784b03abb3a4..b9bd6654f317 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c @@ -771,18 +771,8 @@ int amdgpu_gfx_enable_kgq(struct amdgpu_device *adev, int xcc_id) return r; } -/* amdgpu_gfx_off_ctrl - Handle gfx off feature enable/disable - * - * @adev: amdgpu_device pointer - * @bool enable true: enable gfx off feature, false: disable gfx off feature - * - * 1. gfx off feature will be enabled by gfx ip after gfx cg gp enabled. - * 2. other client can send request to disable gfx off feature, the request should be honored. - * 3. other client can cancel their request of disable gfx off feature - * 4. other client should not send request to enable gfx off feature before disable gfx off feature. - */ - -void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable) +static void amdgpu_gfx_do_off_ctrl(struct amdgpu_device *adev, bool enable, + bool no_delay) { unsigned long delay = GFX_OFF_DELAY_ENABLE; @@ -804,7 +794,7 @@ void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable) if (adev->gfx.gfx_off_req_count == 0 && !adev->gfx.gfx_off_state) { /* If going to s2idle, no need to wait */ - if (adev->in_s0ix) { + if (no_delay) { if (!amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, true, 0)) adev->gfx.gfx_off_state = true; @@ -836,6 +826,43 @@ unlock: mutex_unlock(&adev->gfx.gfx_off_mutex); } +/* amdgpu_gfx_off_ctrl - Handle gfx off feature enable/disable + * + * @adev: amdgpu_device pointer + * @bool enable true: enable gfx off feature, false: disable gfx off feature + * + * 1. gfx off feature will be enabled by gfx ip after gfx cg pg enabled. + * 2. other client can send request to disable gfx off feature, the request should be honored. + * 3. other client can cancel their request of disable gfx off feature + * 4. other client should not send request to enable gfx off feature before disable gfx off feature. + * + * gfx off allow will be delayed by GFX_OFF_DELAY_ENABLE ms. + */ +void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable) +{ + /* If going to s2idle, no need to wait */ + bool no_delay = adev->in_s0ix ? true : false; + + amdgpu_gfx_do_off_ctrl(adev, enable, no_delay); +} + +/* amdgpu_gfx_off_ctrl_immediate - Handle gfx off feature enable/disable + * + * @adev: amdgpu_device pointer + * @bool enable true: enable gfx off feature, false: disable gfx off feature + * + * 1. gfx off feature will be enabled by gfx ip after gfx cg pg enabled. + * 2. other client can send request to disable gfx off feature, the request should be honored. + * 3. other client can cancel their request of disable gfx off feature + * 4. other client should not send request to enable gfx off feature before disable gfx off feature. + * + * gfx off allow will be issued immediately. + */ +void amdgpu_gfx_off_ctrl_immediate(struct amdgpu_device *adev, bool enable) +{ + amdgpu_gfx_do_off_ctrl(adev, enable, true); +} + int amdgpu_set_gfx_off_residency(struct amdgpu_device *adev, bool value) { int r = 0; @@ -1643,11 +1670,13 @@ static ssize_t amdgpu_gfx_set_enforce_isolation(struct device *dev, if (adev->enforce_isolation[i] && !partition_values[i]) { /* Going from enabled to disabled */ amdgpu_vmid_free_reserved(adev, AMDGPU_GFXHUB(i)); - amdgpu_mes_set_enforce_isolation(adev, i, false); + if (adev->enable_mes && adev->gfx.enable_cleaner_shader) + amdgpu_mes_set_enforce_isolation(adev, i, false); } else if (!adev->enforce_isolation[i] && partition_values[i]) { /* Going from disabled to enabled */ amdgpu_vmid_alloc_reserved(adev, AMDGPU_GFXHUB(i)); - amdgpu_mes_set_enforce_isolation(adev, i, true); + if (adev->enable_mes && adev->gfx.enable_cleaner_shader) + amdgpu_mes_set_enforce_isolation(adev, i, true); } adev->enforce_isolation[i] = partition_values[i]; } @@ -2118,6 +2147,63 @@ void amdgpu_gfx_enforce_isolation_ring_end_use(struct amdgpu_ring *ring) amdgpu_gfx_kfd_sch_ctrl(adev, idx, true); } +void amdgpu_gfx_profile_idle_work_handler(struct work_struct *work) +{ + struct amdgpu_device *adev = + container_of(work, struct amdgpu_device, gfx.idle_work.work); + enum PP_SMC_POWER_PROFILE profile; + u32 i, fences = 0; + int r; + + if (adev->gfx.num_gfx_rings) + profile = PP_SMC_POWER_PROFILE_FULLSCREEN3D; + else + profile = PP_SMC_POWER_PROFILE_COMPUTE; + + for (i = 0; i < AMDGPU_MAX_GFX_RINGS; ++i) + fences += amdgpu_fence_count_emitted(&adev->gfx.gfx_ring[i]); + for (i = 0; i < (AMDGPU_MAX_COMPUTE_RINGS * AMDGPU_MAX_GC_INSTANCES); ++i) + fences += amdgpu_fence_count_emitted(&adev->gfx.compute_ring[i]); + if (!fences && !atomic_read(&adev->gfx.total_submission_cnt)) { + r = amdgpu_dpm_switch_power_profile(adev, profile, false); + if (r) + dev_warn(adev->dev, "(%d) failed to disable %s power profile mode\n", r, + profile == PP_SMC_POWER_PROFILE_FULLSCREEN3D ? + "fullscreen 3D" : "compute"); + } else { + schedule_delayed_work(&adev->gfx.idle_work, GFX_PROFILE_IDLE_TIMEOUT); + } +} + +void amdgpu_gfx_profile_ring_begin_use(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + enum PP_SMC_POWER_PROFILE profile; + int r; + + if (adev->gfx.num_gfx_rings) + profile = PP_SMC_POWER_PROFILE_FULLSCREEN3D; + else + profile = PP_SMC_POWER_PROFILE_COMPUTE; + + atomic_inc(&adev->gfx.total_submission_cnt); + + if (!cancel_delayed_work_sync(&adev->gfx.idle_work)) { + r = amdgpu_dpm_switch_power_profile(adev, profile, true); + if (r) + dev_warn(adev->dev, "(%d) failed to disable %s power profile mode\n", r, + profile == PP_SMC_POWER_PROFILE_FULLSCREEN3D ? + "fullscreen 3D" : "compute"); + } +} + +void amdgpu_gfx_profile_ring_end_use(struct amdgpu_ring *ring) +{ + atomic_dec(&ring->adev->gfx.total_submission_cnt); + + schedule_delayed_work(&ring->adev->gfx.idle_work, GFX_PROFILE_IDLE_TIMEOUT); +} + /* * debugfs for to enable/disable gfx job submission to specific core. */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h index 8b5bd63b5773..4b1675f79caa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h @@ -57,6 +57,9 @@ enum amdgpu_gfx_pipe_priority { #define AMDGPU_GFX_QUEUE_PRIORITY_MINIMUM 0 #define AMDGPU_GFX_QUEUE_PRIORITY_MAXIMUM 15 +/* 1 second timeout */ +#define GFX_PROFILE_IDLE_TIMEOUT msecs_to_jiffies(1000) + enum amdgpu_gfx_partition { AMDGPU_SPX_PARTITION_MODE = 0, AMDGPU_DPX_PARTITION_MODE = 1, @@ -476,6 +479,9 @@ struct amdgpu_gfx { bool kfd_sch_inactive[MAX_XCP]; unsigned long enforce_isolation_jiffies[MAX_XCP]; unsigned long enforce_isolation_time[MAX_XCP]; + + atomic_t total_submission_cnt; + struct delayed_work idle_work; }; struct amdgpu_gfx_ras_reg_entry { @@ -547,6 +553,7 @@ int amdgpu_gfx_me_queue_to_bit(struct amdgpu_device *adev, int me, bool amdgpu_gfx_is_me_queue_enabled(struct amdgpu_device *adev, int me, int pipe, int queue); void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable); +void amdgpu_gfx_off_ctrl_immediate(struct amdgpu_device *adev, bool enable); int amdgpu_get_gfx_off_status(struct amdgpu_device *adev, uint32_t *value); int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block); void amdgpu_gfx_ras_fini(struct amdgpu_device *adev); @@ -584,6 +591,11 @@ void amdgpu_gfx_cleaner_shader_init(struct amdgpu_device *adev, void amdgpu_gfx_enforce_isolation_handler(struct work_struct *work); void amdgpu_gfx_enforce_isolation_ring_begin_use(struct amdgpu_ring *ring); void amdgpu_gfx_enforce_isolation_ring_end_use(struct amdgpu_ring *ring); + +void amdgpu_gfx_profile_idle_work_handler(struct work_struct *work); +void amdgpu_gfx_profile_ring_begin_use(struct amdgpu_ring *ring); +void amdgpu_gfx_profile_ring_end_use(struct amdgpu_ring *ring); + void amdgpu_debugfs_gfx_sched_mask_init(struct amdgpu_device *adev); void amdgpu_debugfs_compute_sched_mask_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c index 1c19a65e6553..c6e5c50a3322 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c @@ -591,7 +591,8 @@ int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev) if (ring == &adev->mes.ring[0] || ring == &adev->mes.ring[1] || - ring == &adev->umsch_mm.ring) + ring == &adev->umsch_mm.ring || + ring == &adev->cper.ring_buf) continue; inv_eng = ffs(vm_inv_engs[vmhub]); @@ -851,6 +852,7 @@ void amdgpu_gmc_tmz_set(struct amdgpu_device *adev) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): /* Don't enable it by default yet. */ if (amdgpu_tmz < 1) { @@ -888,6 +890,7 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev) gc_ver == IP_VERSION(9, 4, 2) || gc_ver == IP_VERSION(9, 4, 3) || gc_ver == IP_VERSION(9, 4, 4) || + gc_ver == IP_VERSION(9, 5, 0) || gc_ver >= IP_VERSION(10, 3, 0)); if (!amdgpu_sriov_xnack_support(adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.c index f0765ccde668..8179d0814db9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.c @@ -225,6 +225,25 @@ void amdgpu_i2c_destroy(struct amdgpu_i2c_chan *i2c) kfree(i2c); } +void amdgpu_i2c_init(struct amdgpu_device *adev) +{ + if (!adev->is_atom_fw) { + if (!amdgpu_device_has_dc_support(adev)) { + amdgpu_atombios_i2c_init(adev); + } else { + switch (adev->asic_type) { + case CHIP_POLARIS10: + case CHIP_POLARIS11: + case CHIP_POLARIS12: + amdgpu_atombios_oem_i2c_init(adev, 0x97); + break; + default: + break; + } + } + } +} + /* remove all the buses */ void amdgpu_i2c_fini(struct amdgpu_device *adev) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.h index 21e3d1dad0a1..1d3d3806e0dd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.h @@ -28,6 +28,7 @@ struct amdgpu_i2c_chan *amdgpu_i2c_create(struct drm_device *dev, const struct amdgpu_i2c_bus_rec *rec, const char *name); void amdgpu_i2c_destroy(struct amdgpu_i2c_chan *i2c); +void amdgpu_i2c_init(struct amdgpu_device *adev); void amdgpu_i2c_fini(struct amdgpu_device *adev); struct amdgpu_i2c_chan * amdgpu_i2c_lookup(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c index b6d2eb049f54..dda29132dfb2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c @@ -33,6 +33,7 @@ #define JPEG_IDLE_TIMEOUT msecs_to_jiffies(1000) static void amdgpu_jpeg_idle_work_handler(struct work_struct *work); +static void amdgpu_jpeg_reg_dump_fini(struct amdgpu_device *adev); int amdgpu_jpeg_sw_init(struct amdgpu_device *adev) { @@ -85,6 +86,9 @@ int amdgpu_jpeg_sw_fini(struct amdgpu_device *adev) amdgpu_ring_fini(&adev->jpeg.inst[i].ring_dec[j]); } + if (adev->jpeg.reg_list) + amdgpu_jpeg_reg_dump_fini(adev); + mutex_destroy(&adev->jpeg.jpeg_pg_lock); return 0; @@ -452,3 +456,83 @@ void amdgpu_jpeg_sysfs_reset_mask_fini(struct amdgpu_device *adev) device_remove_file(adev->dev, &dev_attr_jpeg_reset_mask); } } + +int amdgpu_jpeg_reg_dump_init(struct amdgpu_device *adev, + const struct amdgpu_hwip_reg_entry *reg, u32 count) +{ + adev->jpeg.ip_dump = kcalloc(adev->jpeg.num_jpeg_inst * count, + sizeof(uint32_t), GFP_KERNEL); + if (!adev->jpeg.ip_dump) { + DRM_ERROR("Failed to allocate memory for JPEG IP Dump\n"); + return -ENOMEM; + } + adev->jpeg.reg_list = reg; + adev->jpeg.reg_count = count; + + return 0; +} + +static void amdgpu_jpeg_reg_dump_fini(struct amdgpu_device *adev) +{ + kfree(adev->jpeg.ip_dump); + adev->jpeg.reg_list = NULL; + adev->jpeg.reg_count = 0; +} + +void amdgpu_jpeg_dump_ip_state(struct amdgpu_ip_block *ip_block) +{ + struct amdgpu_device *adev = ip_block->adev; + u32 inst_off, inst_id, is_powered; + int i, j; + + if (!adev->jpeg.ip_dump) + return; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + inst_id = GET_INST(JPEG, i); + inst_off = i * adev->jpeg.reg_count; + /* check power status from UVD_JPEG_POWER_STATUS */ + adev->jpeg.ip_dump[inst_off] = + RREG32(SOC15_REG_ENTRY_OFFSET_INST(adev->jpeg.reg_list[0], + inst_id)); + is_powered = ((adev->jpeg.ip_dump[inst_off] & 0x1) != 1); + + if (is_powered) + for (j = 1; j < adev->jpeg.reg_count; j++) + adev->jpeg.ip_dump[inst_off + j] = + RREG32(SOC15_REG_ENTRY_OFFSET_INST(adev->jpeg.reg_list[j], + inst_id)); + } +} + +void amdgpu_jpeg_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p) +{ + struct amdgpu_device *adev = ip_block->adev; + u32 inst_off, is_powered; + int i, j; + + if (!adev->jpeg.ip_dump) + return; + + drm_printf(p, "num_instances:%d\n", adev->jpeg.num_jpeg_inst); + for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) { + if (adev->jpeg.harvest_config & (1 << i)) { + drm_printf(p, "\nHarvested Instance:JPEG%d Skipping dump\n", i); + continue; + } + + inst_off = i * adev->jpeg.reg_count; + is_powered = ((adev->jpeg.ip_dump[inst_off] & 0x1) != 1); + + if (is_powered) { + drm_printf(p, "Active Instance:JPEG%d\n", i); + for (j = 0; j < adev->jpeg.reg_count; j++) + drm_printf(p, "%-50s \t 0x%08x\n", adev->jpeg.reg_list[j].reg_name, + adev->jpeg.ip_dump[inst_off + j]); + } else + drm_printf(p, "\nInactive Instance:JPEG%d\n", i); + } +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h index d9cb343a8708..4f0775e39b54 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h @@ -92,6 +92,14 @@ *adev->jpeg.inst[inst_idx].dpg_sram_curr_addr++ = value; \ } while (0) +struct amdgpu_hwip_reg_entry; + +enum amdgpu_jpeg_caps { + AMDGPU_JPEG_RRMT_ENABLED, +}; + +#define AMDGPU_JPEG_CAPS(caps) BIT(AMDGPU_JPEG_##caps) + struct amdgpu_jpeg_reg{ unsigned jpeg_pitch[AMDGPU_MAX_JPEG_RINGS]; }; @@ -130,6 +138,10 @@ struct amdgpu_jpeg { uint8_t num_inst_per_aid; bool indirect_sram; uint32_t supported_reset; + uint32_t caps; + u32 *ip_dump; + u32 reg_count; + const struct amdgpu_hwip_reg_entry *reg_list; }; int amdgpu_jpeg_sw_init(struct amdgpu_device *adev); @@ -154,5 +166,9 @@ int amdgpu_jpeg_psp_update_sram(struct amdgpu_device *adev, int inst_idx, void amdgpu_debugfs_jpeg_sched_mask_init(struct amdgpu_device *adev); int amdgpu_jpeg_sysfs_reset_mask_init(struct amdgpu_device *adev); void amdgpu_jpeg_sysfs_reset_mask_fini(struct amdgpu_device *adev); +int amdgpu_jpeg_reg_dump_init(struct amdgpu_device *adev, + const struct amdgpu_hwip_reg_entry *reg, u32 count); +void amdgpu_jpeg_dump_ip_state(struct amdgpu_ip_block *ip_block); +void amdgpu_jpeg_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p); #endif /*__AMDGPU_JPEG_H__*/ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index 98528ee94c15..05c73bf7541c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -888,6 +888,15 @@ out: if (adev->gfx.config.ta_cntl2_truncate_coord_mode) dev_info->ids_flags |= AMDGPU_IDS_FLAGS_CONFORMANT_TRUNC_COORD; + if (amdgpu_passthrough(adev)) + dev_info->ids_flags |= (AMDGPU_IDS_FLAGS_MODE_PT << + AMDGPU_IDS_FLAGS_MODE_SHIFT) & + AMDGPU_IDS_FLAGS_MODE_MASK; + else if (amdgpu_sriov_vf(adev)) + dev_info->ids_flags |= (AMDGPU_IDS_FLAGS_MODE_VF << + AMDGPU_IDS_FLAGS_MODE_SHIFT) & + AMDGPU_IDS_FLAGS_MODE_MASK; + vm_size = adev->vm_manager.max_pfn * AMDGPU_GPU_PAGE_SIZE; vm_size -= AMDGPU_VA_RESERVED_TOP; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h index 5e3faefc5510..6da4f946cac0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h @@ -609,6 +609,7 @@ struct amdgpu_i2c_adapter { struct i2c_adapter base; struct ddc_service *ddc_service; + bool oem; }; #define TO_DM_AUX(x) container_of((x), struct amdgpu_dm_dp_aux, aux) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index babe94ade247..0f2eb69ad715 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -44,7 +44,7 @@ #include "amdgpu_securedisplay.h" #include "amdgpu_atomfirmware.h" -#define AMD_VBIOS_FILE_MAX_SIZE_B (1024*1024*3) +#define AMD_VBIOS_FILE_MAX_SIZE_B (1024*1024*16) static int psp_load_smu_fw(struct psp_context *psp); static int psp_rap_terminate(struct psp_context *psp); @@ -193,6 +193,7 @@ static int psp_early_init(struct amdgpu_ip_block *ip_block) case IP_VERSION(11, 0, 9): case IP_VERSION(11, 0, 11): case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): case IP_VERSION(11, 0, 12): case IP_VERSION(11, 0, 13): psp_v11_0_set_psp_funcs(psp); @@ -208,11 +209,15 @@ static int psp_early_init(struct amdgpu_ip_block *ip_block) psp->boot_time_tmr = false; fallthrough; case IP_VERSION(13, 0, 6): - case IP_VERSION(13, 0, 12): case IP_VERSION(13, 0, 14): psp_v13_0_set_psp_funcs(psp); psp->autoload_supported = false; break; + case IP_VERSION(13, 0, 12): + psp_v13_0_set_psp_funcs(psp); + psp->autoload_supported = false; + adev->psp.sup_ifwi_up = !amdgpu_sriov_vf(adev); + break; case IP_VERSION(13, 0, 1): case IP_VERSION(13, 0, 3): case IP_VERSION(13, 0, 5): @@ -246,6 +251,10 @@ static int psp_early_init(struct amdgpu_ip_block *ip_block) case IP_VERSION(14, 0, 3): psp_v14_0_set_psp_funcs(psp); break; + case IP_VERSION(14, 0, 5): + psp_v14_0_set_psp_funcs(psp); + psp->boot_time_tmr = false; + break; default: return -EINVAL; } @@ -533,7 +542,6 @@ static int psp_sw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; struct psp_context *psp = &adev->psp; - struct psp_gfx_cmd_resp *cmd = psp->cmd; psp_memory_training_fini(psp); @@ -543,8 +551,8 @@ static int psp_sw_fini(struct amdgpu_ip_block *ip_block) amdgpu_ucode_release(&psp->cap_fw); amdgpu_ucode_release(&psp->toc_fw); - kfree(cmd); - cmd = NULL; + kfree(psp->cmd); + psp->cmd = NULL; psp_free_shared_bufs(psp); @@ -1786,34 +1794,47 @@ int psp_ras_initialize(struct psp_context *psp) if (ret) dev_warn(adev->dev, "PSP get boot config failed\n"); - if (!amdgpu_ras_is_supported(psp->adev, AMDGPU_RAS_BLOCK__UMC)) { - if (!boot_cfg) { - dev_info(adev->dev, "GECC is disabled\n"); - } else { - /* disable GECC in next boot cycle if ras is - * disabled by module parameter amdgpu_ras_enable - * and/or amdgpu_ras_mask, or boot_config_get call - * is failed - */ - ret = psp_boot_config_set(adev, 0); - if (ret) - dev_warn(adev->dev, "PSP set boot config failed\n"); - else - dev_warn(adev->dev, "GECC will be disabled in next boot cycle if set amdgpu_ras_enable and/or amdgpu_ras_mask to 0x0\n"); - } + if (boot_cfg == 1 && !adev->ras_default_ecc_enabled && + amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC)) { + dev_warn(adev->dev, "GECC is currently enabled, which may affect performance\n"); + dev_warn(adev->dev, + "To disable GECC, please reboot the system and load the amdgpu driver with the parameter amdgpu_ras_enable=0\n"); } else { - if (boot_cfg == 1) { - dev_info(adev->dev, "GECC is enabled\n"); + if ((adev->ras_default_ecc_enabled || amdgpu_ras_enable == 1) && + amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC)) { + if (boot_cfg == 1) { + dev_info(adev->dev, "GECC is enabled\n"); + } else { + /* enable GECC in next boot cycle if it is disabled + * in boot config, or force enable GECC if failed to + * get boot configuration + */ + ret = psp_boot_config_set(adev, BOOT_CONFIG_GECC); + if (ret) + dev_warn(adev->dev, "PSP set boot config failed\n"); + else + dev_warn(adev->dev, "GECC will be enabled in next boot cycle\n"); + } } else { - /* enable GECC in next boot cycle if it is disabled - * in boot config, or force enable GECC if failed to - * get boot configuration - */ - ret = psp_boot_config_set(adev, BOOT_CONFIG_GECC); - if (ret) - dev_warn(adev->dev, "PSP set boot config failed\n"); - else - dev_warn(adev->dev, "GECC will be enabled in next boot cycle\n"); + if (!boot_cfg) { + if (!adev->ras_default_ecc_enabled && + amdgpu_ras_enable != 1 && + amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC)) + dev_warn(adev->dev, "GECC is disabled, set amdgpu_ras_enable=1 to enable GECC in next boot cycle if needed\n"); + else + dev_info(adev->dev, "GECC is disabled\n"); + } else { + /* disable GECC in next boot cycle if ras is + * disabled by module parameter amdgpu_ras_enable + * and/or amdgpu_ras_mask, or boot_config_get call + * is failed + */ + ret = psp_boot_config_set(adev, 0); + if (ret) + dev_warn(adev->dev, "PSP set boot config failed\n"); + else + dev_warn(adev->dev, "GECC will be disabled in next boot cycle if set amdgpu_ras_enable and/or amdgpu_ras_mask to 0x0\n"); + } } } } @@ -3815,9 +3836,10 @@ int psp_init_cap_microcode(struct psp_context *psp, const char *chip_name) if (err == -ENODEV) { dev_warn(adev->dev, "cap microcode does not exist, skip\n"); err = 0; - goto out; + } else { + dev_err(adev->dev, "fail to initialize cap microcode\n"); } - dev_err(adev->dev, "fail to initialize cap microcode\n"); + goto out; } info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CAP]; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index f0924aa3f4e4..3c3312bbfee8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -1864,6 +1864,9 @@ int amdgpu_ras_sysfs_create(struct amdgpu_device *adev, if (!obj || obj->attr_inuse) return -EINVAL; + if (amdgpu_sriov_vf(adev) && !amdgpu_virt_ras_telemetry_block_en(adev, head->block)) + return 0; + get_obj(obj); snprintf(obj->fs_data.sysfs_name, sizeof(obj->fs_data.sysfs_name), @@ -3080,31 +3083,29 @@ static void amdgpu_ras_validate_threshold(struct amdgpu_device *adev, struct amdgpu_ras *con = amdgpu_ras_get_context(adev); /* - * Justification of value bad_page_cnt_threshold in ras structure - * - * Generally, 0 <= amdgpu_bad_page_threshold <= max record length - * in eeprom or amdgpu_bad_page_threshold == -2, introduce two - * scenarios accordingly. - * - * Bad page retirement enablement: - * - If amdgpu_bad_page_threshold = -2, - * bad_page_cnt_threshold = typical value by formula. - * - * - When the value from user is 0 < amdgpu_bad_page_threshold < - * max record length in eeprom, use it directly. - * - * Bad page retirement disablement: - * - If amdgpu_bad_page_threshold = 0, bad page retirement - * functionality is disabled, and bad_page_cnt_threshold will - * take no effect. + * amdgpu_bad_page_threshold is used to config + * the threshold for the number of bad pages. + * -1: Threshold is set to default value + * Driver will issue a warning message when threshold is reached + * and continue runtime services. + * 0: Disable bad page retirement + * Driver will not retire bad pages + * which is intended for debugging purpose. + * -2: Threshold is determined by a formula + * that assumes 1 bad page per 100M of local memory. + * Driver will continue runtime services when threhold is reached. + * 0 < threshold < max number of bad page records in EEPROM, + * A user-defined threshold is set + * Driver will halt runtime services when this custom threshold is reached. */ - - if (amdgpu_bad_page_threshold < 0) { + if (amdgpu_bad_page_threshold == -2) { u64 val = adev->gmc.mc_vram_size; do_div(val, RAS_BAD_PAGE_COVER); con->bad_page_cnt_threshold = min(lower_32_bits(val), max_count); + } else if (amdgpu_bad_page_threshold == -1) { + con->bad_page_cnt_threshold = ((con->reserved_pages_in_bytes) >> 21) << 4; } else { con->bad_page_cnt_threshold = min_t(int, max_count, amdgpu_bad_page_threshold); @@ -3759,8 +3760,9 @@ init_ras_enabled_flag: adev->ras_enabled = amdgpu_ras_enable == 0 ? 0 : adev->ras_hw_enabled & amdgpu_ras_mask; - /* aca is disabled by default */ - adev->aca.is_enabled = false; + /* aca is disabled by default except for psp v13_0_12 */ + adev->aca.is_enabled = + (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12)); /* bad page feature is not applicable to specific app platform */ if (adev->gmc.is_app_apu && @@ -3848,8 +3850,10 @@ static void amdgpu_ras_init_reserved_vram_size(struct amdgpu_device *adev) case IP_VERSION(13, 0, 2): case IP_VERSION(13, 0, 6): case IP_VERSION(13, 0, 12): + con->reserved_pages_in_bytes = AMDGPU_RAS_RESERVED_VRAM_SIZE_DEFAULT; + break; case IP_VERSION(13, 0, 14): - con->reserved_pages_in_bytes = AMDGPU_RAS_RESERVED_VRAM_SIZE; + con->reserved_pages_in_bytes = (AMDGPU_RAS_RESERVED_VRAM_SIZE_DEFAULT << 1); break; default: break; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h index 82db986c36a0..cc4586581dba 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h @@ -65,7 +65,7 @@ struct amdgpu_iv_entry; /* Reserve 8 physical dram row for possible retirement. * In worst cases, it will lose 8 * 2MB memory in vram domain */ -#define AMDGPU_RAS_RESERVED_VRAM_SIZE (16ULL << 20) +#define AMDGPU_RAS_RESERVED_VRAM_SIZE_DEFAULT (16ULL << 20) /* The high three bits indicates socketid */ #define AMDGPU_RAS_GET_FEATURES(val) ((val) & ~AMDGPU_RAS_FEATURES_SOCKETID_MASK) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c index 52c16bfeccaa..83b54efcaa87 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c @@ -177,7 +177,7 @@ static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev, if (!control) return false; - if (amdgpu_atomfirmware_ras_rom_addr(adev, &i2c_addr)) { + if (adev->bios && amdgpu_atomfirmware_ras_rom_addr(adev, &i2c_addr)) { /* The address given by VBIOS is an 8-bit, wire-format * address, i.e. the most significant byte. * @@ -558,16 +558,17 @@ bool amdgpu_ras_eeprom_check_err_threshold(struct amdgpu_device *adev) return false; if (con->eeprom_control.tbl_hdr.header == RAS_TABLE_HDR_BAD) { - if (amdgpu_bad_page_threshold == -1) { + if (con->eeprom_control.ras_num_bad_pages > con->bad_page_cnt_threshold) dev_warn(adev->dev, "RAS records:%d exceed threshold:%d", - con->eeprom_control.ras_num_bad_pages, con->bad_page_cnt_threshold); + con->eeprom_control.ras_num_bad_pages, con->bad_page_cnt_threshold); + if ((amdgpu_bad_page_threshold == -1) || + (amdgpu_bad_page_threshold == -2)) { dev_warn(adev->dev, - "But GPU can be operated due to bad_page_threshold = -1.\n"); + "Please consult AMD Service Action Guide (SAG) for appropriate service procedures.\n"); return false; } else { - dev_warn(adev->dev, "This GPU is in BAD status."); - dev_warn(adev->dev, "Please retire it or set a larger " - "threshold value when reloading driver.\n"); + dev_warn(adev->dev, + "Please consider adjusting the customized threshold.\n"); return true; } } @@ -758,7 +759,8 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control) control->tbl_rai.health_percent = 0; } - if (amdgpu_bad_page_threshold != -1) + if ((amdgpu_bad_page_threshold != -1) && + (amdgpu_bad_page_threshold != -2)) ras->is_rma = true; /* ignore the -ENOTSUPP return value */ @@ -1428,8 +1430,9 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control) res = __verify_ras_table_checksum(control); if (res) - DRM_ERROR("RAS table incorrect checksum or error:%d\n", - res); + dev_err(adev->dev, + "RAS table incorrect checksum or error:%d\n", + res); /* Warn if we are at 90% of the threshold or above */ @@ -1447,8 +1450,9 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control) res = __verify_ras_table_checksum(control); if (res) { - dev_err(adev->dev, "RAS Table incorrect checksum or error:%d\n", - res); + dev_err(adev->dev, + "RAS Table incorrect checksum or error:%d\n", + res); return -EINVAL; } if (ras->bad_page_cnt_threshold > control->ras_num_bad_pages) { @@ -1466,17 +1470,18 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control) res = amdgpu_ras_eeprom_correct_header_tag(control, RAS_TABLE_HDR_VAL); } else { - dev_err(adev->dev, "RAS records:%d exceed threshold:%d", + dev_warn(adev->dev, + "RAS records:%d exceed threshold:%d\n", control->ras_num_bad_pages, ras->bad_page_cnt_threshold); - if (amdgpu_bad_page_threshold == -1) { - dev_warn(adev->dev, "GPU will be initialized due to bad_page_threshold = -1."); + if ((amdgpu_bad_page_threshold == -1) || + (amdgpu_bad_page_threshold == -2)) { res = 0; + dev_warn(adev->dev, + "Please consult AMD Service Action Guide (SAG) for appropriate service procedures\n"); } else { ras->is_rma = true; - dev_err(adev->dev, - "RAS records:%d exceed threshold:%d, " - "GPU will not be initialized. Replace this GPU or increase the threshold", - control->ras_num_bad_pages, ras->bad_page_cnt_threshold); + dev_warn(adev->dev, + "User defined threshold is set, runtime service will be halt when threshold is reached\n"); } } } else { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index a6e28fe3f8d6..f53887e2f528 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -324,20 +324,27 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, /* always set cond_exec_polling to CONTINUE */ *ring->cond_exe_cpu_addr = 1; - r = amdgpu_fence_driver_start_ring(ring, irq_src, irq_type); - if (r) { - dev_err(adev->dev, "failed initializing fences (%d).\n", r); - return r; - } + if (ring->funcs->type != AMDGPU_RING_TYPE_CPER) { + r = amdgpu_fence_driver_start_ring(ring, irq_src, irq_type); + if (r) { + dev_err(adev->dev, "failed initializing fences (%d).\n", r); + return r; + } - max_ibs_dw = ring->funcs->emit_frame_size + - amdgpu_ring_max_ibs(ring->funcs->type) * ring->funcs->emit_ib_size; - max_ibs_dw = (max_ibs_dw + ring->funcs->align_mask) & ~ring->funcs->align_mask; + max_ibs_dw = ring->funcs->emit_frame_size + + amdgpu_ring_max_ibs(ring->funcs->type) * ring->funcs->emit_ib_size; + max_ibs_dw = (max_ibs_dw + ring->funcs->align_mask) & ~ring->funcs->align_mask; - if (WARN_ON(max_ibs_dw > max_dw)) - max_dw = max_ibs_dw; + if (WARN_ON(max_ibs_dw > max_dw)) + max_dw = max_ibs_dw; - ring->ring_size = roundup_pow_of_two(max_dw * 4 * sched_hw_submission); + ring->ring_size = roundup_pow_of_two(max_dw * 4 * sched_hw_submission); + } else { + ring->ring_size = roundup_pow_of_two(max_dw * 4); + ring->count_dw = (ring->ring_size - 4) >> 2; + /* ring buffer is empty now */ + ring->wptr = *ring->rptr_cpu_addr = 0; + } ring->buf_mask = (ring->ring_size / 4) - 1; ring->ptr_mask = ring->funcs->support_64bit_ptrs ? @@ -493,6 +500,7 @@ static ssize_t amdgpu_debugfs_ring_read(struct file *f, char __user *buf, { struct amdgpu_ring *ring = file_inode(f)->i_private; uint32_t value, result, early[3]; + uint64_t p; loff_t i; int r; @@ -502,13 +510,18 @@ static ssize_t amdgpu_debugfs_ring_read(struct file *f, char __user *buf, result = 0; if (*pos < 12) { + if (ring->funcs->type == AMDGPU_RING_TYPE_CPER) + mutex_lock(&ring->adev->cper.ring_lock); + early[0] = amdgpu_ring_get_rptr(ring) & ring->buf_mask; early[1] = amdgpu_ring_get_wptr(ring) & ring->buf_mask; early[2] = ring->wptr & ring->buf_mask; for (i = *pos / 4; i < 3 && size; i++) { r = put_user(early[i], (uint32_t *)buf); - if (r) - return r; + if (r) { + result = r; + goto out; + } buf += 4; result += 4; size -= 4; @@ -516,20 +529,50 @@ static ssize_t amdgpu_debugfs_ring_read(struct file *f, char __user *buf, } } - while (size) { - if (*pos >= (ring->ring_size + 12)) - return result; + if (ring->funcs->type != AMDGPU_RING_TYPE_CPER) { + while (size) { + if (*pos >= (ring->ring_size + 12)) + return result; - value = ring->ring[(*pos - 12)/4]; - r = put_user(value, (uint32_t *)buf); - if (r) - return r; - buf += 4; - result += 4; - size -= 4; - *pos += 4; + value = ring->ring[(*pos - 12)/4]; + r = put_user(value, (uint32_t *)buf); + if (r) + return r; + buf += 4; + result += 4; + size -= 4; + *pos += 4; + } + } else { + p = early[0]; + if (early[0] <= early[1]) + size = (early[1] - early[0]); + else + size = ring->ring_size - (early[0] - early[1]); + + while (size) { + if (p == early[1]) + goto out; + + value = ring->ring[p]; + r = put_user(value, (uint32_t *)buf); + if (r) { + result = r; + goto out; + } + + buf += 4; + result += 4; + size--; + p++; + p &= ring->ptr_mask; + } } +out: + if (ring->funcs->type == AMDGPU_RING_TYPE_CPER) + mutex_unlock(&ring->adev->cper.ring_lock); + return result; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index dee5a1b4e572..7372e4aed6b0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -37,7 +37,7 @@ struct amdgpu_job; struct amdgpu_vm; /* max number of rings */ -#define AMDGPU_MAX_RINGS 124 +#define AMDGPU_MAX_RINGS 132 #define AMDGPU_MAX_HWIP_RINGS 64 #define AMDGPU_MAX_GFX_RINGS 2 #define AMDGPU_MAX_SW_GFX_RINGS 2 @@ -82,6 +82,7 @@ enum amdgpu_ring_type { AMDGPU_RING_TYPE_KIQ, AMDGPU_RING_TYPE_MES, AMDGPU_RING_TYPE_UMSCH_MM, + AMDGPU_RING_TYPE_CPER, }; enum amdgpu_ib_pool_type { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index ff286940ab43..bcb4bcc4ab75 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -309,7 +309,7 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, mutex_lock(&adev->mman.gtt_window_lock); while (src_mm.remaining) { uint64_t from, to, cur_size, tiling_flags; - uint32_t num_type, data_format, max_com; + uint32_t num_type, data_format, max_com, write_compress_disable; struct dma_fence *next; /* Never copy more than 256MiB at once to avoid a timeout */ @@ -340,9 +340,13 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, max_com = AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_MAX_COMPRESSED_BLOCK); num_type = AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_NUMBER_TYPE); data_format = AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_DATA_FORMAT); + write_compress_disable = + AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_WRITE_COMPRESS_DISABLE); copy_flags |= (AMDGPU_COPY_FLAGS_SET(MAX_COMPRESSED, max_com) | AMDGPU_COPY_FLAGS_SET(NUMBER_TYPE, num_type) | - AMDGPU_COPY_FLAGS_SET(DATA_FORMAT, data_format)); + AMDGPU_COPY_FLAGS_SET(DATA_FORMAT, data_format) | + AMDGPU_COPY_FLAGS_SET(WRITE_COMPRESS_DISABLE, + write_compress_disable)); } r = amdgpu_copy_buffer(ring, from, to, cur_size, resv, @@ -1960,10 +1964,19 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) /* Compute GTT size, either based on TTM limit * or whatever the user passed on module init. */ - if (amdgpu_gtt_size == -1) - gtt_size = ttm_tt_pages_limit() << PAGE_SHIFT; - else - gtt_size = (uint64_t)amdgpu_gtt_size << 20; + gtt_size = ttm_tt_pages_limit() << PAGE_SHIFT; + if (amdgpu_gtt_size != -1) { + uint64_t configured_size = (uint64_t)amdgpu_gtt_size << 20; + + drm_warn(&adev->ddev, + "Configuring gttsize via module parameter is deprecated, please use ttm.pages_limit\n"); + if (gtt_size != configured_size) + drm_warn(&adev->ddev, + "GTT size has been set as %llu but TTM size has been set as %llu, this is unusual\n", + configured_size, gtt_size); + + gtt_size = configured_size; + } /* Initialize GTT memory pool */ r = amdgpu_gtt_mgr_init(adev, gtt_size); @@ -1974,6 +1987,11 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) DRM_INFO("amdgpu: %uM of GTT memory ready.\n", (unsigned int)(gtt_size / (1024 * 1024))); + if (adev->flags & AMD_IS_APU) { + if (adev->gmc.real_vram_size < gtt_size) + adev->apu_prefer_gtt = true; + } + /* Initialize doorbell pool on PCI BAR */ r = amdgpu_ttm_init_on_chip(adev, AMDGPU_PL_DOORBELL, adev->doorbell.size / PAGE_SIZE); if (r) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h index 461fb8090ae0..208b7d1d8a27 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h @@ -119,6 +119,8 @@ struct amdgpu_copy_mem { #define AMDGPU_COPY_FLAGS_NUMBER_TYPE_MASK 0x07 #define AMDGPU_COPY_FLAGS_DATA_FORMAT_SHIFT 8 #define AMDGPU_COPY_FLAGS_DATA_FORMAT_MASK 0x3f +#define AMDGPU_COPY_FLAGS_WRITE_COMPRESS_DISABLE_SHIFT 14 +#define AMDGPU_COPY_FLAGS_WRITE_COMPRESS_DISABLE_MASK 0x1 #define AMDGPU_COPY_FLAGS_SET(field, value) \ (((__u32)(value) & AMDGPU_COPY_FLAGS_##field##_MASK) << AMDGPU_COPY_FLAGS_##field##_SHIFT) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c index cf700824b960..3d9e9fdc10b4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c @@ -1216,6 +1216,7 @@ static const char *amdgpu_ucode_legacy_naming(struct amdgpu_device *adev, int bl case IP_VERSION(11, 0, 13): return "beige_goby"; case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): return "vangogh"; case IP_VERSION(12, 0, 1): return "green_sardine"; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c index eafe20d8fe0b..0a1ef95b2866 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c @@ -387,6 +387,45 @@ int amdgpu_umc_fill_error_record(struct ras_err_data *err_data, return 0; } +static int amdgpu_umc_loop_all_aid(struct amdgpu_device *adev, umc_func func, + void *data) +{ + uint32_t umc_node_inst; + uint32_t node_inst; + uint32_t umc_inst; + uint32_t ch_inst; + int ret; + + /* + * This loop is done based on the following - + * umc.active mask = mask of active umc instances across all nodes + * umc.umc_inst_num = maximum number of umc instancess per node + * umc.node_inst_num = maximum number of node instances + * Channel instances are not assumed to be harvested. + */ + dev_dbg(adev->dev, "active umcs :%lx umc_inst per node: %d", + adev->umc.active_mask, adev->umc.umc_inst_num); + for_each_set_bit(umc_node_inst, &(adev->umc.active_mask), + adev->umc.node_inst_num * adev->umc.umc_inst_num) { + node_inst = umc_node_inst / adev->umc.umc_inst_num; + umc_inst = umc_node_inst % adev->umc.umc_inst_num; + LOOP_UMC_CH_INST(ch_inst) { + dev_dbg(adev->dev, + "node_inst :%d umc_inst: %d ch_inst: %d", + node_inst, umc_inst, ch_inst); + ret = func(adev, node_inst, umc_inst, ch_inst, data); + if (ret) { + dev_err(adev->dev, + "Node %d umc %d ch %d func returns %d\n", + node_inst, umc_inst, ch_inst, ret); + return ret; + } + } + } + + return 0; +} + int amdgpu_umc_loop_channels(struct amdgpu_device *adev, umc_func func, void *data) { @@ -395,6 +434,9 @@ int amdgpu_umc_loop_channels(struct amdgpu_device *adev, uint32_t ch_inst = 0; int ret = 0; + if (adev->aid_mask) + return amdgpu_umc_loop_all_aid(adev, func, data); + if (adev->umc.node_inst_num) { LOOP_UMC_EACH_NODE_INST_AND_CH(node_inst, umc_inst, ch_inst) { ret = func(adev, node_inst, umc_inst, ch_inst, data); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c index dde15c6a96e1..cd707d70a0bf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c @@ -32,462 +32,7 @@ #include "amdgpu_umsch_mm.h" #include "umsch_mm_v4_0.h" -struct umsch_mm_test_ctx_data { - uint8_t process_csa[PAGE_SIZE]; - uint8_t vpe_ctx_csa[PAGE_SIZE]; - uint8_t vcn_ctx_csa[PAGE_SIZE]; -}; - -struct umsch_mm_test_mqd_data { - uint8_t vpe_mqd[PAGE_SIZE]; - uint8_t vcn_mqd[PAGE_SIZE]; -}; - -struct umsch_mm_test_ring_data { - uint8_t vpe_ring[PAGE_SIZE]; - uint8_t vpe_ib[PAGE_SIZE]; - uint8_t vcn_ring[PAGE_SIZE]; - uint8_t vcn_ib[PAGE_SIZE]; -}; - -struct umsch_mm_test_queue_info { - uint64_t mqd_addr; - uint64_t csa_addr; - uint32_t doorbell_offset_0; - uint32_t doorbell_offset_1; - enum UMSCH_SWIP_ENGINE_TYPE engine; -}; - -struct umsch_mm_test { - struct amdgpu_bo *ctx_data_obj; - uint64_t ctx_data_gpu_addr; - uint32_t *ctx_data_cpu_addr; - - struct amdgpu_bo *mqd_data_obj; - uint64_t mqd_data_gpu_addr; - uint32_t *mqd_data_cpu_addr; - - struct amdgpu_bo *ring_data_obj; - uint64_t ring_data_gpu_addr; - uint32_t *ring_data_cpu_addr; - - - struct amdgpu_vm *vm; - struct amdgpu_bo_va *bo_va; - uint32_t pasid; - uint32_t vm_cntx_cntl; - uint32_t num_queues; -}; - -static int map_ring_data(struct amdgpu_device *adev, struct amdgpu_vm *vm, - struct amdgpu_bo *bo, struct amdgpu_bo_va **bo_va, - uint64_t addr, uint32_t size) -{ - struct amdgpu_sync sync; - struct drm_exec exec; - int r; - - amdgpu_sync_create(&sync); - - drm_exec_init(&exec, 0, 0); - drm_exec_until_all_locked(&exec) { - r = drm_exec_lock_obj(&exec, &bo->tbo.base); - drm_exec_retry_on_contention(&exec); - if (unlikely(r)) - goto error_fini_exec; - - r = amdgpu_vm_lock_pd(vm, &exec, 0); - drm_exec_retry_on_contention(&exec); - if (unlikely(r)) - goto error_fini_exec; - } - - *bo_va = amdgpu_vm_bo_add(adev, vm, bo); - if (!*bo_va) { - r = -ENOMEM; - goto error_fini_exec; - } - - r = amdgpu_vm_bo_map(adev, *bo_va, addr, 0, size, - AMDGPU_PTE_READABLE | AMDGPU_PTE_WRITEABLE | - AMDGPU_PTE_EXECUTABLE); - - if (r) - goto error_del_bo_va; - - - r = amdgpu_vm_bo_update(adev, *bo_va, false); - if (r) - goto error_del_bo_va; - - amdgpu_sync_fence(&sync, (*bo_va)->last_pt_update); - - r = amdgpu_vm_update_pdes(adev, vm, false); - if (r) - goto error_del_bo_va; - - amdgpu_sync_fence(&sync, vm->last_update); - - amdgpu_sync_wait(&sync, false); - drm_exec_fini(&exec); - - amdgpu_sync_free(&sync); - - return 0; - -error_del_bo_va: - amdgpu_vm_bo_del(adev, *bo_va); - amdgpu_sync_free(&sync); - -error_fini_exec: - drm_exec_fini(&exec); - amdgpu_sync_free(&sync); - return r; -} - -static int unmap_ring_data(struct amdgpu_device *adev, struct amdgpu_vm *vm, - struct amdgpu_bo *bo, struct amdgpu_bo_va *bo_va, - uint64_t addr) -{ - struct drm_exec exec; - long r; - - drm_exec_init(&exec, 0, 0); - drm_exec_until_all_locked(&exec) { - r = drm_exec_lock_obj(&exec, &bo->tbo.base); - drm_exec_retry_on_contention(&exec); - if (unlikely(r)) - goto out_unlock; - - r = amdgpu_vm_lock_pd(vm, &exec, 0); - drm_exec_retry_on_contention(&exec); - if (unlikely(r)) - goto out_unlock; - } - - - r = amdgpu_vm_bo_unmap(adev, bo_va, addr); - if (r) - goto out_unlock; - - amdgpu_vm_bo_del(adev, bo_va); - -out_unlock: - drm_exec_fini(&exec); - - return r; -} - -static void setup_vpe_queue(struct amdgpu_device *adev, - struct umsch_mm_test *test, - struct umsch_mm_test_queue_info *qinfo) -{ - struct MQD_INFO *mqd = (struct MQD_INFO *)test->mqd_data_cpu_addr; - uint64_t ring_gpu_addr = test->ring_data_gpu_addr; - - mqd->rb_base_lo = (ring_gpu_addr >> 8); - mqd->rb_base_hi = (ring_gpu_addr >> 40); - mqd->rb_size = PAGE_SIZE / 4; - mqd->wptr_val = 0; - mqd->rptr_val = 0; - mqd->unmapped = 1; - - if (adev->vpe.collaborate_mode) - memcpy(++mqd, test->mqd_data_cpu_addr, sizeof(struct MQD_INFO)); - - qinfo->mqd_addr = test->mqd_data_gpu_addr; - qinfo->csa_addr = test->ctx_data_gpu_addr + - offsetof(struct umsch_mm_test_ctx_data, vpe_ctx_csa); - qinfo->doorbell_offset_0 = 0; - qinfo->doorbell_offset_1 = 0; -} - -static void setup_vcn_queue(struct amdgpu_device *adev, - struct umsch_mm_test *test, - struct umsch_mm_test_queue_info *qinfo) -{ -} - -static int add_test_queue(struct amdgpu_device *adev, - struct umsch_mm_test *test, - struct umsch_mm_test_queue_info *qinfo) -{ - struct umsch_mm_add_queue_input queue_input = {}; - int r; - - queue_input.process_id = test->pasid; - queue_input.page_table_base_addr = amdgpu_gmc_pd_addr(test->vm->root.bo); - - queue_input.process_va_start = 0; - queue_input.process_va_end = (adev->vm_manager.max_pfn - 1) << AMDGPU_GPU_PAGE_SHIFT; - - queue_input.process_quantum = 100000; /* 10ms */ - queue_input.process_csa_addr = test->ctx_data_gpu_addr + - offsetof(struct umsch_mm_test_ctx_data, process_csa); - - queue_input.context_quantum = 10000; /* 1ms */ - queue_input.context_csa_addr = qinfo->csa_addr; - - queue_input.inprocess_context_priority = CONTEXT_PRIORITY_LEVEL_NORMAL; - queue_input.context_global_priority_level = CONTEXT_PRIORITY_LEVEL_NORMAL; - queue_input.doorbell_offset_0 = qinfo->doorbell_offset_0; - queue_input.doorbell_offset_1 = qinfo->doorbell_offset_1; - - queue_input.engine_type = qinfo->engine; - queue_input.mqd_addr = qinfo->mqd_addr; - queue_input.vm_context_cntl = test->vm_cntx_cntl; - - amdgpu_umsch_mm_lock(&adev->umsch_mm); - r = adev->umsch_mm.funcs->add_queue(&adev->umsch_mm, &queue_input); - amdgpu_umsch_mm_unlock(&adev->umsch_mm); - if (r) - return r; - - return 0; -} - -static int remove_test_queue(struct amdgpu_device *adev, - struct umsch_mm_test *test, - struct umsch_mm_test_queue_info *qinfo) -{ - struct umsch_mm_remove_queue_input queue_input = {}; - int r; - - queue_input.doorbell_offset_0 = qinfo->doorbell_offset_0; - queue_input.doorbell_offset_1 = qinfo->doorbell_offset_1; - queue_input.context_csa_addr = qinfo->csa_addr; - - amdgpu_umsch_mm_lock(&adev->umsch_mm); - r = adev->umsch_mm.funcs->remove_queue(&adev->umsch_mm, &queue_input); - amdgpu_umsch_mm_unlock(&adev->umsch_mm); - if (r) - return r; - - return 0; -} - -static int submit_vpe_queue(struct amdgpu_device *adev, struct umsch_mm_test *test) -{ - struct MQD_INFO *mqd = (struct MQD_INFO *)test->mqd_data_cpu_addr; - uint32_t *ring = test->ring_data_cpu_addr + - offsetof(struct umsch_mm_test_ring_data, vpe_ring) / 4; - uint32_t *ib = test->ring_data_cpu_addr + - offsetof(struct umsch_mm_test_ring_data, vpe_ib) / 4; - uint64_t ib_gpu_addr = test->ring_data_gpu_addr + - offsetof(struct umsch_mm_test_ring_data, vpe_ib); - uint32_t *fence = ib + 2048 / 4; - uint64_t fence_gpu_addr = ib_gpu_addr + 2048; - const uint32_t test_pattern = 0xdeadbeef; - int i; - - ib[0] = VPE_CMD_HEADER(VPE_CMD_OPCODE_FENCE, 0); - ib[1] = lower_32_bits(fence_gpu_addr); - ib[2] = upper_32_bits(fence_gpu_addr); - ib[3] = test_pattern; - - ring[0] = VPE_CMD_HEADER(VPE_CMD_OPCODE_INDIRECT, 0); - ring[1] = (ib_gpu_addr & 0xffffffe0); - ring[2] = upper_32_bits(ib_gpu_addr); - ring[3] = 4; - ring[4] = 0; - ring[5] = 0; - - mqd->wptr_val = (6 << 2); - if (adev->vpe.collaborate_mode) - (++mqd)->wptr_val = (6 << 2); - - WDOORBELL32(adev->umsch_mm.agdb_index[CONTEXT_PRIORITY_LEVEL_NORMAL], mqd->wptr_val); - - for (i = 0; i < adev->usec_timeout; i++) { - if (*fence == test_pattern) - return 0; - udelay(1); - } - - dev_err(adev->dev, "vpe queue submission timeout\n"); - - return -ETIMEDOUT; -} - -static int submit_vcn_queue(struct amdgpu_device *adev, struct umsch_mm_test *test) -{ - return 0; -} - -static int setup_umsch_mm_test(struct amdgpu_device *adev, - struct umsch_mm_test *test) -{ - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; - int r; - - test->vm_cntx_cntl = hub->vm_cntx_cntl; - - test->vm = kzalloc(sizeof(*test->vm), GFP_KERNEL); - if (!test->vm) { - r = -ENOMEM; - return r; - } - - r = amdgpu_vm_init(adev, test->vm, -1); - if (r) - goto error_free_vm; - - r = amdgpu_pasid_alloc(16); - if (r < 0) - goto error_fini_vm; - test->pasid = r; - - r = amdgpu_bo_create_kernel(adev, sizeof(struct umsch_mm_test_ctx_data), - PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, - &test->ctx_data_obj, - &test->ctx_data_gpu_addr, - (void **)&test->ctx_data_cpu_addr); - if (r) - goto error_free_pasid; - - memset(test->ctx_data_cpu_addr, 0, sizeof(struct umsch_mm_test_ctx_data)); - - r = amdgpu_bo_create_kernel(adev, PAGE_SIZE, - PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, - &test->mqd_data_obj, - &test->mqd_data_gpu_addr, - (void **)&test->mqd_data_cpu_addr); - if (r) - goto error_free_ctx_data_obj; - - memset(test->mqd_data_cpu_addr, 0, PAGE_SIZE); - - r = amdgpu_bo_create_kernel(adev, sizeof(struct umsch_mm_test_ring_data), - PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, - &test->ring_data_obj, - NULL, - (void **)&test->ring_data_cpu_addr); - if (r) - goto error_free_mqd_data_obj; - - memset(test->ring_data_cpu_addr, 0, sizeof(struct umsch_mm_test_ring_data)); - - test->ring_data_gpu_addr = AMDGPU_VA_RESERVED_BOTTOM; - r = map_ring_data(adev, test->vm, test->ring_data_obj, &test->bo_va, - test->ring_data_gpu_addr, sizeof(struct umsch_mm_test_ring_data)); - if (r) - goto error_free_ring_data_obj; - - return 0; - -error_free_ring_data_obj: - amdgpu_bo_free_kernel(&test->ring_data_obj, NULL, - (void **)&test->ring_data_cpu_addr); -error_free_mqd_data_obj: - amdgpu_bo_free_kernel(&test->mqd_data_obj, &test->mqd_data_gpu_addr, - (void **)&test->mqd_data_cpu_addr); -error_free_ctx_data_obj: - amdgpu_bo_free_kernel(&test->ctx_data_obj, &test->ctx_data_gpu_addr, - (void **)&test->ctx_data_cpu_addr); -error_free_pasid: - amdgpu_pasid_free(test->pasid); -error_fini_vm: - amdgpu_vm_fini(adev, test->vm); -error_free_vm: - kfree(test->vm); - - return r; -} - -static void cleanup_umsch_mm_test(struct amdgpu_device *adev, - struct umsch_mm_test *test) -{ - unmap_ring_data(adev, test->vm, test->ring_data_obj, - test->bo_va, test->ring_data_gpu_addr); - amdgpu_bo_free_kernel(&test->mqd_data_obj, &test->mqd_data_gpu_addr, - (void **)&test->mqd_data_cpu_addr); - amdgpu_bo_free_kernel(&test->ring_data_obj, NULL, - (void **)&test->ring_data_cpu_addr); - amdgpu_bo_free_kernel(&test->ctx_data_obj, &test->ctx_data_gpu_addr, - (void **)&test->ctx_data_cpu_addr); - amdgpu_pasid_free(test->pasid); - amdgpu_vm_fini(adev, test->vm); - kfree(test->vm); -} - -static int setup_test_queues(struct amdgpu_device *adev, - struct umsch_mm_test *test, - struct umsch_mm_test_queue_info *qinfo) -{ - int i, r; - - for (i = 0; i < test->num_queues; i++) { - if (qinfo[i].engine == UMSCH_SWIP_ENGINE_TYPE_VPE) - setup_vpe_queue(adev, test, &qinfo[i]); - else - setup_vcn_queue(adev, test, &qinfo[i]); - - r = add_test_queue(adev, test, &qinfo[i]); - if (r) - return r; - } - - return 0; -} - -static int submit_test_queues(struct amdgpu_device *adev, - struct umsch_mm_test *test, - struct umsch_mm_test_queue_info *qinfo) -{ - int i, r; - - for (i = 0; i < test->num_queues; i++) { - if (qinfo[i].engine == UMSCH_SWIP_ENGINE_TYPE_VPE) - r = submit_vpe_queue(adev, test); - else - r = submit_vcn_queue(adev, test); - if (r) - return r; - } - - return 0; -} - -static void cleanup_test_queues(struct amdgpu_device *adev, - struct umsch_mm_test *test, - struct umsch_mm_test_queue_info *qinfo) -{ - int i; - - for (i = 0; i < test->num_queues; i++) - remove_test_queue(adev, test, &qinfo[i]); -} - -static int umsch_mm_test(struct amdgpu_device *adev) -{ - struct umsch_mm_test_queue_info qinfo[] = { - { .engine = UMSCH_SWIP_ENGINE_TYPE_VPE }, - }; - struct umsch_mm_test test = { .num_queues = ARRAY_SIZE(qinfo) }; - int r; - - r = setup_umsch_mm_test(adev, &test); - if (r) - return r; - - r = setup_test_queues(adev, &test, qinfo); - if (r) - goto cleanup; - - r = submit_test_queues(adev, &test, qinfo); - if (r) - goto cleanup; - - cleanup_test_queues(adev, &test, qinfo); - cleanup_umsch_mm_test(adev, &test); - - return 0; - -cleanup: - cleanup_test_queues(adev, &test, qinfo); - cleanup_umsch_mm_test(adev, &test); - return r; -} +MODULE_FIRMWARE("amdgpu/umsch_mm_4_0_0.bin"); int amdgpu_umsch_mm_submit_pkt(struct amdgpu_umsch_mm *umsch, void *pkt, int ndws) { @@ -581,14 +126,14 @@ int amdgpu_umsch_mm_init_microcode(struct amdgpu_umsch_mm *umsch) switch (amdgpu_ip_version(adev, VCN_HWIP, 0)) { case IP_VERSION(4, 0, 5): case IP_VERSION(4, 0, 6): - fw_name = "amdgpu/umsch_mm_4_0_0.bin"; + fw_name = "4_0_0"; break; default: - break; + return -EINVAL; } r = amdgpu_ucode_request(adev, &adev->umsch_mm.fw, AMDGPU_UCODE_REQUIRED, - "%s", fw_name); + "amdgpu/umsch_mm_%s.bin", fw_name); if (r) { release_firmware(adev->umsch_mm.fw); adev->umsch_mm.fw = NULL; @@ -792,7 +337,7 @@ static int umsch_mm_late_init(struct amdgpu_ip_block *ip_block) if (amdgpu_in_reset(adev) || adev->in_s0ix || adev->in_suspend) return 0; - return umsch_mm_test(adev); + return 0; } static int umsch_mm_sw_init(struct amdgpu_ip_block *ip_block) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index adaf4388ad28..c92f683ee595 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -238,6 +238,12 @@ #define AMDGPU_DRM_KEY_INJECT_WORKAROUND_VCNFW_ASD_HANDSHAKING 2 +enum amdgpu_vcn_caps { + AMDGPU_VCN_RRMT_ENABLED, +}; + +#define AMDGPU_VCN_CAPS(caps) BIT(AMDGPU_VCN_##caps) + enum fw_queue_mode { FW_QUEUE_RING_RESET = 1, FW_QUEUE_DPG_HOLD_OFF = 2, @@ -345,6 +351,7 @@ struct amdgpu_vcn { uint32_t *ip_dump; uint32_t supported_reset; + uint32_t caps; }; struct amdgpu_fw_shared_rb_ptrs_struct { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index 0af469ec6fcc..e6f0152e5b08 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -1017,6 +1017,7 @@ u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 f void *scratch_reg2; void *scratch_reg3; void *spare_int; + unsigned long flags; if (!adev->gfx.rlc.rlcg_reg_access_supported) { dev_err(adev->dev, @@ -1038,7 +1039,7 @@ u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 f scratch_reg2 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg2; scratch_reg3 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg3; - mutex_lock(&adev->virt.rlcg_reg_lock); + spin_lock_irqsave(&adev->virt.rlcg_reg_lock, flags); if (reg_access_ctrl->spare_int) spare_int = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->spare_int; @@ -1097,7 +1098,7 @@ u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 f ret = readl(scratch_reg0); - mutex_unlock(&adev->virt.rlcg_reg_lock); + spin_unlock_irqrestore(&adev->virt.rlcg_reg_lock, flags); return ret; } @@ -1246,7 +1247,8 @@ amdgpu_ras_block_to_sriov(struct amdgpu_device *adev, enum amdgpu_ras_block bloc case AMDGPU_RAS_BLOCK__MPIO: return RAS_TELEMETRY_GPU_BLOCK_MPIO; default: - dev_err(adev->dev, "Unsupported SRIOV RAS telemetry block 0x%x\n", block); + DRM_WARN_ONCE("Unsupported SRIOV RAS telemetry block 0x%x\n", + block); return RAS_TELEMETRY_GPU_BLOCK_COUNT; } } @@ -1331,3 +1333,17 @@ int amdgpu_virt_ras_telemetry_post_reset(struct amdgpu_device *adev) return 0; } + +bool amdgpu_virt_ras_telemetry_block_en(struct amdgpu_device *adev, + enum amdgpu_ras_block block) +{ + enum amd_sriov_ras_telemetry_gpu_block sriov_block; + + sriov_block = amdgpu_ras_block_to_sriov(adev, block); + + if (sriov_block >= RAS_TELEMETRY_GPU_BLOCK_COUNT || + !amdgpu_sriov_ras_telemetry_block_en(adev, sriov_block)) + return false; + + return true; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h index 5381b8d596e6..0f3ccae5c1ab 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h @@ -279,7 +279,8 @@ struct amdgpu_virt { /* the ucode id to signal the autoload */ uint32_t autoload_ucode_id; - struct mutex rlcg_reg_lock; + /* Spinlock to protect access to the RLCG register interface */ + spinlock_t rlcg_reg_lock; union amd_sriov_ras_caps ras_en_caps; union amd_sriov_ras_caps ras_telemetry_en_caps; @@ -406,4 +407,6 @@ bool amdgpu_virt_get_ras_capability(struct amdgpu_device *adev); int amdgpu_virt_req_ras_err_count(struct amdgpu_device *adev, enum amdgpu_ras_block block, struct ras_err_data *err_data); int amdgpu_virt_ras_telemetry_post_reset(struct amdgpu_device *adev); +bool amdgpu_virt_ras_telemetry_block_en(struct amdgpu_device *adev, + enum amdgpu_ras_block block); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 5c07777d3239..1c0fd95c3820 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -2534,8 +2534,6 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, spin_lock_init(&vm->status_lock); INIT_LIST_HEAD(&vm->freed); INIT_LIST_HEAD(&vm->done); - INIT_LIST_HEAD(&vm->pt_freed); - INIT_WORK(&vm->pt_free_work, amdgpu_vm_pt_free_work); INIT_KFIFO(vm->faults); r = amdgpu_vm_init_entities(adev, vm); @@ -2674,20 +2672,6 @@ unreserve_bo: return r; } -/** - * amdgpu_vm_release_compute - release a compute vm - * @adev: amdgpu_device pointer - * @vm: a vm turned into compute vm by calling amdgpu_vm_make_compute - * - * This is a correspondant of amdgpu_vm_make_compute. It decouples compute - * pasid from vm. Compute should stop use of vm after this call. - */ -void amdgpu_vm_release_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm) -{ - amdgpu_vm_set_pasid(adev, vm, 0); - vm->is_compute_context = false; -} - static int amdgpu_vm_stats_is_zero(struct amdgpu_vm *vm) { for (int i = 0; i < __AMDGPU_PL_NUM; ++i) { @@ -2717,8 +2701,6 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) amdgpu_amdkfd_gpuvm_destroy_cb(adev, vm); - flush_work(&vm->pt_free_work); - root = amdgpu_bo_ref(vm->root.bo); amdgpu_bo_reserve(root, true); amdgpu_vm_set_pasid(adev, vm, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h index a3e128e373bc..f3ad687125ad 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h @@ -374,10 +374,6 @@ struct amdgpu_vm { /* BOs which are invalidated, has been updated in the PTs */ struct list_head done; - /* PT BOs scheduled to free and fill with zero if vm_resv is not hold */ - struct list_head pt_freed; - struct work_struct pt_free_work; - /* contains the page directory */ struct amdgpu_vm_bo_base root; struct dma_fence *last_update; @@ -493,7 +489,6 @@ int amdgpu_vm_set_pasid(struct amdgpu_device *adev, struct amdgpu_vm *vm, long amdgpu_vm_wait_idle(struct amdgpu_vm *vm, long timeout); int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, int32_t xcp_id); int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm); -void amdgpu_vm_release_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm); void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm); int amdgpu_vm_lock_pd(struct amdgpu_vm *vm, struct drm_exec *exec, unsigned int num_fences); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c index b0bf21682115..30022123b0bf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c @@ -547,27 +547,6 @@ static void amdgpu_vm_pt_free(struct amdgpu_vm_bo_base *entry) amdgpu_bo_unref(&entry->bo); } -void amdgpu_vm_pt_free_work(struct work_struct *work) -{ - struct amdgpu_vm_bo_base *entry, *next; - struct amdgpu_vm *vm; - LIST_HEAD(pt_freed); - - vm = container_of(work, struct amdgpu_vm, pt_free_work); - - spin_lock(&vm->status_lock); - list_splice_init(&vm->pt_freed, &pt_freed); - spin_unlock(&vm->status_lock); - - /* flush_work in amdgpu_vm_fini ensure vm->root.bo is valid. */ - amdgpu_bo_reserve(vm->root.bo, true); - - list_for_each_entry_safe(entry, next, &pt_freed, vm_status) - amdgpu_vm_pt_free(entry); - - amdgpu_bo_unreserve(vm->root.bo); -} - /** * amdgpu_vm_pt_free_list - free PD/PT levels * @@ -580,19 +559,15 @@ void amdgpu_vm_pt_free_list(struct amdgpu_device *adev, struct amdgpu_vm_update_params *params) { struct amdgpu_vm_bo_base *entry, *next; - struct amdgpu_vm *vm = params->vm; bool unlocked = params->unlocked; if (list_empty(¶ms->tlb_flush_waitlist)) return; - if (unlocked) { - spin_lock(&vm->status_lock); - list_splice_init(¶ms->tlb_flush_waitlist, &vm->pt_freed); - spin_unlock(&vm->status_lock); - schedule_work(&vm->pt_free_work); - return; - } + /* + * unlocked unmap clear page table leaves, warning to free the page entry. + */ + WARN_ON(unlocked); list_for_each_entry_safe(entry, next, ¶ms->tlb_flush_waitlist, vm_status) amdgpu_vm_pt_free(entry); @@ -900,7 +875,15 @@ int amdgpu_vm_ptes_update(struct amdgpu_vm_update_params *params, incr = (uint64_t)AMDGPU_GPU_PAGE_SIZE << shift; mask = amdgpu_vm_pt_entries_mask(adev, cursor.level); pe_start = ((cursor.pfn >> shift) & mask) * 8; - entry_end = ((uint64_t)mask + 1) << shift; + + if (cursor.level < AMDGPU_VM_PTB && params->unlocked) + /* + * MMU notifier callback unlocked unmap huge page, leave is PDE entry, + * only clear one entry. Next entry search again for PDE or PTE leave. + */ + entry_end = 1ULL << shift; + else + entry_end = ((uint64_t)mask + 1) << shift; entry_end += cursor.pfn & ~(entry_end - 1); entry_end = min(entry_end, end); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c index ff5e52025266..6da8994e0469 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c @@ -28,7 +28,6 @@ #include "amdgpu.h" #include "amdgpu_vm.h" #include "amdgpu_res_cursor.h" -#include "amdgpu_atomfirmware.h" #include "atom.h" #define AMDGPU_MAX_SG_SEGMENT_SIZE (2UL << 30) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c index 74b4349e345a..c98b6b35cfdf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c @@ -1123,10 +1123,12 @@ static int xgmi_v6_4_0_aca_bank_parser(struct aca_handle *handle, struct aca_ban if (ext_error_code != 0 && ext_error_code != 9) count = 0ULL; + bank->aca_err_type = ACA_ERROR_TYPE_UE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE, count); break; case ACA_SMU_TYPE_CE: count = ext_error_code == 6 ? count : 0ULL; + bank->aca_err_type = ACA_ERROR_TYPE_CE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_CE, count); break; default: diff --git a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c index e157d6d857b6..3c07517be09a 100644 --- a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c +++ b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c @@ -77,7 +77,8 @@ static void aqua_vanjaram_set_xcp_id(struct amdgpu_device *adev, ring->xcp_id = AMDGPU_XCP_NO_PARTITION; if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) adev->gfx.enforce_isolation[0].xcp_id = ring->xcp_id; - if (adev->xcp_mgr->mode == AMDGPU_XCP_MODE_NONE) + if ((adev->xcp_mgr->mode == AMDGPU_XCP_MODE_NONE) || + (ring->funcs->type == AMDGPU_RING_TYPE_CPER)) return; inst_mask = 1 << inst_idx; @@ -559,8 +560,11 @@ static bool __aqua_vanjaram_is_valid_mode(struct amdgpu_xcp_mgr *xcp_mgr, adev->gmc.num_mem_partitions == 4) && (num_xccs_per_xcp >= 2); case AMDGPU_CPX_PARTITION_MODE: + /* (num_xcc > 1) because 1 XCC is considered SPX, not CPX. + * (num_xcc % adev->gmc.num_mem_partitions) == 0 because + * num_compute_partitions can't be less than num_mem_partitions + */ return ((num_xcc > 1) && - (adev->gmc.num_mem_partitions == 1 || adev->gmc.num_mem_partitions == 4) && (num_xcc % adev->gmc.num_mem_partitions) == 0); default: return false; diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_dp.c b/drivers/gpu/drm/amd/amdgpu/atombios_dp.c index 622634c08c7b..521b9faab180 100644 --- a/drivers/gpu/drm/amd/amdgpu/atombios_dp.c +++ b/drivers/gpu/drm/amd/amdgpu/atombios_dp.c @@ -430,7 +430,7 @@ void amdgpu_atombios_dp_set_link_config(struct drm_connector *connector, } int amdgpu_atombios_dp_mode_valid_helper(struct drm_connector *connector, - struct drm_display_mode *mode) + const struct drm_display_mode *mode) { struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector); struct amdgpu_connector_atom_dig *dig_connector; diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_dp.h b/drivers/gpu/drm/amd/amdgpu/atombios_dp.h index f59d85eaddf0..3e24acf8133f 100644 --- a/drivers/gpu/drm/amd/amdgpu/atombios_dp.h +++ b/drivers/gpu/drm/amd/amdgpu/atombios_dp.h @@ -32,7 +32,7 @@ int amdgpu_atombios_dp_get_panel_mode(struct drm_encoder *encoder, void amdgpu_atombios_dp_set_link_config(struct drm_connector *connector, const struct drm_display_mode *mode); int amdgpu_atombios_dp_mode_valid_helper(struct drm_connector *connector, - struct drm_display_mode *mode); + const struct drm_display_mode *mode); bool amdgpu_atombios_dp_needs_link_train(struct amdgpu_connector *amdgpu_connector); void amdgpu_atombios_dp_set_rx_power_state(struct drm_connector *connector, u8 power_state); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c index 5ba263fe5512..f54617c6c071 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c @@ -40,7 +40,6 @@ #include "ivsrcid/gfx/irqsrcs_gfx_10_1.h" #include "soc15.h" -#include "soc15d.h" #include "soc15_common.h" #include "clearstate_gfx10.h" #include "v10_structs.h" @@ -3790,12 +3789,65 @@ static void gfx10_kiq_invalidate_tlbs(struct amdgpu_ring *kiq_ring, gfx_v10_0_ring_invalidate_tlbs(kiq_ring, pasid, flush_type, all_hub, 1); } +static void gfx_v10_0_kiq_reset_hw_queue(struct amdgpu_ring *kiq_ring, uint32_t queue_type, + uint32_t me_id, uint32_t pipe_id, uint32_t queue_id, + uint32_t xcc_id, uint32_t vmid) +{ + struct amdgpu_device *adev = kiq_ring->adev; + unsigned i; + uint32_t tmp; + + /* enter save mode */ + amdgpu_gfx_rlc_enter_safe_mode(adev, xcc_id); + mutex_lock(&adev->srbm_mutex); + nv_grbm_select(adev, me_id, pipe_id, queue_id, 0); + + if (queue_type == AMDGPU_RING_TYPE_COMPUTE) { + WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 0x2); + WREG32_SOC15(GC, 0, mmSPI_COMPUTE_QUEUE_RESET, 0x1); + /* wait till dequeue take effects */ + for (i = 0; i < adev->usec_timeout; i++) { + if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + if (i >= adev->usec_timeout) + dev_err(adev->dev, "fail to wait on hqd deactive\n"); + } else if (queue_type == AMDGPU_RING_TYPE_GFX) { + WREG32_SOC15(GC, 0, mmGRBM_GFX_INDEX, + (uint32_t)(0x1 << GRBM_GFX_INDEX__SE_BROADCAST_WRITES__SHIFT)); + tmp = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid); + if (pipe_id == 0) + tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE0_QUEUES, 1 << queue_id); + else + tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE1_QUEUES, 1 << queue_id); + WREG32_SOC15(GC, 0, mmCP_VMID_RESET, tmp); + + /* wait till dequeue take effects */ + for (i = 0; i < adev->usec_timeout; i++) { + if (!(RREG32_SOC15(GC, 0, mmCP_GFX_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + if (i >= adev->usec_timeout) + dev_err(adev->dev, "failed to wait on gfx hqd deactivate\n"); + } else { + dev_err(adev->dev, "reset queue_type(%d) not supported\n", queue_type); + } + + nv_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + /* exit safe mode */ + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); +} + static const struct kiq_pm4_funcs gfx_v10_0_kiq_pm4_funcs = { .kiq_set_resources = gfx10_kiq_set_resources, .kiq_map_queues = gfx10_kiq_map_queues, .kiq_unmap_queues = gfx10_kiq_unmap_queues, .kiq_query_status = gfx10_kiq_query_status, .kiq_invalidate_tlbs = gfx10_kiq_invalidate_tlbs, + .kiq_reset_hw_queue = gfx_v10_0_kiq_reset_hw_queue, .set_resources_size = 8, .map_queues_size = 7, .unmap_queues_size = 6, @@ -4701,6 +4753,8 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block) int xcc_id = 0; struct amdgpu_device *adev = ip_block->adev; + INIT_DELAYED_WORK(&adev->gfx.idle_work, amdgpu_gfx_profile_idle_work_handler); + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(10, 1, 10): case IP_VERSION(10, 1, 1): @@ -4739,6 +4793,22 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block) break; } switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + adev->gfx.cleaner_shader_ptr = gfx_10_1_10_cleaner_shader_hex; + adev->gfx.cleaner_shader_size = sizeof(gfx_10_1_10_cleaner_shader_hex); + if (adev->gfx.me_fw_version >= 101 && + adev->gfx.pfp_fw_version >= 158 && + adev->gfx.mec_fw_version >= 152) { + adev->gfx.enable_cleaner_shader = true; + r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size); + if (r) { + adev->gfx.enable_cleaner_shader = false; + dev_err(adev->dev, "Failed to initialize cleaner shader\n"); + } + } + break; case IP_VERSION(10, 3, 0): case IP_VERSION(10, 3, 2): case IP_VERSION(10, 3, 4): @@ -7467,6 +7537,8 @@ static int gfx_v10_0_hw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + cancel_delayed_work_sync(&adev->gfx.idle_work); + amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0); @@ -8431,9 +8503,9 @@ static int gfx_v10_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void gfx_v10_0_get_clockgating_state(void *handle, u64 *flags) +static void gfx_v10_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; /* AMD_CG_SUPPORT_GFX_FGCG */ @@ -9748,6 +9820,20 @@ static void gfx_v10_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring) amdgpu_ring_write(ring, 0); /* RESERVED field, programmed to zero */ } +static void gfx_v10_0_ring_begin_use(struct amdgpu_ring *ring) +{ + amdgpu_gfx_profile_ring_begin_use(ring); + + amdgpu_gfx_enforce_isolation_ring_begin_use(ring); +} + +static void gfx_v10_0_ring_end_use(struct amdgpu_ring *ring) +{ + amdgpu_gfx_profile_ring_end_use(ring); + + amdgpu_gfx_enforce_isolation_ring_end_use(ring); +} + static const struct amd_ip_funcs gfx_v10_0_ip_funcs = { .name = "gfx_v10_0", .early_init = gfx_v10_0_early_init, @@ -9823,8 +9909,8 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = { .emit_mem_sync = gfx_v10_0_emit_mem_sync, .reset = gfx_v10_0_reset_kgq, .emit_cleaner_shader = gfx_v10_0_ring_emit_cleaner_shader, - .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use, - .end_use = amdgpu_gfx_enforce_isolation_ring_end_use, + .begin_use = gfx_v10_0_ring_begin_use, + .end_use = gfx_v10_0_ring_end_use, }; static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = { @@ -9864,8 +9950,8 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = { .emit_mem_sync = gfx_v10_0_emit_mem_sync, .reset = gfx_v10_0_reset_kcq, .emit_cleaner_shader = gfx_v10_0_ring_emit_cleaner_shader, - .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use, - .end_use = amdgpu_gfx_enforce_isolation_ring_end_use, + .begin_use = gfx_v10_0_ring_begin_use, + .end_use = gfx_v10_0_ring_end_use, }; static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h index 663c2572d440..5255378af53c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h @@ -21,6 +21,41 @@ * OTHER DEALINGS IN THE SOFTWARE. */ +/* Define the cleaner shader gfx_10_1_10 */ +static const u32 gfx_10_1_10_cleaner_shader_hex[] = { + 0xb0804004, 0xbf8a0000, + 0xbf068100, 0xbf840023, + 0xbe8203b8, 0xbefc0380, + 0x7e008480, 0x7e028480, + 0x7e048480, 0x7e068480, + 0x7e088480, 0x7e0a8480, + 0x7e0c8480, 0x7e0e8480, + 0xbefc0302, 0x80828802, + 0xbf84fff5, 0xbe8203ff, + 0x80000000, 0x87020102, + 0xbf840012, 0xbefe03c1, + 0xbeff03c1, 0xd7650001, + 0x0001007f, 0xd7660001, + 0x0002027e, 0x16020288, + 0xbe8203bf, 0xbefc03c1, + 0xd9382000, 0x00020201, + 0xd9386040, 0x00040401, + 0xd70f6a01, 0x000202ff, + 0x00000400, 0x80828102, + 0xbf84fff7, 0xbefc03ff, + 0x00000068, 0xbe803080, + 0xbe813080, 0xbe823080, + 0xbe833080, 0x80fc847c, + 0xbf84fffa, 0xbeea0480, + 0xbeec0480, 0xbeee0480, + 0xbef00480, 0xbef20480, + 0xbef40480, 0xbef60480, + 0xbef80480, 0xbefa0480, + 0xbf810000, 0xbf9f0000, + 0xbf9f0000, 0xbf9f0000, + 0xbf9f0000, 0xbf9f0000, +}; + /* Define the cleaner shader gfx_10_3_0 */ static const u32 gfx_10_3_0_cleaner_shader_hex[] = { 0xb0804004, 0xbf8a0000, diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm b/drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm new file mode 100644 index 000000000000..9ba3359253c9 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm @@ -0,0 +1,126 @@ +/* SPDX-License-Identifier: MIT */ +/* + * Copyright 2025 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. + */ + +// This shader is to clean LDS, SGPRs and VGPRs. It is first 64 Dwords or 256 bytes of 256 Dwords cleaner shader. + +// GFX10.1 : Clear SGPRs, VGPRs and LDS +// Launch 32 waves per CU (16 per SIMD) as a workgroup (threadgroup) to fill every wave slot +// Waves are "wave32" and have 64 VGPRs each, which uses all 1024 VGPRs per SIMD +// Waves are launched in "CU" mode, and the workgroup shares 64KB of LDS (half of the WGP's LDS) +// It takes 2 workgroups to use all of LDS: one on each CU of the WGP +// Each wave clears SGPRs 0 - 107 +// Each wave clears VGPRs 0 - 63 +// The first wave of the workgroup clears its 64KB of LDS +// The shader starts with "S_BARRIER" to ensure SPI has launched all waves of the workgroup +// before any wave in the workgroup could end. Without this, it is possible not all SGPRs get cleared. + + +shader main + asic(GFX10.1) + type(CS) + wave_size(32) +// Note: original source code from SQ team + +// +// Create 32 waves in a threadgroup (CS waves) +// Each allocates 64 VGPRs +// The workgroup allocates all of LDS (64kbytes) +// +// Takes about 2500 clocks to run. +// (theorhetical fastest = 1024clks vgpr + 640lds = 1660 clks) +// + S_BARRIER + s_cmp_eq_u32 s0, 1 // Bit0 is set, sgpr0 is set then clear VGPRS and LDS as FW set COMPUTE_USER_DATA_0 + s_cbranch_scc0 label_0023 // Clean VGPRs and LDS if sgpr0 of wave is set, scc = (s0 == 1) + + s_mov_b32 s2, 0x00000038 // Loop 64/8=8 times (loop unrolled for performance) + s_mov_b32 m0, 0 + // + // CLEAR VGPRs + // +label_0005: + v_movreld_b32 v0, 0 + v_movreld_b32 v1, 0 + v_movreld_b32 v2, 0 + v_movreld_b32 v3, 0 + v_movreld_b32 v4, 0 + v_movreld_b32 v5, 0 + v_movreld_b32 v6, 0 + v_movreld_b32 v7, 0 + s_mov_b32 m0, s2 + s_sub_u32 s2, s2, 8 + s_cbranch_scc0 label_0005 + // + s_mov_b32 s2, 0x80000000 // Bit31 is first_wave + s_and_b32 s2, s2, s0 // sgpr0 has tg_size (first_wave) term as in ucode only COMPUTE_PGM_RSRC2.tg_size_en is set + s_cbranch_scc0 label_0023 // Clean LDS if its first wave of ThreadGroup/WorkGroup + // CLEAR LDS + // + s_mov_b32 exec_lo, 0xffffffff + s_mov_b32 exec_hi, 0xffffffff + v_mbcnt_lo_u32_b32 v1, exec_hi, 0 // Set V1 to thread-ID (0..63) + v_mbcnt_hi_u32_b32 v1, exec_lo, v1 // Set V1 to thread-ID (0..63) + v_mul_u32_u24 v1, 0x00000008, v1 // * 8, so each thread is a double-dword address (8byte) + s_mov_b32 s2, 0x00000003f // 64 loop iterations + s_mov_b32 m0, 0xffffffff + // Clear all of LDS space + // Each FirstWave of WorkGroup clears 64kbyte block + +label_001F: + ds_write2_b64 v1, v[2:3], v[2:3] offset1:32 + ds_write2_b64 v1, v[4:5], v[4:5] offset0:64 offset1:96 + v_add_co_u32 v1, vcc, 0x00000400, v1 + s_sub_u32 s2, s2, 1 + s_cbranch_scc0 label_001F + + // + // CLEAR SGPRs + // +label_0023: + s_mov_b32 m0, 0x00000068 // Loop 108/4=27 times (loop unrolled for performance) +label_sgpr_loop: + s_movreld_b32 s0, 0 + s_movreld_b32 s1, 0 + s_movreld_b32 s2, 0 + s_movreld_b32 s3, 0 + s_sub_u32 m0, m0, 4 + s_cbranch_scc0 label_sgpr_loop + + //clear vcc + s_mov_b64 vcc, 0 //clear vcc + //s_setreg_imm32_b32 hw_reg_shader_flat_scratch_lo, 0 //clear flat scratch lo SGPR + //s_setreg_imm32_b32 hw_reg_shader_flat_scratch_hi, 0 //clear flat scratch hi SGPR + s_mov_b64 ttmp0, 0 //Clear ttmp0 and ttmp1 + s_mov_b64 ttmp2, 0 //Clear ttmp2 and ttmp3 + s_mov_b64 ttmp4, 0 //Clear ttmp4 and ttmp5 + s_mov_b64 ttmp6, 0 //Clear ttmp6 and ttmp7 + s_mov_b64 ttmp8, 0 //Clear ttmp8 and ttmp9 + s_mov_b64 ttmp10, 0 //Clear ttmp10 and ttmp11 + s_mov_b64 ttmp12, 0 //Clear ttmp12 and ttmp13 + s_mov_b64 ttmp14, 0 //Clear ttmp14 and ttmp15 + + s_endpgm + +end + + diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index 56c06b72a70a..57f1e2b50e5a 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -29,7 +29,6 @@ #include "amdgpu_gfx.h" #include "amdgpu_psp.h" #include "amdgpu_smu.h" -#include "amdgpu_atomfirmware.h" #include "imu_v11_0.h" #include "soc21.h" #include "nvd.h" @@ -42,7 +41,6 @@ #include "ivsrcid/gfx/irqsrcs_gfx_11_0_0.h" #include "soc15.h" -#include "soc15d.h" #include "clearstate_gfx11.h" #include "v11_structs.h" #include "gfx_v11_0.h" @@ -98,6 +96,10 @@ MODULE_FIRMWARE("amdgpu/gc_11_5_2_pfp.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_2_me.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_2_mec.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_2_rlc.bin"); +MODULE_FIRMWARE("amdgpu/gc_11_5_3_pfp.bin"); +MODULE_FIRMWARE("amdgpu/gc_11_5_3_me.bin"); +MODULE_FIRMWARE("amdgpu/gc_11_5_3_mec.bin"); +MODULE_FIRMWARE("amdgpu/gc_11_5_3_rlc.bin"); static const struct amdgpu_hwip_reg_entry gc_reg_list_11_0[] = { SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS), @@ -1087,6 +1089,7 @@ static int gfx_v11_0_gpu_early_init(struct amdgpu_device *adev) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): adev->gfx.config.max_hw_contexts = 8; adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; adev->gfx.config.sc_prim_fifo_size_backend = 0x100; @@ -1552,6 +1555,8 @@ static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block) int xcc_id = 0; struct amdgpu_device *adev = ip_block->adev; + INIT_DELAYED_WORK(&adev->gfx.idle_work, amdgpu_gfx_profile_idle_work_handler); + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(11, 0, 0): case IP_VERSION(11, 0, 2): @@ -1568,6 +1573,7 @@ static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): adev->gfx.me.num_me = 1; adev->gfx.me.num_pipe_per_me = 1; adev->gfx.me.num_queue_per_pipe = 1; @@ -2926,7 +2932,8 @@ static int gfx_v11_0_wait_for_rlc_autoload_complete(struct amdgpu_device *adev) IP_VERSION(11, 0, 4) || amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 5, 0) || amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 5, 1) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 5, 2)) + amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 5, 2) || + amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 5, 3)) bootload_status = RREG32_SOC15(GC, 0, regRLC_RLCS_BOOTLOAD_STATUS_gc_11_0_1); else @@ -4734,6 +4741,8 @@ static int gfx_v11_0_hw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + cancel_delayed_work_sync(&adev->gfx.idle_work); + amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0); @@ -5448,6 +5457,7 @@ static void gfx_v11_cntl_power_gating(struct amdgpu_device *adev, bool enable) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): WREG32_SOC15(GC, 0, regRLC_PG_DELAY_3, RLC_PG_DELAY_3_DEFAULT_GC_11_0_1); break; default: @@ -5485,6 +5495,7 @@ static int gfx_v11_0_set_powergating_state(struct amdgpu_ip_block *ip_block, case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): if (!enable) amdgpu_gfx_off_ctrl(adev, false); @@ -5518,6 +5529,7 @@ static int gfx_v11_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): gfx_v11_0_update_gfx_clock_gating(adev, state == AMD_CG_STATE_GATE); break; @@ -5528,9 +5540,9 @@ static int gfx_v11_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void gfx_v11_0_get_clockgating_state(void *handle, u64 *flags) +static void gfx_v11_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; /* AMD_CG_SUPPORT_GFX_MGCG */ @@ -6826,6 +6838,20 @@ static void gfx_v11_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring) amdgpu_ring_write(ring, 0); /* RESERVED field, programmed to zero */ } +static void gfx_v11_0_ring_begin_use(struct amdgpu_ring *ring) +{ + amdgpu_gfx_profile_ring_begin_use(ring); + + amdgpu_gfx_enforce_isolation_ring_begin_use(ring); +} + +static void gfx_v11_0_ring_end_use(struct amdgpu_ring *ring) +{ + amdgpu_gfx_profile_ring_end_use(ring); + + amdgpu_gfx_enforce_isolation_ring_end_use(ring); +} + static const struct amd_ip_funcs gfx_v11_0_ip_funcs = { .name = "gfx_v11_0", .early_init = gfx_v11_0_early_init, @@ -6900,8 +6926,8 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_gfx = { .emit_mem_sync = gfx_v11_0_emit_mem_sync, .reset = gfx_v11_0_reset_kgq, .emit_cleaner_shader = gfx_v11_0_ring_emit_cleaner_shader, - .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use, - .end_use = amdgpu_gfx_enforce_isolation_ring_end_use, + .begin_use = gfx_v11_0_ring_begin_use, + .end_use = gfx_v11_0_ring_end_use, }; static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_compute = { @@ -6942,8 +6968,8 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_compute = { .emit_mem_sync = gfx_v11_0_emit_mem_sync, .reset = gfx_v11_0_reset_kcq, .emit_cleaner_shader = gfx_v11_0_ring_emit_cleaner_shader, - .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use, - .end_use = amdgpu_gfx_enforce_isolation_ring_end_use, + .begin_use = gfx_v11_0_ring_begin_use, + .end_use = gfx_v11_0_ring_end_use, }; static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c index 2523221a2519..47490309045f 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c @@ -29,7 +29,6 @@ #include "amdgpu_gfx.h" #include "amdgpu_psp.h" #include "amdgpu_smu.h" -#include "amdgpu_atomfirmware.h" #include "imu_v12_0.h" #include "soc24.h" #include "nvd.h" @@ -40,7 +39,6 @@ #include "ivsrcid/gfx/irqsrcs_gfx_11_0_0.h" #include "soc15.h" -#include "soc15d.h" #include "clearstate_gfx12.h" #include "v12_structs.h" #include "gfx_v12_0.h" @@ -1331,6 +1329,8 @@ static int gfx_v12_0_sw_init(struct amdgpu_ip_block *ip_block) int xcc_id = 0; struct amdgpu_device *adev = ip_block->adev; + INIT_DELAYED_WORK(&adev->gfx.idle_work, amdgpu_gfx_profile_idle_work_handler); + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(12, 0, 0): case IP_VERSION(12, 0, 1): @@ -3648,6 +3648,8 @@ static int gfx_v12_0_hw_fini(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; uint32_t tmp; + cancel_delayed_work_sync(&adev->gfx.idle_work); + amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0); @@ -4147,9 +4149,9 @@ static int gfx_v12_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void gfx_v12_0_get_clockgating_state(void *handle, u64 *flags) +static void gfx_v12_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; /* AMD_CG_SUPPORT_GFX_MGCG */ @@ -5280,6 +5282,20 @@ static int gfx_v12_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid) return amdgpu_ring_test_ring(ring); } +static void gfx_v12_0_ring_begin_use(struct amdgpu_ring *ring) +{ + amdgpu_gfx_profile_ring_begin_use(ring); + + amdgpu_gfx_enforce_isolation_ring_begin_use(ring); +} + +static void gfx_v12_0_ring_end_use(struct amdgpu_ring *ring) +{ + amdgpu_gfx_profile_ring_end_use(ring); + + amdgpu_gfx_enforce_isolation_ring_end_use(ring); +} + static const struct amd_ip_funcs gfx_v12_0_ip_funcs = { .name = "gfx_v12_0", .early_init = gfx_v12_0_early_init, @@ -5345,8 +5361,8 @@ static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_gfx = { .emit_mem_sync = gfx_v12_0_emit_mem_sync, .reset = gfx_v12_0_reset_kgq, .emit_cleaner_shader = gfx_v12_0_ring_emit_cleaner_shader, - .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use, - .end_use = amdgpu_gfx_enforce_isolation_ring_end_use, + .begin_use = gfx_v12_0_ring_begin_use, + .end_use = gfx_v12_0_ring_end_use, }; static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_compute = { @@ -5384,8 +5400,8 @@ static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_compute = { .emit_mem_sync = gfx_v12_0_emit_mem_sync, .reset = gfx_v12_0_reset_kcq, .emit_cleaner_shader = gfx_v12_0_ring_emit_cleaner_shader, - .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use, - .end_use = amdgpu_gfx_enforce_isolation_ring_end_use, + .begin_use = gfx_v12_0_ring_begin_use, + .end_use = gfx_v12_0_ring_end_use, }; static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 6a025438f9d0..6add76ef75e8 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -5452,9 +5452,9 @@ static int gfx_v8_0_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void gfx_v8_0_get_clockgating_state(void *handle, u64 *flags) +static void gfx_v8_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index fa572b40989e..e144bce938d5 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -5241,7 +5241,7 @@ static int gfx_v9_0_set_powergating_state(struct amdgpu_ip_block *ip_block, case IP_VERSION(9, 1, 0): case IP_VERSION(9, 3, 0): if (!enable) - amdgpu_gfx_off_ctrl(adev, false); + amdgpu_gfx_off_ctrl_immediate(adev, false); if (adev->pg_flags & AMD_PG_SUPPORT_RLC_SMU_HS) { gfx_v9_0_enable_sck_slow_down_on_power_up(adev, true); @@ -5263,10 +5263,10 @@ static int gfx_v9_0_set_powergating_state(struct amdgpu_ip_block *ip_block, gfx_v9_0_update_gfx_mg_power_gating(adev, enable); if (enable) - amdgpu_gfx_off_ctrl(adev, true); + amdgpu_gfx_off_ctrl_immediate(adev, true); break; case IP_VERSION(9, 2, 1): - amdgpu_gfx_off_ctrl(adev, enable); + amdgpu_gfx_off_ctrl_immediate(adev, enable); break; default: break; @@ -5301,9 +5301,9 @@ static int gfx_v9_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void gfx_v9_0_get_clockgating_state(void *handle, u64 *flags) +static void gfx_v9_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) @@ -7437,6 +7437,38 @@ static void gfx_v9_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring) amdgpu_ring_write(ring, 0); /* RESERVED field, programmed to zero */ } +static void gfx_v9_0_ring_begin_use_compute(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_ip_block *gfx_block = + amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX); + + amdgpu_gfx_enforce_isolation_ring_begin_use(ring); + + /* Raven and PCO APUs seem to have stability issues + * with compute and gfxoff and gfx pg. Disable gfx pg during + * submission and allow again afterwards. + */ + if (gfx_block && amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 1, 0)) + gfx_v9_0_set_powergating_state(gfx_block, AMD_PG_STATE_UNGATE); +} + +static void gfx_v9_0_ring_end_use_compute(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_ip_block *gfx_block = + amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX); + + /* Raven and PCO APUs seem to have stability issues + * with compute and gfxoff and gfx pg. Disable gfx pg during + * submission and allow again afterwards. + */ + if (gfx_block && amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 1, 0)) + gfx_v9_0_set_powergating_state(gfx_block, AMD_PG_STATE_GATE); + + amdgpu_gfx_enforce_isolation_ring_end_use(ring); +} + static const struct amd_ip_funcs gfx_v9_0_ip_funcs = { .name = "gfx_v9_0", .early_init = gfx_v9_0_early_init, @@ -7613,8 +7645,8 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = { .emit_wave_limit = gfx_v9_0_emit_wave_limit, .reset = gfx_v9_0_reset_kcq, .emit_cleaner_shader = gfx_v9_0_ring_emit_cleaner_shader, - .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use, - .end_use = amdgpu_gfx_enforce_isolation_ring_end_use, + .begin_use = gfx_v9_0_ring_begin_use_compute, + .end_use = gfx_v9_0_ring_end_use_compute, }; static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c index f53b379d8971..6028afd81690 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c @@ -27,7 +27,6 @@ #include "amdgpu_gfx.h" #include "soc15.h" #include "soc15d.h" -#include "amdgpu_atomfirmware.h" #include "amdgpu_pm.h" #include "gc/gc_9_4_1_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c index 2ba185875baa..c88564de50cd 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c @@ -891,10 +891,12 @@ static int gfx_v9_4_3_aca_bank_parser(struct aca_handle *handle, switch (type) { case ACA_SMU_TYPE_UE: + bank->aca_err_type = ACA_ERROR_TYPE_UE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE, 1ULL); break; case ACA_SMU_TYPE_CE: + bank->aca_err_type = ACA_ERROR_TYPE_CE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_CE, ACA_REG__MISC0__ERRCNT(misc0)); break; @@ -942,21 +944,12 @@ static int gfx_v9_4_3_gpu_early_init(struct amdgpu_device *adev) adev->gfx.funcs = &gfx_v9_4_3_gfx_funcs; adev->gfx.ras = &gfx_v9_4_3_ras; - switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { - case IP_VERSION(9, 4, 3): - case IP_VERSION(9, 4, 4): - case IP_VERSION(9, 5, 0): - adev->gfx.config.max_hw_contexts = 8; - adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; - adev->gfx.config.sc_prim_fifo_size_backend = 0x100; - adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; - adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0; - gb_addr_config = RREG32_SOC15(GC, GET_INST(GC, 0), regGB_ADDR_CONFIG); - break; - default: - BUG(); - break; - } + adev->gfx.config.max_hw_contexts = 8; + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0; + gb_addr_config = RREG32_SOC15(GC, GET_INST(GC, 0), regGB_ADDR_CONFIG); adev->gfx.config.gb_addr_config = gb_addr_config; @@ -2795,22 +2788,16 @@ static int gfx_v9_4_3_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; num_xcc = NUM_XCC(adev->gfx.xcc_mask); - switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { - case IP_VERSION(9, 4, 3): - case IP_VERSION(9, 4, 4): - for (i = 0; i < num_xcc; i++) - gfx_v9_4_3_xcc_update_gfx_clock_gating( - adev, state == AMD_CG_STATE_GATE, i); - break; - default: - break; - } + for (i = 0; i < num_xcc; i++) + gfx_v9_4_3_xcc_update_gfx_clock_gating( + adev, state == AMD_CG_STATE_GATE, i); + return 0; } -static void gfx_v9_4_3_get_clockgating_state(void *handle, u64 *flags) +static void gfx_v9_4_3_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) @@ -4867,34 +4854,13 @@ static void gfx_v9_4_3_set_rlc_funcs(struct amdgpu_device *adev) static void gfx_v9_4_3_set_gds_init(struct amdgpu_device *adev) { - /* init asci gds info */ - switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { - case IP_VERSION(9, 4, 3): - case IP_VERSION(9, 4, 4): - case IP_VERSION(9, 5, 0): - /* 9.4.3 removed all the GDS internal memory, - * only support GWS opcode in kernel, like barrier - * semaphore.etc */ - adev->gds.gds_size = 0; - break; - default: - adev->gds.gds_size = 0x10000; - break; - } - - switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { - case IP_VERSION(9, 4, 3): - case IP_VERSION(9, 4, 4): - case IP_VERSION(9, 5, 0): - /* deprecated for 9.4.3, no usage at all */ - adev->gds.gds_compute_max_wave_id = 0; - break; - default: - /* this really depends on the chip */ - adev->gds.gds_compute_max_wave_id = 0x7ff; - break; - } + /* 9.4.3 variants removed all the GDS internal memory, + * only support GWS opcode in kernel, like barrier + * semaphore.etc */ + /* init asic gds info */ + adev->gds.gds_size = 0; + adev->gds.gds_compute_max_wave_id = 0; adev->gds.gws_size = 64; adev->gds.oa_size = 16; } diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c index 5470cef7e9bd..cb25f7f0dfc1 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c @@ -313,6 +313,16 @@ gfxhub_v1_2_xcc_disable_identity_aperture(struct amdgpu_device *adev, } } +static inline bool +gfxhub_v1_2_per_process_xnack_support(struct amdgpu_device *adev) +{ + /* + * TODO: Check if this function is really needed, so far only 9.4.3 + * variants use GFXHUB 1.2 + */ + return !!adev->aid_mask; +} + static void gfxhub_v1_2_xcc_setup_vmid_config(struct amdgpu_device *adev, uint32_t xcc_mask) { @@ -355,7 +365,7 @@ static void gfxhub_v1_2_xcc_setup_vmid_config(struct amdgpu_device *adev, PAGE_TABLE_BLOCK_SIZE, block_size); /* Send no-retry XNACK on fault to suppress VM fault storm. - * On 9.4.2 and 9.4.3, XNACK can be enabled in + * On 9.4.3 variants, XNACK can be enabled in * the SQ per-process. * Retry faults need to be enabled for that to work. */ @@ -363,14 +373,8 @@ static void gfxhub_v1_2_xcc_setup_vmid_config(struct amdgpu_device *adev, tmp, VM_CONTEXT1_CNTL, RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, !adev->gmc.noretry || - amdgpu_ip_version(adev, GC_HWIP, 0) == - IP_VERSION(9, 4, 2) || - amdgpu_ip_version(adev, GC_HWIP, 0) == - IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == - IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == - IP_VERSION(9, 5, 0)); + gfxhub_v1_2_per_process_xnack_support( + adev)); WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_CONTEXT1_CNTL, i * hub->ctx_distance, tmp); WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c index 9bedca9a79c6..edf6cf42f141 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c @@ -1115,9 +1115,9 @@ static int gmc_v10_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return athub_v2_0_set_clockgating(adev, state); } -static void gmc_v10_0_get_clockgating_state(void *handle, u64 *flags) +static void gmc_v10_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 3) || amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 4)) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c index 72751ab4c766..f86d0650a05e 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c @@ -579,6 +579,7 @@ static void gmc_v11_0_set_mmhub_funcs(struct amdgpu_device *adev) break; case IP_VERSION(3, 3, 0): case IP_VERSION(3, 3, 1): + case IP_VERSION(3, 3, 2): adev->mmhub.funcs = &mmhub_v3_3_funcs; break; default: @@ -596,6 +597,7 @@ static void gmc_v11_0_set_gfxhub_funcs(struct amdgpu_device *adev) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): adev->gfxhub.funcs = &gfxhub_v11_5_0_funcs; break; default: @@ -759,6 +761,7 @@ static int gmc_v11_0_sw_init(struct amdgpu_ip_block *ip_block) case IP_VERSION(11, 5, 0): case IP_VERSION(11, 5, 1): case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask); set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask); /* @@ -1009,9 +1012,9 @@ static int gmc_v11_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return athub_v3_0_set_clockgating(adev, state); } -static void gmc_v11_0_get_clockgating_state(void *handle, u64 *flags) +static void gmc_v11_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; adev->mmhub.funcs->get_clockgating(adev, flags); diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c index b749f1c3f6a9..c6d45d0fb9d1 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c @@ -1009,9 +1009,9 @@ static int gmc_v12_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return athub_v4_1_0_set_clockgating(adev, state); } -static void gmc_v12_0_get_clockgating_state(void *handle, u64 *flags) +static void gmc_v12_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; adev->mmhub.funcs->get_clockgating(adev, flags); diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index d06585207c33..744081652d42 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -1686,9 +1686,9 @@ static int gmc_v8_0_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void gmc_v8_0_get_clockgating_state(void *handle, u64 *flags) +static void gmc_v8_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index 291549765c38..a80f3e2bcba8 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -411,6 +411,11 @@ static const uint32_t ecc_umc_mcumc_ctrl_mask_addrs[] = { (0x001d43e0 + 0x00001800), }; +static inline bool gmc_v9_0_is_multi_chiplet(struct amdgpu_device *adev) +{ + return !!adev->aid_mask; +} + static int gmc_v9_0_ecc_interrupt_state(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned int type, @@ -647,9 +652,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, addr, entry->client_id, soc15_ih_clientid_name[entry->client_id]); - if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) + if (gmc_v9_0_is_multi_chiplet(adev)) dev_err(adev->dev, " cookie node_id %d fault from die %s%d%s\n", node_id, node_id % 4 == 3 ? "RSV" : "AID", node_id / 4, node_id % 4 == 1 ? ".XCD0" : node_id % 4 == 2 ? ".XCD1" : ""); @@ -798,9 +801,7 @@ static bool gmc_v9_0_use_invalidate_semaphore(struct amdgpu_device *adev, uint32_t vmhub) { if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) + gmc_v9_0_is_multi_chiplet(adev)) return false; return ((vmhub == AMDGPU_MMHUB0(0) || @@ -1278,9 +1279,8 @@ static void gmc_v9_0_override_vm_pte_flags(struct amdgpu_device *adev, /* Only GFX 9.4.3 APUs associate GPUs with NUMA nodes. Local system * memory can use more efficient MTYPEs. */ - if (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 3) && - amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 4) && - amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 5, 0)) + if (!(adev->flags & AMD_IS_APU) || + amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 3)) return; /* Only direct-mapped memory allows us to determine the NUMA node from @@ -1498,13 +1498,13 @@ static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev) adev->umc.channel_idx_tbl = &umc_v6_7_channel_idx_tbl_second[0][0]; break; case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 5, 0): adev->umc.max_ras_err_cnt_per_query = UMC_V12_0_TOTAL_CHANNEL_NUM(adev) * UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL; adev->umc.channel_inst_num = UMC_V12_0_CHANNEL_INSTANCE_NUM; adev->umc.umc_inst_num = UMC_V12_0_UMC_INSTANCE_NUM; adev->umc.node_inst_num /= UMC_V12_0_UMC_INSTANCE_NUM; adev->umc.channel_offs = UMC_V12_0_PER_CHANNEL_OFFSET; - adev->umc.active_mask = adev->aid_mask; adev->umc.retire_unit = UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL; if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) adev->umc.ras = &umc_v12_0_ras; @@ -1524,6 +1524,7 @@ static void gmc_v9_0_set_mmhub_funcs(struct amdgpu_device *adev) adev->mmhub.funcs = &mmhub_v1_7_funcs; break; case IP_VERSION(1, 8, 0): + case IP_VERSION(1, 8, 1): adev->mmhub.funcs = &mmhub_v1_8_funcs; break; default: @@ -1556,9 +1557,7 @@ static void gmc_v9_0_set_mmhub_ras_funcs(struct amdgpu_device *adev) static void gmc_v9_0_set_gfxhub_funcs(struct amdgpu_device *adev) { - if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) + if (gmc_v9_0_is_multi_chiplet(adev)) adev->gfxhub.funcs = &gfxhub_v1_2_funcs; else adev->gfxhub.funcs = &gfxhub_v1_0_funcs; @@ -1595,23 +1594,39 @@ static void gmc_v9_0_set_xgmi_ras_funcs(struct amdgpu_device *adev) static void gmc_v9_0_init_nps_details(struct amdgpu_device *adev) { + enum amdgpu_memory_partition mode; + uint32_t supp_modes; + int i; + adev->gmc.supported_nps_modes = 0; if (amdgpu_sriov_vf(adev) || (adev->flags & AMD_IS_APU)) return; - /*TODO: Check PSP version also which supports NPS switch. Otherwise keep + mode = gmc_v9_0_get_memory_partition(adev, &supp_modes); + + /* Mode detected by hardware and supported modes available */ + if ((mode != UNKNOWN_MEMORY_PARTITION_MODE) && supp_modes) { + for (i = AMDGPU_NPS1_PARTITION_MODE; + supp_modes && i <= AMDGPU_NPS8_PARTITION_MODE; i++) { + if (supp_modes & BIT(i - 1)) + adev->gmc.supported_nps_modes |= BIT(i); + supp_modes &= supp_modes - 1; + } + } else { + /*TODO: Check PSP version also which supports NPS switch. Otherwise keep * supported modes as 0. */ - switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { - case IP_VERSION(9, 4, 3): - case IP_VERSION(9, 4, 4): - adev->gmc.supported_nps_modes = - BIT(AMDGPU_NPS1_PARTITION_MODE) | - BIT(AMDGPU_NPS4_PARTITION_MODE); - break; - default: - break; + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { + case IP_VERSION(9, 4, 3): + case IP_VERSION(9, 4, 4): + adev->gmc.supported_nps_modes = + BIT(AMDGPU_NPS1_PARTITION_MODE) | + BIT(AMDGPU_NPS4_PARTITION_MODE); + break; + default: + break; + } } } @@ -1625,9 +1640,7 @@ static int gmc_v9_0_early_init(struct amdgpu_ip_block *ip_block) */ if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 0) || amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) + gmc_v9_0_is_multi_chiplet(adev)) adev->gmc.xgmi.supported = true; if (amdgpu_ip_version(adev, XGMI_HWIP, 0) == IP_VERSION(6, 1, 0)) { @@ -1636,8 +1649,7 @@ static int gmc_v9_0_early_init(struct amdgpu_ip_block *ip_block) adev->smuio.funcs->is_host_gpu_xgmi_supported(adev); } - if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)) { + if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3)) { enum amdgpu_pkg_type pkg_type = adev->smuio.funcs->get_pkg_type(adev); /* On GFXIP 9.4.3. APU, there is no physical VRAM domain present @@ -2078,9 +2090,7 @@ static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block) spin_lock_init(&adev->gmc.invalidate_lock); - if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) { + if (gmc_v9_0_is_multi_chiplet(adev)) { gmc_v9_4_3_init_vram_info(adev); } else if (!adev->bios) { if (adev->flags & AMD_IS_APU) { @@ -2230,9 +2240,7 @@ static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block) amdgpu_gmc_get_vbios_allocations(adev); - if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) { + if (gmc_v9_0_is_multi_chiplet(adev)) { r = gmc_v9_0_init_mem_ranges(adev); if (r) return r; @@ -2261,9 +2269,7 @@ static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block) adev->vm_manager.first_kfd_vmid = (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1) || amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) ? + gmc_v9_0_is_multi_chiplet(adev)) ? 3 : 8; @@ -2275,9 +2281,7 @@ static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block) if (r) return r; - if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) + if (gmc_v9_0_is_multi_chiplet(adev)) amdgpu_gmc_sysfs_init(adev); return 0; @@ -2287,9 +2291,7 @@ static int gmc_v9_0_sw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)) + if (gmc_v9_0_is_multi_chiplet(adev)) amdgpu_gmc_sysfs_fini(adev); amdgpu_gmc_ras_fini(adev); @@ -2571,9 +2573,9 @@ static int gmc_v9_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void gmc_v9_0_get_clockgating_state(void *handle, u64 *flags) +static void gmc_v9_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; adev->mmhub.funcs->get_clockgating(adev, flags); diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c index 194026e9be33..f1dc13b3ab38 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "hdp_v4_0.h" #include "amdgpu_ras.h" diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c index d3962d469088..43195c079748 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "hdp_v5_0.h" #include "hdp/hdp_5_0_0_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c index f52552c5fa27..fcb8dd2876bc 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "hdp_v5_2.h" #include "hdp/hdp_5_2_1_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c index 6948fe9956ce..a88d25a06c29 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "hdp_v6_0.h" #include "hdp/hdp_6_0_0_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c index 63820329f67e..49f7eb4fbd11 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "hdp_v7_0.h" #include "hdp/hdp_7_0_0_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c index f8a485164437..7198ddfaa8f4 100644 --- a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c @@ -768,9 +768,9 @@ static int ih_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void ih_v6_0_get_clockgating_state(void *handle, u64 *flags) +static void ih_v6_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; if (!RREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL)) *flags |= AMD_CG_SUPPORT_IH_CG; diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c index dd0042efceec..342b166c136d 100644 --- a/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c +++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c @@ -749,9 +749,9 @@ static int ih_v6_1_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void ih_v6_1_get_clockgating_state(void *handle, u64 *flags) +static void ih_v6_1_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; if (!RREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL)) *flags |= AMD_CG_SUPPORT_IH_CG; diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c b/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c index 8f9b15c171f3..71c1c77035e0 100644 --- a/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c @@ -739,9 +739,9 @@ static int ih_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void ih_v7_0_get_clockgating_state(void *handle, u64 *flags) +static void ih_v7_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; if (!RREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL)) *flags |= AMD_CG_SUPPORT_IH_CG; diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c index aeca5c08ea2f..cfa91d709d49 100644 --- a/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c @@ -39,6 +39,7 @@ MODULE_FIRMWARE("amdgpu/gc_11_0_4_imu.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_0_imu.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_1_imu.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_2_imu.bin"); +MODULE_FIRMWARE("amdgpu/gc_11_5_3_imu.bin"); static int imu_v11_0_init_microcode(struct amdgpu_device *adev) { diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c index 7c9251c03815..75843a0e3bfb 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c @@ -33,6 +33,22 @@ #include "vcn/vcn_2_0_0_sh_mask.h" #include "ivsrcid/vcn/irqsrcs_vcn_2_0.h" +static const struct amdgpu_hwip_reg_entry jpeg_reg_list_2_0[] = { + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_POWER_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_INT_STAT), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_CNTL), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_SIZE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_ADDR_MODE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_GFX10_ADDR_CONFIG), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_Y_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_UV_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_UV_PITCH), +}; + static void jpeg_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev); static void jpeg_v2_0_set_irq_funcs(struct amdgpu_device *adev); static int jpeg_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block, @@ -98,7 +114,14 @@ static int jpeg_v2_0_sw_init(struct amdgpu_ip_block *ip_block) adev->jpeg.internal.jpeg_pitch[0] = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH); - return 0; + r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_2_0, ARRAY_SIZE(jpeg_reg_list_2_0)); + if (r) + return r; + + adev->jpeg.supported_reset = AMDGPU_RESET_TYPE_PER_QUEUE; + r = amdgpu_jpeg_sysfs_reset_mask_init(adev); + + return r; } /** @@ -117,6 +140,8 @@ static int jpeg_v2_0_sw_fini(struct amdgpu_ip_block *ip_block) if (r) return r; + amdgpu_jpeg_sysfs_reset_mask_fini(adev); + r = amdgpu_jpeg_sw_fini(adev); return r; @@ -739,6 +764,13 @@ static int jpeg_v2_0_process_interrupt(struct amdgpu_device *adev, return 0; } +static int jpeg_v2_0_ring_reset(struct amdgpu_ring *ring, unsigned int vmid) +{ + jpeg_v2_0_stop(ring->adev); + jpeg_v2_0_start(ring->adev); + return amdgpu_ring_test_helper(ring); +} + static const struct amd_ip_funcs jpeg_v2_0_ip_funcs = { .name = "jpeg_v2_0", .early_init = jpeg_v2_0_early_init, @@ -752,6 +784,8 @@ static const struct amd_ip_funcs jpeg_v2_0_ip_funcs = { .wait_for_idle = jpeg_v2_0_wait_for_idle, .set_clockgating_state = jpeg_v2_0_set_clockgating_state, .set_powergating_state = jpeg_v2_0_set_powergating_state, + .dump_ip_state = amdgpu_jpeg_dump_ip_state, + .print_ip_state = amdgpu_jpeg_print_ip_state, }; static const struct amdgpu_ring_funcs jpeg_v2_0_dec_ring_vm_funcs = { @@ -782,6 +816,7 @@ static const struct amdgpu_ring_funcs jpeg_v2_0_dec_ring_vm_funcs = { .emit_wreg = jpeg_v2_0_dec_ring_emit_wreg, .emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = jpeg_v2_0_ring_reset, }; static void jpeg_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c index 11f6af2646e7..0a2c1dee2430 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c @@ -36,6 +36,22 @@ #define JPEG25_MAX_HW_INSTANCES_ARCTURUS 2 +static const struct amdgpu_hwip_reg_entry jpeg_reg_list_2_5[] = { + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_POWER_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_INT_STAT), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_CNTL), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_SIZE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_ADDR_MODE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_GFX10_ADDR_CONFIG), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_Y_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_UV_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_UV_PITCH), +}; + static void jpeg_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev); static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev); static int jpeg_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block, @@ -147,7 +163,14 @@ static int jpeg_v2_5_sw_init(struct amdgpu_ip_block *ip_block) if (r) return r; - return 0; + r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_2_5, ARRAY_SIZE(jpeg_reg_list_2_5)); + if (r) + return r; + + adev->jpeg.supported_reset = AMDGPU_RESET_TYPE_PER_QUEUE; + r = amdgpu_jpeg_sysfs_reset_mask_init(adev); + + return r; } /** @@ -166,6 +189,8 @@ static int jpeg_v2_5_sw_fini(struct amdgpu_ip_block *ip_block) if (r) return r; + amdgpu_jpeg_sysfs_reset_mask_fini(adev); + r = amdgpu_jpeg_sw_fini(adev); return r; @@ -310,6 +335,44 @@ static void jpeg_v2_5_enable_clock_gating(struct amdgpu_device *adev, int inst) WREG32_SOC15(JPEG, inst, mmJPEG_CGC_GATE, data); } +static void jpeg_v2_5_start_inst(struct amdgpu_device *adev, int i) +{ + struct amdgpu_ring *ring = adev->jpeg.inst[i].ring_dec; + /* disable anti hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS), 0, + ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); + + /* JPEG disable CGC */ + jpeg_v2_5_disable_clock_gating(adev, i); + + /* MJPEG global tiling registers */ + WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX8_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* enable JMI channel */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL), 0, + ~UVD_JMI_CNTL__SOFT_RESET_MASK); + + /* enable System Interrupt for JRBC */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmJPEG_SYS_INT_EN), + JPEG_SYS_INT_EN__DJRBC_MASK, + ~JPEG_SYS_INT_EN__DJRBC_MASK); + + WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_VMID, 0); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L)); + WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW, + lower_32_bits(ring->gpu_addr)); + WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH, + upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_RPTR, 0); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_WPTR, 0); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, 0x00000002L); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_SIZE, ring->ring_size / 4); + ring->wptr = RREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_WPTR); +} + /** * jpeg_v2_5_start - start JPEG block * @@ -319,52 +382,33 @@ static void jpeg_v2_5_enable_clock_gating(struct amdgpu_device *adev, int inst) */ static int jpeg_v2_5_start(struct amdgpu_device *adev) { - struct amdgpu_ring *ring; int i; for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { if (adev->jpeg.harvest_config & (1 << i)) continue; + jpeg_v2_5_start_inst(adev, i); - ring = adev->jpeg.inst[i].ring_dec; - /* disable anti hang mechanism */ - WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS), 0, - ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); - - /* JPEG disable CGC */ - jpeg_v2_5_disable_clock_gating(adev, i); - - /* MJPEG global tiling registers */ - WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX8_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX10_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - - /* enable JMI channel */ - WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL), 0, - ~UVD_JMI_CNTL__SOFT_RESET_MASK); - - /* enable System Interrupt for JRBC */ - WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmJPEG_SYS_INT_EN), - JPEG_SYS_INT_EN__DJRBC_MASK, - ~JPEG_SYS_INT_EN__DJRBC_MASK); - - WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_VMID, 0); - WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L)); - WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW, - lower_32_bits(ring->gpu_addr)); - WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH, - upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_RPTR, 0); - WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_WPTR, 0); - WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, 0x00000002L); - WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_SIZE, ring->ring_size / 4); - ring->wptr = RREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_WPTR); } return 0; } +static void jpeg_v2_5_stop_inst(struct amdgpu_device *adev, int i) +{ + /* reset JMI */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL), + UVD_JMI_CNTL__SOFT_RESET_MASK, + ~UVD_JMI_CNTL__SOFT_RESET_MASK); + + jpeg_v2_5_enable_clock_gating(adev, i); + + /* enable anti hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS), + UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK, + ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); +} + /** * jpeg_v2_5_stop - stop JPEG block * @@ -379,18 +423,7 @@ static int jpeg_v2_5_stop(struct amdgpu_device *adev) for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { if (adev->jpeg.harvest_config & (1 << i)) continue; - - /* reset JMI */ - WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL), - UVD_JMI_CNTL__SOFT_RESET_MASK, - ~UVD_JMI_CNTL__SOFT_RESET_MASK); - - jpeg_v2_5_enable_clock_gating(adev, i); - - /* enable anti hang mechanism */ - WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS), - UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK, - ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); + jpeg_v2_5_stop_inst(adev, i); } return 0; @@ -610,6 +643,13 @@ static int jpeg_v2_5_process_interrupt(struct amdgpu_device *adev, return 0; } +static int jpeg_v2_5_ring_reset(struct amdgpu_ring *ring, unsigned int vmid) +{ + jpeg_v2_5_stop_inst(ring->adev, ring->me); + jpeg_v2_5_start_inst(ring->adev, ring->me); + return amdgpu_ring_test_helper(ring); +} + static const struct amd_ip_funcs jpeg_v2_5_ip_funcs = { .name = "jpeg_v2_5", .early_init = jpeg_v2_5_early_init, @@ -623,6 +663,8 @@ static const struct amd_ip_funcs jpeg_v2_5_ip_funcs = { .wait_for_idle = jpeg_v2_5_wait_for_idle, .set_clockgating_state = jpeg_v2_5_set_clockgating_state, .set_powergating_state = jpeg_v2_5_set_powergating_state, + .dump_ip_state = amdgpu_jpeg_dump_ip_state, + .print_ip_state = amdgpu_jpeg_print_ip_state, }; static const struct amd_ip_funcs jpeg_v2_6_ip_funcs = { @@ -638,6 +680,8 @@ static const struct amd_ip_funcs jpeg_v2_6_ip_funcs = { .wait_for_idle = jpeg_v2_5_wait_for_idle, .set_clockgating_state = jpeg_v2_5_set_clockgating_state, .set_powergating_state = jpeg_v2_5_set_powergating_state, + .dump_ip_state = amdgpu_jpeg_dump_ip_state, + .print_ip_state = amdgpu_jpeg_print_ip_state, }; static const struct amdgpu_ring_funcs jpeg_v2_5_dec_ring_vm_funcs = { @@ -668,6 +712,7 @@ static const struct amdgpu_ring_funcs jpeg_v2_5_dec_ring_vm_funcs = { .emit_wreg = jpeg_v2_0_dec_ring_emit_wreg, .emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = jpeg_v2_5_ring_reset, }; static const struct amdgpu_ring_funcs jpeg_v2_6_dec_ring_vm_funcs = { @@ -698,6 +743,7 @@ static const struct amdgpu_ring_funcs jpeg_v2_6_dec_ring_vm_funcs = { .emit_wreg = jpeg_v2_0_dec_ring_emit_wreg, .emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = jpeg_v2_5_ring_reset, }; static void jpeg_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c index 4eca65ea9053..9faa9c6809df 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c @@ -34,6 +34,22 @@ #define mmUVD_JPEG_PITCH_INTERNAL_OFFSET 0x401f +static const struct amdgpu_hwip_reg_entry jpeg_reg_list_3_0[] = { + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_POWER_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_INT_STAT), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_CNTL), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_SIZE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_ADDR_MODE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_GFX10_ADDR_CONFIG), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_Y_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_UV_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_UV_PITCH), +}; + static void jpeg_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev); static void jpeg_v3_0_set_irq_funcs(struct amdgpu_device *adev); static int jpeg_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block, @@ -112,7 +128,14 @@ static int jpeg_v3_0_sw_init(struct amdgpu_ip_block *ip_block) adev->jpeg.internal.jpeg_pitch[0] = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH); - return 0; + r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_3_0, ARRAY_SIZE(jpeg_reg_list_3_0)); + if (r) + return r; + + adev->jpeg.supported_reset = AMDGPU_RESET_TYPE_PER_QUEUE; + r = amdgpu_jpeg_sysfs_reset_mask_init(adev); + + return r; } /** @@ -131,6 +154,8 @@ static int jpeg_v3_0_sw_fini(struct amdgpu_ip_block *ip_block) if (r) return r; + amdgpu_jpeg_sysfs_reset_mask_fini(adev); + r = amdgpu_jpeg_sw_fini(adev); return r; @@ -530,6 +555,13 @@ static int jpeg_v3_0_process_interrupt(struct amdgpu_device *adev, return 0; } +static int jpeg_v3_0_ring_reset(struct amdgpu_ring *ring, unsigned int vmid) +{ + jpeg_v3_0_stop(ring->adev); + jpeg_v3_0_start(ring->adev); + return amdgpu_ring_test_helper(ring); +} + static const struct amd_ip_funcs jpeg_v3_0_ip_funcs = { .name = "jpeg_v3_0", .early_init = jpeg_v3_0_early_init, @@ -543,6 +575,8 @@ static const struct amd_ip_funcs jpeg_v3_0_ip_funcs = { .wait_for_idle = jpeg_v3_0_wait_for_idle, .set_clockgating_state = jpeg_v3_0_set_clockgating_state, .set_powergating_state = jpeg_v3_0_set_powergating_state, + .dump_ip_state = amdgpu_jpeg_dump_ip_state, + .print_ip_state = amdgpu_jpeg_print_ip_state, }; static const struct amdgpu_ring_funcs jpeg_v3_0_dec_ring_vm_funcs = { @@ -573,6 +607,7 @@ static const struct amdgpu_ring_funcs jpeg_v3_0_dec_ring_vm_funcs = { .emit_wreg = jpeg_v2_0_dec_ring_emit_wreg, .emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = jpeg_v3_0_ring_reset, }; static void jpeg_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c index 0aef1f64afd0..292d4a234ea6 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c @@ -36,13 +36,28 @@ #define regUVD_JPEG_PITCH_INTERNAL_OFFSET 0x401f +static const struct amdgpu_hwip_reg_entry jpeg_reg_list_4_0[] = { + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_CNTL), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_SIZE), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH), +}; + static int jpeg_v4_0_start_sriov(struct amdgpu_device *adev); static void jpeg_v4_0_set_dec_ring_funcs(struct amdgpu_device *adev); static void jpeg_v4_0_set_irq_funcs(struct amdgpu_device *adev); static int jpeg_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state); static void jpeg_v4_0_set_ras_funcs(struct amdgpu_device *adev); - static void jpeg_v4_0_dec_ring_set_wptr(struct amdgpu_ring *ring); /** @@ -123,14 +138,15 @@ static int jpeg_v4_0_sw_init(struct amdgpu_ip_block *ip_block) r = amdgpu_jpeg_ras_sw_init(adev); if (r) return r; - /* TODO: Add queue reset mask when FW fully supports it */ - adev->jpeg.supported_reset = - amdgpu_get_soft_full_reset_mask(&adev->jpeg.inst[0].ring_dec[0]); - r = amdgpu_jpeg_sysfs_reset_mask_init(adev); + + r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_4_0, ARRAY_SIZE(jpeg_reg_list_4_0)); if (r) return r; - return 0; + adev->jpeg.supported_reset = AMDGPU_RESET_TYPE_PER_QUEUE; + r = amdgpu_jpeg_sysfs_reset_mask_init(adev); + + return r; } /** @@ -704,6 +720,16 @@ static int jpeg_v4_0_process_interrupt(struct amdgpu_device *adev, return 0; } +static int jpeg_v4_0_ring_reset(struct amdgpu_ring *ring, unsigned int vmid) +{ + if (amdgpu_sriov_vf(ring->adev)) + return -EINVAL; + + jpeg_v4_0_stop(ring->adev); + jpeg_v4_0_start(ring->adev); + return amdgpu_ring_test_helper(ring); +} + static const struct amd_ip_funcs jpeg_v4_0_ip_funcs = { .name = "jpeg_v4_0", .early_init = jpeg_v4_0_early_init, @@ -717,6 +743,8 @@ static const struct amd_ip_funcs jpeg_v4_0_ip_funcs = { .wait_for_idle = jpeg_v4_0_wait_for_idle, .set_clockgating_state = jpeg_v4_0_set_clockgating_state, .set_powergating_state = jpeg_v4_0_set_powergating_state, + .dump_ip_state = amdgpu_jpeg_dump_ip_state, + .print_ip_state = amdgpu_jpeg_print_ip_state, }; static const struct amdgpu_ring_funcs jpeg_v4_0_dec_ring_vm_funcs = { @@ -747,6 +775,7 @@ static const struct amdgpu_ring_funcs jpeg_v4_0_dec_ring_vm_funcs = { .emit_wreg = jpeg_v2_0_dec_ring_emit_wreg, .emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = jpeg_v4_0_ring_reset, }; static void jpeg_v4_0_set_dec_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c index 88f9771c1686..0588bb80f41e 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c @@ -59,10 +59,53 @@ static int amdgpu_ih_srcid_jpeg[] = { VCN_4_0__SRCID__JPEG7_DECODE }; +static const struct amdgpu_hwip_reg_entry jpeg_reg_list_4_0_3[] = { + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_SYS_INT_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_STATUS), +}; + static inline bool jpeg_v4_0_3_normalizn_reqd(struct amdgpu_device *adev) { - return amdgpu_sriov_vf(adev) || - (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)); + return (adev->jpeg.caps & AMDGPU_JPEG_CAPS(RRMT_ENABLED)) == 0; +} + +static inline int jpeg_v4_0_3_core_reg_offset(u32 pipe) +{ + if (pipe) + return ((0x40 * pipe) - 0xc80); + else + return 0; } /** @@ -144,10 +187,8 @@ static int jpeg_v4_0_3_sw_init(struct amdgpu_ip_block *ip_block) adev->jpeg.internal.jpeg_pitch[j] = regUVD_JRBC0_UVD_JRBC_SCRATCH0_INTERNAL_OFFSET; adev->jpeg.inst[i].external.jpeg_pitch[j] = - SOC15_REG_OFFSET1( - JPEG, jpeg_inst, - regUVD_JRBC0_UVD_JRBC_SCRATCH0, - (j ? (0x40 * j - 0xc80) : 0)); + SOC15_REG_OFFSET1(JPEG, jpeg_inst, regUVD_JRBC0_UVD_JRBC_SCRATCH0, + jpeg_v4_0_3_core_reg_offset(j)); } } @@ -159,13 +200,17 @@ static int jpeg_v4_0_3_sw_init(struct amdgpu_ip_block *ip_block) } } - /* TODO: Add queue reset mask when FW fully supports it */ - adev->jpeg.supported_reset = - amdgpu_get_soft_full_reset_mask(&adev->jpeg.inst[0].ring_dec[0]); - r = amdgpu_jpeg_sysfs_reset_mask_init(adev); + r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_4_0_3, ARRAY_SIZE(jpeg_reg_list_4_0_3)); if (r) return r; + if (!amdgpu_sriov_vf(adev)) { + adev->jpeg.supported_reset = AMDGPU_RESET_TYPE_PER_QUEUE; + r = amdgpu_jpeg_sysfs_reset_mask_init(adev); + if (r) + return r; + } + return 0; } @@ -185,7 +230,9 @@ static int jpeg_v4_0_3_sw_fini(struct amdgpu_ip_block *ip_block) if (r) return r; - amdgpu_jpeg_sysfs_reset_mask_fini(adev); + if (!amdgpu_sriov_vf(adev)) + amdgpu_jpeg_sysfs_reset_mask_fini(adev); + r = amdgpu_jpeg_sw_fini(adev); return r; @@ -331,6 +378,11 @@ static int jpeg_v4_0_3_hw_init(struct amdgpu_ip_block *ip_block) } } } else { + /* This flag is not set for VF, assumed to be disabled always */ + if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) & + 0x100) + adev->jpeg.caps |= AMDGPU_JPEG_CAPS(RRMT_ENABLED); + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { jpeg_inst = GET_INST(JPEG, i); @@ -475,6 +527,75 @@ static void jpeg_v4_0_3_enable_clock_gating(struct amdgpu_device *adev, int inst WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_GATE, data); } +static void jpeg_v4_0_3_start_inst(struct amdgpu_device *adev, int inst) +{ + int jpeg_inst = GET_INST(JPEG, inst); + + WREG32_SOC15(JPEG, jpeg_inst, regUVD_PGFSM_CONFIG, + 1 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT); + SOC15_WAIT_ON_RREG(JPEG, jpeg_inst, regUVD_PGFSM_STATUS, + UVD_PGFSM_STATUS__UVDJ_PWR_ON << + UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT, + UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK); + + /* disable anti hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS), + 0, ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); + + /* JPEG disable CGC */ + jpeg_v4_0_3_disable_clock_gating(adev, inst); + + /* MJPEG global tiling registers */ + WREG32_SOC15(JPEG, jpeg_inst, regJPEG_DEC_GFX8_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(JPEG, jpeg_inst, regJPEG_DEC_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* enable JMI channel */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL), 0, + ~UVD_JMI_CNTL__SOFT_RESET_MASK); +} + +static void jpeg_v4_0_3_start_jrbc(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + int jpeg_inst = GET_INST(JPEG, ring->me); + int reg_offset = jpeg_v4_0_3_core_reg_offset(ring->pipe); + + /* enable System Interrupt for JRBC */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regJPEG_SYS_INT_EN), + JPEG_SYS_INT_EN__DJRBC0_MASK << ring->pipe, + ~(JPEG_SYS_INT_EN__DJRBC0_MASK << ring->pipe)); + + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JMI0_UVD_LMI_JRBC_RB_VMID, + reg_offset, 0); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_CNTL, + reg_offset, + (0x00000001L | 0x00000002L)); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW, + reg_offset, lower_32_bits(ring->gpu_addr)); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH, + reg_offset, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_RPTR, + reg_offset, 0); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_WPTR, + reg_offset, 0); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_CNTL, + reg_offset, 0x00000002L); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_SIZE, + reg_offset, ring->ring_size / 4); + ring->wptr = RREG32_SOC15_OFFSET(JPEG, jpeg_inst, regUVD_JRBC0_UVD_JRBC_RB_WPTR, + reg_offset); +} + /** * jpeg_v4_0_3_start - start JPEG block * @@ -485,84 +606,42 @@ static void jpeg_v4_0_3_enable_clock_gating(struct amdgpu_device *adev, int inst static int jpeg_v4_0_3_start(struct amdgpu_device *adev) { struct amdgpu_ring *ring; - int i, j, jpeg_inst; + int i, j; for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { - jpeg_inst = GET_INST(JPEG, i); - - WREG32_SOC15(JPEG, jpeg_inst, regUVD_PGFSM_CONFIG, - 1 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT); - SOC15_WAIT_ON_RREG( - JPEG, jpeg_inst, regUVD_PGFSM_STATUS, - UVD_PGFSM_STATUS__UVDJ_PWR_ON - << UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT, - UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK); - - /* disable anti hang mechanism */ - WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, - regUVD_JPEG_POWER_STATUS), - 0, ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); - - /* JPEG disable CGC */ - jpeg_v4_0_3_disable_clock_gating(adev, i); - - /* MJPEG global tiling registers */ - WREG32_SOC15(JPEG, jpeg_inst, regJPEG_DEC_GFX8_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - WREG32_SOC15(JPEG, jpeg_inst, regJPEG_DEC_GFX10_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - - /* enable JMI channel */ - WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL), 0, - ~UVD_JMI_CNTL__SOFT_RESET_MASK); - + jpeg_v4_0_3_start_inst(adev, i); for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { - unsigned int reg_offset = (j?(0x40 * j - 0xc80):0); - ring = &adev->jpeg.inst[i].ring_dec[j]; - - /* enable System Interrupt for JRBC */ - WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, - regJPEG_SYS_INT_EN), - JPEG_SYS_INT_EN__DJRBC0_MASK << j, - ~(JPEG_SYS_INT_EN__DJRBC0_MASK << j)); - - WREG32_SOC15_OFFSET(JPEG, jpeg_inst, - regUVD_JMI0_UVD_LMI_JRBC_RB_VMID, - reg_offset, 0); - WREG32_SOC15_OFFSET(JPEG, jpeg_inst, - regUVD_JRBC0_UVD_JRBC_RB_CNTL, - reg_offset, - (0x00000001L | 0x00000002L)); - WREG32_SOC15_OFFSET( - JPEG, jpeg_inst, - regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW, - reg_offset, lower_32_bits(ring->gpu_addr)); - WREG32_SOC15_OFFSET( - JPEG, jpeg_inst, - regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH, - reg_offset, upper_32_bits(ring->gpu_addr)); - WREG32_SOC15_OFFSET(JPEG, jpeg_inst, - regUVD_JRBC0_UVD_JRBC_RB_RPTR, - reg_offset, 0); - WREG32_SOC15_OFFSET(JPEG, jpeg_inst, - regUVD_JRBC0_UVD_JRBC_RB_WPTR, - reg_offset, 0); - WREG32_SOC15_OFFSET(JPEG, jpeg_inst, - regUVD_JRBC0_UVD_JRBC_RB_CNTL, - reg_offset, 0x00000002L); - WREG32_SOC15_OFFSET(JPEG, jpeg_inst, - regUVD_JRBC0_UVD_JRBC_RB_SIZE, - reg_offset, ring->ring_size / 4); - ring->wptr = RREG32_SOC15_OFFSET( - JPEG, jpeg_inst, regUVD_JRBC0_UVD_JRBC_RB_WPTR, - reg_offset); + jpeg_v4_0_3_start_jrbc(ring); } } return 0; } +static void jpeg_v4_0_3_stop_inst(struct amdgpu_device *adev, int inst) +{ + int jpeg_inst = GET_INST(JPEG, inst); + /* reset JMI */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL), + UVD_JMI_CNTL__SOFT_RESET_MASK, + ~UVD_JMI_CNTL__SOFT_RESET_MASK); + + jpeg_v4_0_3_enable_clock_gating(adev, inst); + + /* enable anti hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS), + UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK, + ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); + + WREG32_SOC15(JPEG, jpeg_inst, regUVD_PGFSM_CONFIG, + 2 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT); + SOC15_WAIT_ON_RREG(JPEG, jpeg_inst, regUVD_PGFSM_STATUS, + UVD_PGFSM_STATUS__UVDJ_PWR_OFF << + UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT, + UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK); +} + /** * jpeg_v4_0_3_stop - stop JPEG block * @@ -572,31 +651,10 @@ static int jpeg_v4_0_3_start(struct amdgpu_device *adev) */ static int jpeg_v4_0_3_stop(struct amdgpu_device *adev) { - int i, jpeg_inst; + int i; - for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { - jpeg_inst = GET_INST(JPEG, i); - /* reset JMI */ - WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL), - UVD_JMI_CNTL__SOFT_RESET_MASK, - ~UVD_JMI_CNTL__SOFT_RESET_MASK); - - jpeg_v4_0_3_enable_clock_gating(adev, i); - - /* enable anti hang mechanism */ - WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, - regUVD_JPEG_POWER_STATUS), - UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK, - ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); - - WREG32_SOC15(JPEG, jpeg_inst, regUVD_PGFSM_CONFIG, - 2 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT); - SOC15_WAIT_ON_RREG( - JPEG, jpeg_inst, regUVD_PGFSM_STATUS, - UVD_PGFSM_STATUS__UVDJ_PWR_OFF - << UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT, - UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK); - } + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) + jpeg_v4_0_3_stop_inst(adev, i); return 0; } @@ -612,9 +670,8 @@ static uint64_t jpeg_v4_0_3_dec_ring_get_rptr(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - return RREG32_SOC15_OFFSET( - JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC0_UVD_JRBC_RB_RPTR, - ring->pipe ? (0x40 * ring->pipe - 0xc80) : 0); + return RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC0_UVD_JRBC_RB_RPTR, + jpeg_v4_0_3_core_reg_offset(ring->pipe)); } /** @@ -630,11 +687,9 @@ static uint64_t jpeg_v4_0_3_dec_ring_get_wptr(struct amdgpu_ring *ring) if (ring->use_doorbell) return adev->wb.wb[ring->wptr_offs]; - else - return RREG32_SOC15_OFFSET( - JPEG, GET_INST(JPEG, ring->me), - regUVD_JRBC0_UVD_JRBC_RB_WPTR, - ring->pipe ? (0x40 * ring->pipe - 0xc80) : 0); + + return RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC0_UVD_JRBC_RB_WPTR, + jpeg_v4_0_3_core_reg_offset(ring->pipe)); } static void jpeg_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring) @@ -659,10 +714,8 @@ static void jpeg_v4_0_3_dec_ring_set_wptr(struct amdgpu_ring *ring) adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr); WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr)); } else { - WREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), - regUVD_JRBC0_UVD_JRBC_RB_WPTR, - (ring->pipe ? (0x40 * ring->pipe - 0xc80) : - 0), + WREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC0_UVD_JRBC_RB_WPTR, + jpeg_v4_0_3_core_reg_offset(ring->pipe), lower_32_bits(ring->wptr)); } } @@ -915,13 +968,9 @@ static bool jpeg_v4_0_3_is_idle(void *handle) for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { - unsigned int reg_offset = (j?(0x40 * j - 0xc80):0); - - ret &= ((RREG32_SOC15_OFFSET( - JPEG, GET_INST(JPEG, i), - regUVD_JRBC0_UVD_JRBC_STATUS, - reg_offset) & - UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK) == + ret &= ((RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, i), + regUVD_JRBC0_UVD_JRBC_STATUS, jpeg_v4_0_3_core_reg_offset(j)) & + UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK) == UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK); } } @@ -937,13 +986,10 @@ static int jpeg_v4_0_3_wait_for_idle(struct amdgpu_ip_block *ip_block) for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { - unsigned int reg_offset = (j?(0x40 * j - 0xc80):0); - - ret &= SOC15_WAIT_ON_RREG_OFFSET( - JPEG, GET_INST(JPEG, i), - regUVD_JRBC0_UVD_JRBC_STATUS, reg_offset, + ret &= (SOC15_WAIT_ON_RREG_OFFSET(JPEG, GET_INST(JPEG, i), + regUVD_JRBC0_UVD_JRBC_STATUS, jpeg_v4_0_3_core_reg_offset(j), UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK, - UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK); + UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK)); } } return ret; @@ -1055,6 +1101,45 @@ static int jpeg_v4_0_3_process_interrupt(struct amdgpu_device *adev, return 0; } +static void jpeg_v4_0_3_core_stall_reset(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + int jpeg_inst = GET_INST(JPEG, ring->me); + int reg_offset = jpeg_v4_0_3_core_reg_offset(ring->pipe); + + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JMI0_UVD_JMI_CLIENT_STALL, + reg_offset, 0x1F); + SOC15_WAIT_ON_RREG(JPEG, jpeg_inst, + regUVD_JMI0_UVD_JMI_CLIENT_CLEAN_STATUS, + 0x1F, 0x1F); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JMI0_JPEG_LMI_DROP, + reg_offset, 0x1F); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regJPEG_CORE_RST_CTRL, + reg_offset, 1 << ring->pipe); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JMI0_UVD_JMI_CLIENT_STALL, + reg_offset, 0x00); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JMI0_JPEG_LMI_DROP, + reg_offset, 0x00); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regJPEG_CORE_RST_CTRL, + reg_offset, 0x00); +} + +static int jpeg_v4_0_3_ring_reset(struct amdgpu_ring *ring, unsigned int vmid) +{ + if (amdgpu_sriov_vf(ring->adev)) + return -EOPNOTSUPP; + + jpeg_v4_0_3_core_stall_reset(ring); + jpeg_v4_0_3_start_jrbc(ring); + return amdgpu_ring_test_helper(ring); +} + static const struct amd_ip_funcs jpeg_v4_0_3_ip_funcs = { .name = "jpeg_v4_0_3", .early_init = jpeg_v4_0_3_early_init, @@ -1068,6 +1153,8 @@ static const struct amd_ip_funcs jpeg_v4_0_3_ip_funcs = { .wait_for_idle = jpeg_v4_0_3_wait_for_idle, .set_clockgating_state = jpeg_v4_0_3_set_clockgating_state, .set_powergating_state = jpeg_v4_0_3_set_powergating_state, + .dump_ip_state = amdgpu_jpeg_dump_ip_state, + .print_ip_state = amdgpu_jpeg_print_ip_state, }; static const struct amdgpu_ring_funcs jpeg_v4_0_3_dec_ring_vm_funcs = { @@ -1099,6 +1186,7 @@ static const struct amdgpu_ring_funcs jpeg_v4_0_3_dec_ring_vm_funcs = { .emit_wreg = jpeg_v4_0_3_dec_ring_emit_wreg, .emit_reg_wait = jpeg_v4_0_3_dec_ring_emit_reg_wait, .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, + .reset = jpeg_v4_0_3_ring_reset, }; static void jpeg_v4_0_3_set_dec_ring_funcs(struct amdgpu_device *adev) @@ -1245,10 +1333,12 @@ static int jpeg_v4_0_3_aca_bank_parser(struct aca_handle *handle, struct aca_ban misc0 = bank->regs[ACA_REG_IDX_MISC0]; switch (type) { case ACA_SMU_TYPE_UE: + bank->aca_err_type = ACA_ERROR_TYPE_UE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE, 1ULL); break; case ACA_SMU_TYPE_CE: + bank->aca_err_type = ACA_ERROR_TYPE_CE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_CE, ACA_REG__MISC0__ERRCNT(misc0)); break; diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c index 6b3656984957..cbba1d9e8367 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c @@ -46,11 +46,26 @@ #define regJPEG_CGC_GATE_INTERNAL_OFFSET 0x4160 #define regUVD_NO_OP_INTERNAL_OFFSET 0x0029 +static const struct amdgpu_hwip_reg_entry jpeg_reg_list_4_0_5[] = { + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_CNTL), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_SIZE), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH), +}; + static void jpeg_v4_0_5_set_dec_ring_funcs(struct amdgpu_device *adev); static void jpeg_v4_0_5_set_irq_funcs(struct amdgpu_device *adev); static int jpeg_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state); - static void jpeg_v4_0_5_dec_ring_set_wptr(struct amdgpu_ring *ring); static int amdgpu_ih_clientid_jpeg[] = { @@ -58,6 +73,8 @@ static int amdgpu_ih_clientid_jpeg[] = { SOC15_IH_CLIENTID_VCN1 }; + + /** * jpeg_v4_0_5_early_init - set function pointers * @@ -153,6 +170,10 @@ static int jpeg_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block) adev->jpeg.inst[i].external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, i, regUVD_JPEG_PITCH); } + r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_4_0_5, ARRAY_SIZE(jpeg_reg_list_4_0_5)); + if (r) + return r; + /* TODO: Add queue reset mask when FW fully supports it */ adev->jpeg.supported_reset = amdgpu_get_soft_full_reset_mask(&adev->jpeg.inst[0].ring_dec[0]); @@ -759,6 +780,8 @@ static const struct amd_ip_funcs jpeg_v4_0_5_ip_funcs = { .wait_for_idle = jpeg_v4_0_5_wait_for_idle, .set_clockgating_state = jpeg_v4_0_5_set_clockgating_state, .set_powergating_state = jpeg_v4_0_5_set_powergating_state, + .dump_ip_state = amdgpu_jpeg_dump_ip_state, + .print_ip_state = amdgpu_jpeg_print_ip_state, }; static const struct amdgpu_ring_funcs jpeg_v4_0_5_dec_ring_vm_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c index d5cf0f2799d4..4a55e0cf39e4 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c @@ -34,6 +34,22 @@ #include "ivsrcid/vcn/irqsrcs_vcn_5_0.h" #include "jpeg_v5_0_0.h" +static const struct amdgpu_hwip_reg_entry jpeg_reg_list_5_0[] = { + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_CNTL), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_SIZE), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH), +}; + static void jpeg_v5_0_0_set_dec_ring_funcs(struct amdgpu_device *adev); static void jpeg_v5_0_0_set_irq_funcs(struct amdgpu_device *adev); static int jpeg_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block, @@ -100,6 +116,10 @@ static int jpeg_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block) adev->jpeg.internal.jpeg_pitch[0] = regUVD_JPEG_PITCH_INTERNAL_OFFSET; adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_PITCH); + r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_5_0, ARRAY_SIZE(jpeg_reg_list_5_0)); + if (r) + return r; + /* TODO: Add queue reset mask when FW fully supports it */ adev->jpeg.supported_reset = amdgpu_get_soft_full_reset_mask(&adev->jpeg.inst[0].ring_dec[0]); @@ -637,6 +657,8 @@ static const struct amd_ip_funcs jpeg_v5_0_0_ip_funcs = { .wait_for_idle = jpeg_v5_0_0_wait_for_idle, .set_clockgating_state = jpeg_v5_0_0_set_clockgating_state, .set_powergating_state = jpeg_v5_0_0_set_powergating_state, + .dump_ip_state = amdgpu_jpeg_dump_ip_state, + .print_ip_state = amdgpu_jpeg_print_ip_state, }; static const struct amdgpu_ring_funcs jpeg_v5_0_0_dec_ring_vm_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c index 40d4c32a8c2a..6e3f522e9133 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c @@ -52,6 +52,47 @@ static int amdgpu_ih_srcid_jpeg[] = { VCN_5_0__SRCID__JPEG9_DECODE, }; +static const struct amdgpu_hwip_reg_entry jpeg_reg_list_5_0_1[] = { + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC8_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC8_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC8_UVD_JRBC_STATUS), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC9_UVD_JRBC_RB_RPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC9_UVD_JRBC_RB_WPTR), + SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC9_UVD_JRBC_STATUS), +}; + static int jpeg_v5_0_1_core_reg_offset(u32 pipe) { if (pipe <= AMDGPU_MAX_JPEG_RINGS_4_0_3) @@ -145,6 +186,10 @@ static int jpeg_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block) } } + r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_5_0_1, ARRAY_SIZE(jpeg_reg_list_5_0_1)); + if (r) + return r; + return 0; } @@ -635,8 +680,8 @@ static const struct amd_ip_funcs jpeg_v5_0_1_ip_funcs = { .post_soft_reset = NULL, .set_clockgating_state = jpeg_v5_0_1_set_clockgating_state, .set_powergating_state = jpeg_v5_0_1_set_powergating_state, - .dump_ip_state = NULL, - .print_ip_state = NULL, + .dump_ip_state = amdgpu_jpeg_dump_ip_state, + .print_ip_state = amdgpu_jpeg_print_ip_state, }; static const struct amdgpu_ring_funcs jpeg_v5_0_1_dec_ring_vm_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h index 8ce146c00bb6..9de3272ef47f 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h @@ -26,4 +26,65 @@ extern const struct amdgpu_ip_block_version jpeg_v5_0_1_ip_block; +#define regUVD_JRBC0_UVD_JRBC_RB_WPTR 0x0640 +#define regUVD_JRBC0_UVD_JRBC_RB_WPTR_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_STATUS 0x0649 +#define regUVD_JRBC0_UVD_JRBC_STATUS_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_RB_RPTR 0x064a +#define regUVD_JRBC0_UVD_JRBC_RB_RPTR_BASE_IDX 1 +#define regUVD_JRBC1_UVD_JRBC_RB_WPTR 0x0000 +#define regUVD_JRBC1_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_STATUS 0x0009 +#define regUVD_JRBC1_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_RB_RPTR 0x000a +#define regUVD_JRBC1_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_RB_WPTR 0x0040 +#define regUVD_JRBC2_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_STATUS 0x0049 +#define regUVD_JRBC2_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_RB_RPTR 0x004a +#define regUVD_JRBC2_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_RB_WPTR 0x0080 +#define regUVD_JRBC3_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_STATUS 0x0089 +#define regUVD_JRBC3_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_RB_RPTR 0x008a +#define regUVD_JRBC3_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_RB_WPTR 0x00c0 +#define regUVD_JRBC4_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_STATUS 0x00c9 +#define regUVD_JRBC4_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_RB_RPTR 0x00ca +#define regUVD_JRBC4_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_RB_WPTR 0x0100 +#define regUVD_JRBC5_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_STATUS 0x0109 +#define regUVD_JRBC5_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_RB_RPTR 0x010a +#define regUVD_JRBC5_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_RB_WPTR 0x0140 +#define regUVD_JRBC6_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_STATUS 0x0149 +#define regUVD_JRBC6_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_RB_RPTR 0x014a +#define regUVD_JRBC6_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_RB_WPTR 0x0180 +#define regUVD_JRBC7_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_STATUS 0x0189 +#define regUVD_JRBC7_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_RB_RPTR 0x018a +#define regUVD_JRBC7_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC8_UVD_JRBC_RB_WPTR 0x01c0 +#define regUVD_JRBC8_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC8_UVD_JRBC_STATUS 0x01c9 +#define regUVD_JRBC8_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC8_UVD_JRBC_RB_RPTR 0x01ca +#define regUVD_JRBC8_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC9_UVD_JRBC_RB_WPTR 0x0440 +#define regUVD_JRBC9_UVD_JRBC_RB_WPTR_BASE_IDX 1 +#define regUVD_JRBC9_UVD_JRBC_STATUS 0x0449 +#define regUVD_JRBC9_UVD_JRBC_STATUS_BASE_IDX 1 +#define regUVD_JRBC9_UVD_JRBC_RB_RPTR 0x044a +#define regUVD_JRBC9_UVD_JRBC_RB_RPTR_BASE_IDX 1 + #endif /* __JPEG_V5_0_0_H__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c index 65f389eb65e5..747b05d9b3cf 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c @@ -54,6 +54,8 @@ MODULE_FIRMWARE("amdgpu/gc_11_5_1_mes_2.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_1_mes1.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_2_mes_2.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_2_mes1.bin"); +MODULE_FIRMWARE("amdgpu/gc_11_5_3_mes_2.bin"); +MODULE_FIRMWARE("amdgpu/gc_11_5_3_mes1.bin"); static int mes_v11_0_hw_init(struct amdgpu_ip_block *ip_block); static int mes_v11_0_hw_fini(struct amdgpu_ip_block *ip_block); @@ -62,6 +64,7 @@ static int mes_v11_0_kiq_hw_fini(struct amdgpu_device *adev); #define MES_EOP_SIZE 2048 #define GFX_MES_DRAM_SIZE 0x80000 +#define MES11_HW_RESOURCE_1_SIZE (128 * AMDGPU_GPU_PAGE_SIZE) static void mes_v11_0_ring_set_wptr(struct amdgpu_ring *ring) { @@ -730,9 +733,6 @@ static int mes_v11_0_set_hw_resources(struct amdgpu_mes *mes) static int mes_v11_0_set_hw_resources_1(struct amdgpu_mes *mes) { - int size = 128 * PAGE_SIZE; - int ret = 0; - struct amdgpu_device *adev = mes->adev; union MESAPI_SET_HW_RESOURCES_1 mes_set_hw_res_pkt; memset(&mes_set_hw_res_pkt, 0, sizeof(mes_set_hw_res_pkt)); @@ -740,19 +740,11 @@ static int mes_v11_0_set_hw_resources_1(struct amdgpu_mes *mes) mes_set_hw_res_pkt.header.opcode = MES_SCH_API_SET_HW_RSRC_1; mes_set_hw_res_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS; mes_set_hw_res_pkt.enable_mes_info_ctx = 1; - - ret = amdgpu_bo_create_kernel(adev, size, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, - &mes->resource_1, - &mes->resource_1_gpu_addr, - &mes->resource_1_addr); - if (ret) { - dev_err(adev->dev, "(%d) failed to create mes resource_1 bo\n", ret); - return ret; - } - mes_set_hw_res_pkt.mes_info_ctx_mc_addr = mes->resource_1_gpu_addr; - mes_set_hw_res_pkt.mes_info_ctx_size = mes->resource_1->tbo.base.size; + mes_set_hw_res_pkt.mes_info_ctx_size = MES11_HW_RESOURCE_1_SIZE; + mes_set_hw_res_pkt.cleaner_shader_fence_mc_addr = + mes->resource_1_gpu_addr + MES11_HW_RESOURCE_1_SIZE; + return mes_v11_0_submit_pkt_and_poll_completion(mes, &mes_set_hw_res_pkt, sizeof(mes_set_hw_res_pkt), offsetof(union MESAPI_SET_HW_RESOURCES_1, api_status)); @@ -1424,6 +1416,21 @@ static int mes_v11_0_sw_init(struct amdgpu_ip_block *ip_block) if (r) return r; + if (amdgpu_sriov_is_mes_info_enable(adev) || + adev->gfx.enable_cleaner_shader) { + r = amdgpu_bo_create_kernel(adev, + MES11_HW_RESOURCE_1_SIZE + AMDGPU_GPU_PAGE_SIZE, + PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM, + &adev->mes.resource_1, + &adev->mes.resource_1_gpu_addr, + &adev->mes.resource_1_addr); + if (r) { + dev_err(adev->dev, "(%d) failed to create mes resource_1 bo\n", r); + return r; + } + } + return 0; } @@ -1432,6 +1439,12 @@ static int mes_v11_0_sw_fini(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int pipe; + if (amdgpu_sriov_is_mes_info_enable(adev) || + adev->gfx.enable_cleaner_shader) { + amdgpu_bo_free_kernel(&adev->mes.resource_1, &adev->mes.resource_1_gpu_addr, + &adev->mes.resource_1_addr); + } + for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) { kfree(adev->mes.mqd_backup[pipe]); @@ -1619,7 +1632,8 @@ static int mes_v11_0_hw_init(struct amdgpu_ip_block *ip_block) if (r) goto failure; - if (amdgpu_sriov_is_mes_info_enable(adev)) { + if (amdgpu_sriov_is_mes_info_enable(adev) || + adev->gfx.enable_cleaner_shader) { r = mes_v11_0_set_hw_resources_1(&adev->mes); if (r) { DRM_ERROR("failed mes_v11_0_set_hw_resources_1, r=%d\n", r); @@ -1651,11 +1665,6 @@ failure: static int mes_v11_0_hw_fini(struct amdgpu_ip_block *ip_block) { - struct amdgpu_device *adev = ip_block->adev; - if (amdgpu_sriov_is_mes_info_enable(adev)) { - amdgpu_bo_free_kernel(&adev->mes.resource_1, &adev->mes.resource_1_gpu_addr, - &adev->mes.resource_1_addr); - } return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c index 901e924e69ad..8d7cf40030f2 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c @@ -686,6 +686,8 @@ static int mes_v12_0_set_hw_resources_1(struct amdgpu_mes *mes, int pipe) mes_set_hw_res_1_pkt.header.opcode = MES_SCH_API_SET_HW_RSRC_1; mes_set_hw_res_1_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS; mes_set_hw_res_1_pkt.mes_kiq_unmap_timeout = 0xa; + mes_set_hw_res_1_pkt.cleaner_shader_fence_mc_addr = + mes->resource_1_gpu_addr; return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe, &mes_set_hw_res_1_pkt, sizeof(mes_set_hw_res_1_pkt), @@ -1525,6 +1527,18 @@ static int mes_v12_0_sw_init(struct amdgpu_ip_block *ip_block) return r; } + if (adev->enable_uni_mes) { + r = amdgpu_bo_create_kernel(adev, AMDGPU_GPU_PAGE_SIZE, PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM, + &adev->mes.resource_1, + &adev->mes.resource_1_gpu_addr, + &adev->mes.resource_1_addr); + if (r) { + dev_err(adev->dev, "(%d) failed to create mes resource_1 bo\n", r); + return r; + } + } + return 0; } @@ -1533,6 +1547,11 @@ static int mes_v12_0_sw_fini(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int pipe; + if (adev->enable_uni_mes) + amdgpu_bo_free_kernel(&adev->mes.resource_1, + &adev->mes.resource_1_gpu_addr, + &adev->mes.resource_1_addr); + for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) { kfree(adev->mes.mqd_backup[pipe]); diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c index 9689e2b5d4e5..2adee2b94c37 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c @@ -172,6 +172,30 @@ static void mmhub_v1_7_init_tlb_regs(struct amdgpu_device *adev) WREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL, tmp); } +/* Set snoop bit for SDMA so that SDMA writes probe-invalidates RW lines */ +static void mmhub_v1_7_init_snoop_override_regs(struct amdgpu_device *adev) +{ + uint32_t tmp; + int i; + uint32_t distance = regDAGB1_WRCLI_GPU_SNOOP_OVERRIDE - + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE; + + for (i = 0; i < 5; i++) { /* DAGB instances */ + tmp = RREG32_SOC15_OFFSET(MMHUB, 0, + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, i * distance); + tmp |= (1 << 15); /* SDMA client is BIT15 */ + WREG32_SOC15_OFFSET(MMHUB, 0, + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, i * distance, tmp); + + tmp = RREG32_SOC15_OFFSET(MMHUB, 0, + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, i * distance); + tmp |= (1 << 15); + WREG32_SOC15_OFFSET(MMHUB, 0, + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, i * distance, tmp); + } + +} + static void mmhub_v1_7_init_cache_regs(struct amdgpu_device *adev) { uint32_t tmp; @@ -337,6 +361,7 @@ static int mmhub_v1_7_gart_enable(struct amdgpu_device *adev) mmhub_v1_7_init_system_aperture_regs(adev); mmhub_v1_7_init_tlb_regs(adev); mmhub_v1_7_init_cache_regs(adev); + mmhub_v1_7_init_snoop_override_regs(adev); mmhub_v1_7_enable_system_domain(adev); mmhub_v1_7_disable_identity_aperture(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c index e646e5cef0a2..58d22f0d5a68 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c @@ -213,6 +213,32 @@ static void mmhub_v1_8_init_tlb_regs(struct amdgpu_device *adev) } } +/* Set snoop bit for SDMA so that SDMA writes probe-invalidates RW lines */ +static void mmhub_v1_8_init_snoop_override_regs(struct amdgpu_device *adev) +{ + uint32_t tmp, inst_mask; + int i, j; + uint32_t distance = regDAGB1_WRCLI_GPU_SNOOP_OVERRIDE - + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE; + + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + for (j = 0; j < 5; j++) { /* DAGB instances */ + tmp = RREG32_SOC15_OFFSET(MMHUB, i, + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, j * distance); + tmp |= (1 << 15); /* SDMA client is BIT15 */ + WREG32_SOC15_OFFSET(MMHUB, i, + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, j * distance, tmp); + + tmp = RREG32_SOC15_OFFSET(MMHUB, i, + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, j * distance); + tmp |= (1 << 15); + WREG32_SOC15_OFFSET(MMHUB, i, + regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, j * distance, tmp); + } + } +} + static void mmhub_v1_8_init_cache_regs(struct amdgpu_device *adev) { uint32_t tmp, inst_mask; @@ -418,6 +444,7 @@ static int mmhub_v1_8_gart_enable(struct amdgpu_device *adev) mmhub_v1_8_init_system_aperture_regs(adev); mmhub_v1_8_init_tlb_regs(adev); mmhub_v1_8_init_cache_regs(adev); + mmhub_v1_8_init_snoop_override_regs(adev); mmhub_v1_8_enable_system_domain(adev); mmhub_v1_8_disable_identity_aperture(adev); @@ -719,10 +746,12 @@ static int mmhub_v1_8_aca_bank_parser(struct aca_handle *handle, struct aca_bank misc0 = bank->regs[ACA_REG_IDX_MISC0]; switch (type) { case ACA_SMU_TYPE_UE: + bank->aca_err_type = ACA_ERROR_TYPE_UE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE, 1ULL); break; case ACA_SMU_TYPE_CE: + bank->aca_err_type = ACA_ERROR_TYPE_CE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_CE, ACA_REG__MISC0__ERRCNT(misc0)); break; diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.c index b4ce3375d3fd..bc3d6c2fc87a 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.c @@ -103,6 +103,7 @@ mmhub_v3_3_print_l2_protection_fault_status(struct amdgpu_device *adev, switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) { case IP_VERSION(3, 3, 0): case IP_VERSION(3, 3, 1): + case IP_VERSION(3, 3, 2): mmhub_cid = cid < ARRAY_SIZE(mmhub_client_ids_v3_3) ? mmhub_client_ids_v3_3[cid][rw] : cid == 0x140 ? "UMSCH" : NULL; diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c index ff1b58e44689..fe0710b55c3a 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c @@ -198,6 +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; @@ -392,6 +422,7 @@ static 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); diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c index 62cdfe10e6f4..1c727ccd03b1 100644 --- a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c @@ -682,9 +682,9 @@ static int navi10_ih_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void navi10_ih_get_clockgating_state(void *handle, u64 *flags) +static void navi10_ih_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; if (!RREG32_SOC15(OSSSYS, 0, mmIH_CLK_CTRL)) *flags |= AMD_CG_SUPPORT_IH_CG; diff --git a/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c index c92875ceb31f..9900fe5c3bc3 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c +++ b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbif_v6_3_1.h" #include "nbif/nbif_6_3_1_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c index 739fce4fa8fd..04041b398781 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbio_v2_3.h" #include "nbio/nbio_2_3_default.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c index a54052dea8bf..f89e5f40e1a5 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbio_v4_3.h" #include "nbio/nbio_4_3_0_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c index 34180c6070dd..e911368c1aeb 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbio_v6_1.h" #include "nbio/nbio_6_1_default.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c index d1032e9992b4..1569a1e934ec 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbio_v7_0.h" #include "nbio/nbio_7_0_default.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c index 41421da63a08..2ece3ae75ec1 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbio_v7_11.h" #include "nbio/nbio_7_11_0_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c index a766e2d90cd0..acc5f363684a 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbio_v7_2.h" #include "nbio/nbio_7_2_0_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c index a26a9be58eac..d5002ff931d8 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbio_v7_4.h" #include "amdgpu_ras.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.c index 3fb6d2aa7e3b..2ee60b8746a6 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbio_v7_7.h" #include "nbio/nbio_7_7_0_offset.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c index 8a0a63ac88d2..f23cb79110d6 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c @@ -21,7 +21,6 @@ * */ #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "nbio_v7_9.h" #include "amdgpu_ras.h" diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c index 47db483c3516..5d4a4e7fd97f 100644 --- a/drivers/gpu/drm/amd/amdgpu/nv.c +++ b/drivers/gpu/drm/amd/amdgpu/nv.c @@ -454,6 +454,7 @@ nv_asic_reset_method(struct amdgpu_device *adev) switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) { case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): case IP_VERSION(13, 0, 1): case IP_VERSION(13, 0, 3): case IP_VERSION(13, 0, 5): @@ -1077,9 +1078,9 @@ static int nv_common_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void nv_common_get_clockgating_state(void *handle, u64 *flags) +static void nv_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; if (amdgpu_sriov_vf(adev)) *flags = 0; diff --git a/drivers/gpu/drm/amd/amdgpu/nvd.h b/drivers/gpu/drm/amd/amdgpu/nvd.h index 631dafb92299..56f1bfac0b20 100644 --- a/drivers/gpu/drm/amd/amdgpu/nvd.h +++ b/drivers/gpu/drm/amd/amdgpu/nvd.h @@ -64,6 +64,24 @@ #define PACKET3_INDIRECT_BUFFER_CNST_END 0x19 #define PACKET3_ATOMIC_GDS 0x1D #define PACKET3_ATOMIC_MEM 0x1E +#define PACKET3_ATOMIC_MEM__ATOMIC(x) ((((unsigned)(x)) & 0x7F) << 0) +#define PACKET3_ATOMIC_MEM__COMMAND(x) ((((unsigned)(x)) & 0xF) << 8) +#define PACKET3_ATOMIC_MEM__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_ATOMIC_MEM__ADDR_LO(x) (((unsigned)(x))) +#define PACKET3_ATOMIC_MEM__ADDR_HI(x) (((unsigned)(x))) +#define PACKET3_ATOMIC_MEM__SRC_DATA_LO(x) (((unsigned)(x))) +#define PACKET3_ATOMIC_MEM__SRC_DATA_HI(x) (((unsigned)(x))) +#define PACKET3_ATOMIC_MEM__CMP_DATA_LO(x) (((unsigned)(x))) +#define PACKET3_ATOMIC_MEM__CMP_DATA_HI(x) (((unsigned)(x))) +#define PACKET3_ATOMIC_MEM__LOOP_INTERVAL(x) ((((unsigned)(x)) & 0x1FFF) << 0) +#define PACKET3_ATOMIC_MEM__COMMAND__SINGLE_PASS_ATOMIC 0 +#define PACKET3_ATOMIC_MEM__COMMAND__LOOP_UNTIL_COMPARE_SATISFIED 1 +#define PACKET3_ATOMIC_MEM__COMMAND__WAIT_FOR_WRITE_CONFIRMATION 2 +#define PACKET3_ATOMIC_MEM__COMMAND__SEND_AND_CONTINUE 3 +#define PACKET3_ATOMIC_MEM__CACHE_POLICY__LRU 0 +#define PACKET3_ATOMIC_MEM__CACHE_POLICY__STREAM 1 +#define PACKET3_ATOMIC_MEM__CACHE_POLICY__NOA 2 +#define PACKET3_ATOMIC_MEM__CACHE_POLICY__BYPASS 3 #define PACKET3_OCCLUSION_QUERY 0x1F #define PACKET3_SET_PREDICATION 0x20 #define PACKET3_REG_RMW 0x21 @@ -105,6 +123,38 @@ * 1 - pfp * 2 - ce */ +#define PACKET3_WRITE_DATA__DST_SEL(x) ((((unsigned)(x)) & 0xF) << 8) +#define PACKET3_WRITE_DATA__ADDR_INCR(x) ((((unsigned)(x)) & 0x1) << 16) +#define PACKET3_WRITE_DATA__WR_CONFIRM(x) ((((unsigned)(x)) & 0x1) << 20) +#define PACKET3_WRITE_DATA__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_WRITE_DATA__DST_MMREG_ADDR(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_WRITE_DATA__DST_GDS_ADDR(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_WRITE_DATA__DST_MEM_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_WRITE_DATA__DST_MEM_ADDR_HI(x) ((unsigned)(x)) +#define PACKET3_WRITE_DATA__MODE(x) ((((unsigned)(x)) & 0x1) << 21) +#define PACKET3_WRITE_DATA__AID_ID(x) ((((unsigned)(x)) & 0x3) << 22) +#define PACKET3_WRITE_DATA__TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 24) +#define PACKET3_WRITE_DATA__DST_MMREG_ADDR_LO(x) ((unsigned)(x)) +#define PACKET3_WRITE_DATA__DST_MMREG_ADDR_HI(x) ((((unsigned)(x)) & 0xFF) << 0) +#define PACKET3_WRITE_DATA__DST_SEL__MEM_MAPPED_REGISTER 0 +#define PACKET3_WRITE_DATA__DST_SEL__TC_L2 2 +#define PACKET3_WRITE_DATA__DST_SEL__GDS 3 +#define PACKET3_WRITE_DATA__DST_SEL__MEMORY 5 +#define PACKET3_WRITE_DATA__DST_SEL__MEMORY_MAPPED_ADC_PERSISTENT_STATE 6 +#define PACKET3_WRITE_DATA__ADDR_INCR__INCREMENT_ADDRESS 0 +#define PACKET3_WRITE_DATA__ADDR_INCR__DO_NOT_INCREMENT_ADDRESS 1 +#define PACKET3_WRITE_DATA__WR_CONFIRM__DO_NOT_WAIT_FOR_WRITE_CONFIRMATION 0 +#define PACKET3_WRITE_DATA__WR_CONFIRM__WAIT_FOR_WRITE_CONFIRMATION 1 +#define PACKET3_WRITE_DATA__MODE__PF_VF_DISABLED 0 +#define PACKET3_WRITE_DATA__MODE__PF_VF_ENABLED 1 +#define PACKET3_WRITE_DATA__TEMPORAL__RT 0 +#define PACKET3_WRITE_DATA__TEMPORAL__NT 1 +#define PACKET3_WRITE_DATA__TEMPORAL__HT 2 +#define PACKET3_WRITE_DATA__TEMPORAL__LU 3 +#define PACKET3_WRITE_DATA__CACHE_POLICY__LRU 0 +#define PACKET3_WRITE_DATA__CACHE_POLICY__STREAM 1 +#define PACKET3_WRITE_DATA__CACHE_POLICY__NOA 2 +#define PACKET3_WRITE_DATA__CACHE_POLICY__BYPASS 3 #define PACKET3_DRAW_INDEX_INDIRECT_MULTI 0x38 #define PACKET3_MEM_SEMAPHORE 0x39 # define PACKET3_SEM_USE_MAILBOX (0x1 << 16) @@ -135,6 +185,42 @@ /* 0 - me * 1 - pfp */ +#define PACKET3_WAIT_REG_MEM__FUNCTION(x) ((((unsigned)(x)) & 0x7) << 0) +#define PACKET3_WAIT_REG_MEM__MEM_SPACE(x) ((((unsigned)(x)) & 0x3) << 4) +#define PACKET3_WAIT_REG_MEM__OPERATION(x) ((((unsigned)(x)) & 0x3) << 6) +#define PACKET3_WAIT_REG_MEM__MES_INTR_PIPE(x) ((((unsigned)(x)) & 0x3) << 22) +#define PACKET3_WAIT_REG_MEM__MES_ACTION(x) ((((unsigned)(x)) & 0x1) << 24) +#define PACKET3_WAIT_REG_MEM__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_WAIT_REG_MEM__TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_WAIT_REG_MEM__MEM_POLL_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_WAIT_REG_MEM__REG_POLL_ADDR(x) ((((unsigned)(x)) & 0X3FFFF) << 0) +#define PACKET3_WAIT_REG_MEM__REG_WRITE_ADDR1(x) ((((unsigned)(x)) & 0X3FFFF) << 0) +#define PACKET3_WAIT_REG_MEM__MEM_POLL_ADDR_HI(x) ((unsigned)(x)) +#define PACKET3_WAIT_REG_MEM__REG_WRITE_ADDR2(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_WAIT_REG_MEM__REFERENCE(x) ((unsigned)(x)) +#define PACKET3_WAIT_REG_MEM__MASK(x) ((unsigned)(x)) +#define PACKET3_WAIT_REG_MEM__POLL_INTERVAL(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_WAIT_REG_MEM__OPTIMIZE_ACE_OFFLOAD_MODE(x) ((((unsigned)(x)) & 0x1) << 31) +#define PACKET3_WAIT_REG_MEM__FUNCTION__ALWAYS_PASS 0 +#define PACKET3_WAIT_REG_MEM__FUNCTION__LESS_THAN_REF_VALUE 1 +#define PACKET3_WAIT_REG_MEM__FUNCTION__LESS_THAN_EQUAL_TO_THE_REF_VALUE 2 +#define PACKET3_WAIT_REG_MEM__FUNCTION__EQUAL_TO_THE_REFERENCE_VALUE 3 +#define PACKET3_WAIT_REG_MEM__FUNCTION__NOT_EQUAL_REFERENCE_VALUE 4 +#define PACKET3_WAIT_REG_MEM__FUNCTION__GREATER_THAN_OR_EQUAL_REFERENCE_VALUE 5 +#define PACKET3_WAIT_REG_MEM__FUNCTION__GREATER_THAN_REFERENCE_VALUE 6 +#define PACKET3_WAIT_REG_MEM__MEM_SPACE__REGISTER_SPACE 0 +#define PACKET3_WAIT_REG_MEM__MEM_SPACE__MEMORY_SPACE 1 +#define PACKET3_WAIT_REG_MEM__OPERATION__WAIT_REG_MEM 0 +#define PACKET3_WAIT_REG_MEM__OPERATION__WR_WAIT_WR_REG 1 +#define PACKET3_WAIT_REG_MEM__OPERATION__WAIT_MEM_PREEMPTABLE 3 +#define PACKET3_WAIT_REG_MEM__CACHE_POLICY__LRU 0 +#define PACKET3_WAIT_REG_MEM__CACHE_POLICY__STREAM 1 +#define PACKET3_WAIT_REG_MEM__CACHE_POLICY__NOA 2 +#define PACKET3_WAIT_REG_MEM__CACHE_POLICY__BYPASS 3 +#define PACKET3_WAIT_REG_MEM__TEMPORAL__RT 0 +#define PACKET3_WAIT_REG_MEM__TEMPORAL__NT 1 +#define PACKET3_WAIT_REG_MEM__TEMPORAL__HT 2 +#define PACKET3_WAIT_REG_MEM__TEMPORAL__LU 3 #define PACKET3_INDIRECT_BUFFER 0x3F #define INDIRECT_BUFFER_VALID (1 << 23) #define INDIRECT_BUFFER_CACHE_POLICY(x) ((x) << 28) @@ -144,8 +230,94 @@ */ #define INDIRECT_BUFFER_PRE_ENB(x) ((x) << 21) #define INDIRECT_BUFFER_PRE_RESUME(x) ((x) << 30) +#define PACKET3_INDIRECT_BUFFER__IB_BASE_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_INDIRECT_BUFFER__IB_BASE_HI(x) ((unsigned)(x)) +#define PACKET3_INDIRECT_BUFFER__IB_SIZE(x) ((((unsigned)(x)) & 0xFFFFF) << 0) +#define PACKET3_INDIRECT_BUFFER__CHAIN(x) ((((unsigned)(x)) & 0x1) << 20) +#define PACKET3_INDIRECT_BUFFER__OFFLOAD_POLLING(x) ((((unsigned)(x)) & 0x1) << 21) +#define PACKET3_INDIRECT_BUFFER__VALID(x) ((((unsigned)(x)) & 0x1) << 23) +#define PACKET3_INDIRECT_BUFFER__VMID(x) ((((unsigned)(x)) & 0xF) << 24) +#define PACKET3_INDIRECT_BUFFER__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 28) +#define PACKET3_INDIRECT_BUFFER__TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 28) +#define PACKET3_INDIRECT_BUFFER__PRIV(x) ((((unsigned)(x)) & 0x1) << 31) +#define PACKET3_INDIRECT_BUFFER__TEMPORAL__RT 0 +#define PACKET3_INDIRECT_BUFFER__TEMPORAL__NT 1 +#define PACKET3_INDIRECT_BUFFER__TEMPORAL__HT 2 +#define PACKET3_INDIRECT_BUFFER__TEMPORAL__LU 3 +#define PACKET3_INDIRECT_BUFFER__CACHE_POLICY__LRU 0 +#define PACKET3_INDIRECT_BUFFER__CACHE_POLICY__STREAM 1 +#define PACKET3_INDIRECT_BUFFER__CACHE_POLICY__NOA 2 +#define PACKET3_INDIRECT_BUFFER__CACHE_POLICY__BYPASS 3 #define PACKET3_COND_INDIRECT_BUFFER 0x3F #define PACKET3_COPY_DATA 0x40 +#define PACKET3_COPY_DATA__SRC_SEL(x) ((((unsigned)(x)) & 0xF) << 0) +#define PACKET3_COPY_DATA__DST_SEL(x) ((((unsigned)(x)) & 0xF) << 8) +#define PACKET3_COPY_DATA__SRC_CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 13) +#define PACKET3_COPY_DATA__SRC_TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 13) +#define PACKET3_COPY_DATA__COUNT_SEL(x) ((((unsigned)(x)) & 0x1) << 16) +#define PACKET3_COPY_DATA__WR_CONFIRM(x) ((((unsigned)(x)) & 0x1) << 20) +#define PACKET3_COPY_DATA__DST_CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_COPY_DATA__PQ_EXE_STATUS(x) ((((unsigned)(x)) & 0x1) << 29) +#define PACKET3_COPY_DATA__SRC_REG_OFFSET(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_COPY_DATA__SRC_32B_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_COPY_DATA__SRC_64B_ADDR_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3) +#define PACKET3_COPY_DATA__SRC_GDS_ADDR_LO(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_COPY_DATA__IMM_DATA(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__SRC_MEMTC_ADDR_HI(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__SRC_IMM_DATA(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__DST_REG_OFFSET(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_COPY_DATA__DST_32B_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_COPY_DATA__DST_64B_ADDR_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3) +#define PACKET3_COPY_DATA__DST_GDS_ADDR_LO(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_COPY_DATA__DST_ADDR_HI(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__MODE(x) ((((unsigned)(x)) & 0x1) << 21) +#define PACKET3_COPY_DATA__AID_ID(x) ((((unsigned)(x)) & 0x3) << 23) +#define PACKET3_COPY_DATA__DST_TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_COPY_DATA__SRC_REG_OFFSET_LO(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__SRC_REG_OFFSET_HI(x) ((((unsigned)(x)) & 0xFF) << 0) +#define PACKET3_COPY_DATA__DST_REG_OFFSET_LO(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__DST_REG_OFFSET_HI(x) ((((unsigned)(x)) & 0xFF) << 0) +#define PACKET3_COPY_DATA__SRC_SEL__MEM_MAPPED_REGISTER 0 +#define PACKET3_COPY_DATA__SRC_SEL__TC_L2_OBSOLETE 1 +#define PACKET3_COPY_DATA__SRC_SEL__TC_L2 2 +#define PACKET3_COPY_DATA__SRC_SEL__GDS 3 +#define PACKET3_COPY_DATA__SRC_SEL__PERFCOUNTERS 4 +#define PACKET3_COPY_DATA__SRC_SEL__IMMEDIATE_DATA 5 +#define PACKET3_COPY_DATA__SRC_SEL__ATOMIC_RETURN_DATA 6 +#define PACKET3_COPY_DATA__SRC_SEL__GDS_ATOMIC_RETURN_DATA0 7 +#define PACKET3_COPY_DATA__SRC_SEL__GDS_ATOMIC_RETURN_DATA1 8 +#define PACKET3_COPY_DATA__SRC_SEL__GPU_CLOCK_COUNT 9 +#define PACKET3_COPY_DATA__SRC_SEL__SYSTEM_CLOCK_COUNT 10 +#define PACKET3_COPY_DATA__DST_SEL__MEM_MAPPED_REGISTER 0 +#define PACKET3_COPY_DATA__DST_SEL__TC_L2 2 +#define PACKET3_COPY_DATA__DST_SEL__GDS 3 +#define PACKET3_COPY_DATA__DST_SEL__PERFCOUNTERS 4 +#define PACKET3_COPY_DATA__DST_SEL__TC_L2_OBSOLETE 5 +#define PACKET3_COPY_DATA__DST_SEL__MEM_MAPPED_REG_DC 6 +#define PACKET3_COPY_DATA__SRC_TEMPORAL__RT 0 +#define PACKET3_COPY_DATA__SRC_TEMPORAL__NT 1 +#define PACKET3_COPY_DATA__SRC_TEMPORAL__HT 2 +#define PACKET3_COPY_DATA__SRC_TEMPORAL__LU 3 +#define PACKET3_COPY_DATA__SRC_CACHE_POLICY__LRU 0 +#define PACKET3_COPY_DATA__SRC_CACHE_POLICY__STREAM 1 +#define PACKET3_COPY_DATA__SRC_CACHE_POLICY__NOA 2 +#define PACKET3_COPY_DATA__SRC_CACHE_POLICY__BYPASS 3 +#define PACKET3_COPY_DATA__COUNT_SEL__32_BITS_OF_DATA 0 +#define PACKET3_COPY_DATA__COUNT_SEL__64_BITS_OF_DATA 1 +#define PACKET3_COPY_DATA__WR_CONFIRM__DO_NOT_WAIT_FOR_CONFIRMATION 0 +#define PACKET3_COPY_DATA__WR_CONFIRM__WAIT_FOR_CONFIRMATION 1 +#define PACKET3_COPY_DATA__MODE__PF_VF_DISABLED 0 +#define PACKET3_COPY_DATA__MODE__PF_VF_ENABLED 1 +#define PACKET3_COPY_DATA__DST_TEMPORAL__RT 0 +#define PACKET3_COPY_DATA__DST_TEMPORAL__NT 1 +#define PACKET3_COPY_DATA__DST_TEMPORAL__HT 2 +#define PACKET3_COPY_DATA__DST_TEMPORAL__LU 3 +#define PACKET3_COPY_DATA__DST_CACHE_POLICY__LRU 0 +#define PACKET3_COPY_DATA__DST_CACHE_POLICY__STREAM 1 +#define PACKET3_COPY_DATA__DST_CACHE_POLICY__NOA 2 +#define PACKET3_COPY_DATA__DST_CACHE_POLICY__BYPASS 3 +#define PACKET3_COPY_DATA__PQ_EXE_STATUS__DEFAULT 0 +#define PACKET3_COPY_DATA__PQ_EXE_STATUS__PHASE_UPDATE 1 #define PACKET3_CP_DMA 0x41 #define PACKET3_PFP_SYNC_ME 0x42 #define PACKET3_SURFACE_SYNC 0x43 @@ -160,6 +332,23 @@ * 3 - SAMPLE_STREAMOUTSTAT* * 4 - *S_PARTIAL_FLUSH */ +#define PACKET3_EVENT_WRITE__EVENT_TYPE(x) ((((unsigned)(x)) & 0x3F) << 0) +#define PACKET3_EVENT_WRITE__EVENT_INDEX(x) ((((unsigned)(x)) & 0xF) << 8) +#define PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE(x) ((((unsigned)(x)) & 0x3) << 29) +#define PACKET3_EVENT_WRITE__OFFLOAD_ENABLE(x) ((((unsigned)(x)) & 0x1) << 0) +#define PACKET3_EVENT_WRITE__ADDRESS_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3) +#define PACKET3_EVENT_WRITE__ADDRESS_HI(x) ((unsigned)(x)) +#define PACKET3_EVENT_WRITE__EVENT_INDEX__OTHER 0 +#define PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_PIPELINESTAT 2 +#define PACKET3_EVENT_WRITE__EVENT_INDEX__CS_PARTIAL_FLUSH 4 +#define PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_STREAMOUTSTATS 8 +#define PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_STREAMOUTSTATS1 9 +#define PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_STREAMOUTSTATS2 10 +#define PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_STREAMOUTSTATS3 11 +#define PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE__LEGACY_MODE 0 +#define PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE__MIXED_MODE1 1 +#define PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE__NEW_MODE 2 +#define PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE__MIXED_MODE3 3 #define PACKET3_EVENT_WRITE_EOP 0x47 #define PACKET3_EVENT_WRITE_EOS 0x48 #define PACKET3_RELEASE_MEM 0x49 @@ -304,6 +493,12 @@ * 2: REVERSE */ #define PACKET3_ACQUIRE_MEM_GCR_RANGE_IS_PA (1 << 18) +#define PACKET3_ACQUIRE_MEM__COHER_SIZE(x) ((unsigned)(x)) +#define PACKET3_ACQUIRE_MEM__COHER_SIZE_HI(x) ((((unsigned)(x)) & 0xFF) << 0) +#define PACKET3_ACQUIRE_MEM__COHER_BASE_LO(x) ((unsigned)(x)) +#define PACKET3_ACQUIRE_MEM__COHER_BASE_HI(x) ((((unsigned)(x)) & 0xFFFFFF) << 0) +#define PACKET3_ACQUIRE_MEM__POLL_INTERVAL(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_ACQUIRE_MEM__GCR_CNTL(x) ((((unsigned)(x)) & 0x7FFFF) << 0) #define PACKET3_REWIND 0x59 #define PACKET3_INTERRUPT 0x5A #define PACKET3_GEN_PDEPTE 0x5B @@ -330,11 +525,17 @@ #define PACKET3_SET_SH_REG 0x76 #define PACKET3_SET_SH_REG_START 0x00002c00 #define PACKET3_SET_SH_REG_END 0x00003000 +#define PACKET3_SET_SH_REG__REG_OFFSET(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_SET_SH_REG__VMID_SHIFT(x) ((((unsigned)(x)) & 0x1F) << 23) +#define PACKET3_SET_SH_REG__INDEX(x) ((((unsigned)(x)) & 0xF) << 28) +#define PACKET3_SET_SH_REG__INDEX__DEFAULT 0 +#define PACKET3_SET_SH_REG__INDEX__INSERT_VMID 1 #define PACKET3_SET_SH_REG_OFFSET 0x77 #define PACKET3_SET_QUEUE_REG 0x78 #define PACKET3_SET_UCONFIG_REG 0x79 #define PACKET3_SET_UCONFIG_REG_START 0x0000c000 #define PACKET3_SET_UCONFIG_REG_END 0x0000c400 +#define PACKET3_SET_UCONFIG_REG__REG_OFFSET(x) ((((unsigned)(x)) & 0xFFFF) << 0) #define PACKET3_SET_UCONFIG_REG_INDEX 0x7A #define PACKET3_FORWARD_HEADER 0x7C #define PACKET3_SCRATCH_RAM_WRITE 0x7D @@ -369,6 +570,7 @@ # define PACKET3_INVALIDATE_TLBS_DST_SEL(x) ((x) << 0) # define PACKET3_INVALIDATE_TLBS_ALL_HUB(x) ((x) << 4) # define PACKET3_INVALIDATE_TLBS_PASID(x) ((x) << 5) +# define PACKET3_INVALIDATE_TLBS_FLUSH_TYPE(x) ((x) << 29) #define PACKET3_AQL_PACKET 0x99 #define PACKET3_DMA_DATA_FILL_MULTI 0x9A #define PACKET3_SET_SH_REG_INDEX 0x9B @@ -462,6 +664,12 @@ # define PACKET3_QUERY_STATUS_ENG_SEL(x) ((x) << 25) #define PACKET3_RUN_LIST 0xA5 #define PACKET3_MAP_PROCESS_VM 0xA6 + +#define PACKET3_RUN_CLEANER_SHADER 0xD2 +/* 1. header + * 2. RESERVED [31:0] + */ + /* GFX11 */ #define PACKET3_SET_Q_PREEMPTION_MODE 0xF0 # define PACKET3_SET_Q_PREEMPTION_MODE_IB_VMID(x) ((x) << 0) diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c index 2395f1856962..bb5dfc410a66 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c @@ -129,6 +129,7 @@ static int psp_v11_0_init_microcode(struct psp_context *psp) err = psp_init_ta_microcode(psp, ucode_prefix); break; case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): err = psp_init_asd_microcode(psp, ucode_prefix); if (err) return err; diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c index 4d33c95a5116..7c49c3f3c388 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c @@ -35,6 +35,8 @@ MODULE_FIRMWARE("amdgpu/psp_14_0_2_sos.bin"); MODULE_FIRMWARE("amdgpu/psp_14_0_2_ta.bin"); MODULE_FIRMWARE("amdgpu/psp_14_0_3_sos.bin"); MODULE_FIRMWARE("amdgpu/psp_14_0_3_ta.bin"); +MODULE_FIRMWARE("amdgpu/psp_14_0_5_toc.bin"); +MODULE_FIRMWARE("amdgpu/psp_14_0_5_ta.bin"); /* For large FW files the time to complete can be very long */ #define USBC_PD_POLLING_LIMIT_S 240 @@ -72,6 +74,14 @@ static int psp_v14_0_init_microcode(struct psp_context *psp) if (err) return err; break; + case IP_VERSION(14, 0, 5): + err = psp_init_toc_microcode(psp, ucode_prefix); + if (err) + return err; + err = psp_init_ta_microcode(psp, ucode_prefix); + if (err) + return err; + break; default: BUG(); } diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c index c611328671ed..6e75a4a85f74 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c @@ -1514,9 +1514,9 @@ static int sdma_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void sdma_v3_0_get_clockgating_state(void *handle, u64 *flags) +static void sdma_v3_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c index b48d9c0b2e1c..d31ee01383df 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c @@ -2331,9 +2331,9 @@ static int sdma_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void sdma_v4_0_get_clockgating_state(void *handle, u64 *flags) +static void sdma_v4_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) @@ -2381,7 +2381,6 @@ static void sdma_v4_0_dump_ip_state(struct amdgpu_ip_block *ip_block) if (!adev->sdma.ip_dump) return; - amdgpu_gfx_off_ctrl(adev, false); for (i = 0; i < adev->sdma.num_instances; i++) { instance_offset = i * reg_count; for (j = 0; j < reg_count; j++) @@ -2389,7 +2388,6 @@ static void sdma_v4_0_dump_ip_state(struct amdgpu_ip_block *ip_block) RREG32(sdma_v4_0_get_reg_offset(adev, i, sdma_reg_list_4_0[j].reg_offset)); } - amdgpu_gfx_off_ctrl(adev, true); } const struct amd_ip_funcs sdma_v4_0_ip_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c index 5e0066cd6c51..23a6bb16a0b1 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c @@ -1920,9 +1920,9 @@ static int sdma_v4_4_2_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void sdma_v4_4_2_get_clockgating_state(void *handle, u64 *flags) +static void sdma_v4_4_2_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) @@ -2392,10 +2392,12 @@ static int sdma_v4_4_2_aca_bank_parser(struct aca_handle *handle, struct aca_ban misc0 = bank->regs[ACA_REG_IDX_MISC0]; switch (type) { case ACA_SMU_TYPE_UE: + bank->aca_err_type = ACA_ERROR_TYPE_UE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE, 1ULL); break; case ACA_SMU_TYPE_CE: + bank->aca_err_type = ACA_ERROR_TYPE_CE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_CE, ACA_REG__MISC0__ERRCNT(misc0)); break; diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c index b764550834a0..377efb2b8d0e 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c @@ -1883,9 +1883,9 @@ static int sdma_v5_0_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void sdma_v5_0_get_clockgating_state(void *handle, u64 *flags) +static void sdma_v5_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c index b1818e87889a..ce05d895f977 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c @@ -1847,9 +1847,9 @@ static int sdma_v5_2_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void sdma_v5_2_get_clockgating_state(void *handle, u64 *flags) +static void sdma_v5_2_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c index 1a023b45f0be..373703d1596d 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c @@ -51,6 +51,7 @@ MODULE_FIRMWARE("amdgpu/sdma_6_0_3.bin"); MODULE_FIRMWARE("amdgpu/sdma_6_1_0.bin"); MODULE_FIRMWARE("amdgpu/sdma_6_1_1.bin"); MODULE_FIRMWARE("amdgpu/sdma_6_1_2.bin"); +MODULE_FIRMWARE("amdgpu/sdma_6_1_3.bin"); #define SDMA1_REG_OFFSET 0x600 #define SDMA0_HYP_DEC_REG_START 0x5880 @@ -1613,7 +1614,7 @@ static int sdma_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void sdma_v6_0_get_clockgating_state(void *handle, u64 *flags) +static void sdma_v6_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { } diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c index 9c17df2cf37b..9eb8f4f9f302 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c @@ -1592,7 +1592,7 @@ static int sdma_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void sdma_v7_0_get_clockgating_state(void *handle, u64 *flags) +static void sdma_v7_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { } @@ -1741,11 +1741,12 @@ static void sdma_v7_0_emit_copy_buffer(struct amdgpu_ib *ib, uint32_t byte_count, uint32_t copy_flags) { - uint32_t num_type, data_format, max_com; + uint32_t num_type, data_format, max_com, write_cm; max_com = AMDGPU_COPY_FLAGS_GET(copy_flags, MAX_COMPRESSED); data_format = AMDGPU_COPY_FLAGS_GET(copy_flags, DATA_FORMAT); num_type = AMDGPU_COPY_FLAGS_GET(copy_flags, NUMBER_TYPE); + write_cm = AMDGPU_COPY_FLAGS_GET(copy_flags, WRITE_COMPRESS_DISABLE) ? 2 : 1; ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_COPY) | SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) | @@ -1762,7 +1763,7 @@ static void sdma_v7_0_emit_copy_buffer(struct amdgpu_ib *ib, if ((copy_flags & (AMDGPU_COPY_FLAGS_READ_DECOMPRESSED | AMDGPU_COPY_FLAGS_WRITE_COMPRESSED))) ib->ptr[ib->length_dw++] = SDMA_DCC_DATA_FORMAT(data_format) | SDMA_DCC_NUM_TYPE(num_type) | ((copy_flags & AMDGPU_COPY_FLAGS_READ_DECOMPRESSED) ? SDMA_DCC_READ_CM(2) : 0) | - ((copy_flags & AMDGPU_COPY_FLAGS_WRITE_COMPRESSED) ? SDMA_DCC_WRITE_CM(1) : 0) | + ((copy_flags & AMDGPU_COPY_FLAGS_WRITE_COMPRESSED) ? SDMA_DCC_WRITE_CM(write_cm) : 0) | SDMA_DCC_MAX_COM(max_com) | SDMA_DCC_MAX_UCOM(1); else ib->ptr[ib->length_dw++] = 0; diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index a59b4c36cad7..2c6d2099e215 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -28,7 +28,6 @@ #include <drm/amdgpu_drm.h> #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "amdgpu_ih.h" #include "amdgpu_uvd.h" #include "amdgpu_vce.h" @@ -605,12 +604,10 @@ soc15_asic_reset_method(struct amdgpu_device *adev) static bool soc15_need_reset_on_resume(struct amdgpu_device *adev) { /* Will reset for the following suspend abort cases. - * 1) Only reset on APU side, dGPU hasn't checked yet. - * 2) S3 suspend aborted in the normal S3 suspend or - * performing pm core test. + * 1) S3 suspend aborted in the normal S3 suspend + * 2) S3 suspend aborted in performing pm core test. */ - if (adev->flags & AMD_IS_APU && adev->in_s3 && - !pm_resume_via_firmware()) + if (adev->in_s3 && !pm_resume_via_firmware()) return true; else return false; @@ -1464,9 +1461,9 @@ static int soc15_common_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void soc15_common_get_clockgating_state(void *handle, u64 *flags) +static void soc15_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/soc15d.h b/drivers/gpu/drm/amd/amdgpu/soc15d.h index b9cbeb389edc..a5000c171c02 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15d.h +++ b/drivers/gpu/drm/amd/amdgpu/soc15d.h @@ -93,11 +93,25 @@ #define PACKET3_DISPATCH_INDIRECT 0x16 #define PACKET3_ATOMIC_GDS 0x1D #define PACKET3_ATOMIC_MEM 0x1E +#define PACKET3_ATOMIC_MEM__ATOMIC(x) ((((unsigned)(x)) & 0x3F) << 0) +#define PACKET3_ATOMIC_MEM__COMMAND(x) ((((unsigned)(x)) & 0xF) << 8) +#define PACKET3_ATOMIC_MEM__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_ATOMIC_MEM__ADDR_LO(x) (((unsigned)(x)) << 0) +#define PACKET3_ATOMIC_MEM__ADDR_HI(x) (((unsigned)(x)) << 0) +#define PACKET3_ATOMIC_MEM__SRC_DATA_LO(x) (((unsigned)(x)) << 0) +#define PACKET3_ATOMIC_MEM__SRC_DATA_HI(x) (((unsigned)(x)) << 0) +#define PACKET3_ATOMIC_MEM__CMP_DATA_LO(x) (((unsigned)(x)) << 0) +#define PACKET3_ATOMIC_MEM__CMP_DATA_HI(x) (((unsigned)(x)) << 0) +#define PACKET3_ATOMIC_MEM__LOOP_INTERVAL(x) ((((unsigned)(x)) & 0x1FFF) << 0) +#define PACKET3_ATOMIC_MEM__COMMAND__SINGLE_PASS_ATOMIC 0 +#define PACKET3_ATOMIC_MEM__COMMAND__LOOP_UNTIL_COMPARE_SATISFIED 1 #define PACKET3_OCCLUSION_QUERY 0x1F #define PACKET3_SET_PREDICATION 0x20 #define PACKET3_REG_RMW 0x21 #define PACKET3_COND_EXEC 0x22 #define PACKET3_PRED_EXEC 0x23 +#define PACKET3_PRED_EXEC__EXEC_COUNT(x) ((((unsigned)(x)) & 0x3FFF) << 0) +#define PACKET3_PRED_EXEC__VIRTUAL_XCC_ID_SELECT(x) ((((unsigned)(x)) & 0xFF) << 24) #define PACKET3_DRAW_INDIRECT 0x24 #define PACKET3_DRAW_INDEX_INDIRECT 0x25 #define PACKET3_INDEX_BASE 0x26 @@ -132,6 +146,28 @@ * 1 - pfp * 2 - ce */ +#define PACKET3_WRITE_DATA__DST_SEL(x) ((((unsigned)(x)) & 0xF) << 8) +#define PACKET3_WRITE_DATA__ADDR_INCR(x) ((((unsigned)(x)) & 0x1) << 16) +#define PACKET3_WRITE_DATA__RESUME_VF_MI300(x) ((((unsigned)(x)) & 0x1) << 19) +#define PACKET3_WRITE_DATA__WR_CONFIRM(x) ((((unsigned)(x)) & 0x1) << 20) +#define PACKET3_WRITE_DATA__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_WRITE_DATA__DST_MMREG_ADDR(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_WRITE_DATA__DST_GDS_ADDR(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_WRITE_DATA__DST_MEM_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_WRITE_DATA__DST_MEM_ADDR_HI(x) ((unsigned)(x)) +#define PACKET3_WRITE_DATA__DST_SEL__MEM_MAPPED_REGISTER 0 +#define PACKET3_WRITE_DATA__DST_SEL__TC_L2 2 +#define PACKET3_WRITE_DATA__DST_SEL__GDS 3 +#define PACKET3_WRITE_DATA__DST_SEL__MEMORY 5 +#define PACKET3_WRITE_DATA__DST_SEL__MEMORY_MAPPED_ADC_PERSISTENT_STATE 6 +#define PACKET3_WRITE_DATA__ADDR_INCR__INCREMENT_ADDRESS 0 +#define PACKET3_WRITE_DATA__ADDR_INCR__DO_NOT_INCREMENT_ADDRESS 1 +#define PACKET3_WRITE_DATA__WR_CONFIRM__DO_NOT_WAIT_FOR_WRITE_CONFIRMATION 0 +#define PACKET3_WRITE_DATA__WR_CONFIRM__WAIT_FOR_WRITE_CONFIRMATION 1 +#define PACKET3_WRITE_DATA__CACHE_POLICY__LRU 0 +#define PACKET3_WRITE_DATA__CACHE_POLICY__STREAM 1 +#define PACKET3_WRITE_DATA__CACHE_POLICY__NOA 2 +#define PACKET3_WRITE_DATA__CACHE_POLICY__BYPASS 3 #define PACKET3_DRAW_INDEX_INDIRECT_MULTI 0x38 #define PACKET3_MEM_SEMAPHORE 0x39 # define PACKET3_SEM_USE_MAILBOX (0x1 << 16) @@ -160,6 +196,33 @@ /* 0 - me * 1 - pfp */ +#define PACKET3_WAIT_REG_MEM__FUNCTION(x) ((((unsigned)(x)) & 0x7) << 0) +#define PACKET3_WAIT_REG_MEM__MEM_SPACE(x) ((((unsigned)(x)) & 0x3) << 4) +#define PACKET3_WAIT_REG_MEM__OPERATION(x) ((((unsigned)(x)) & 0x3) << 6) +#define PACKET3_WAIT_REG_MEM__MES_INTR_PIPE(x) ((((unsigned)(x)) & 0x3) << 22) +#define PACKET3_WAIT_REG_MEM__MES_ACTION(x) ((((unsigned)(x)) & 0x1) << 24) +#define PACKET3_WAIT_REG_MEM__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_WAIT_REG_MEM__MEM_POLL_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_WAIT_REG_MEM__REG_POLL_ADDR(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_WAIT_REG_MEM__REG_WRITE_ADDR1(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_WAIT_REG_MEM__MEM_POLL_ADDR_HI(x) ((unsigned)(x)) +#define PACKET3_WAIT_REG_MEM__REG_WRITE_ADDR2(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_WAIT_REG_MEM__REFERENCE(x) ((unsigned)(x)) +#define PACKET3_WAIT_REG_MEM__MASK(x) ((unsigned)(x)) +#define PACKET3_WAIT_REG_MEM__POLL_INTERVAL(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_WAIT_REG_MEM__OPTIMIZE_ACE_OFFLOAD_MODE(x) ((((unsigned)(x)) & 0x1) << 31) +#define PACKET3_WAIT_REG_MEM__FUNCTION__ALWAYS_PASS 0 +#define PACKET3_WAIT_REG_MEM__FUNCTION__LESS_THAN_REF_VALUE 1 +#define PACKET3_WAIT_REG_MEM__FUNCTION__LESS_THAN_EQUAL_TO_THE_REF_VALUE 2 +#define PACKET3_WAIT_REG_MEM__FUNCTION__EQUAL_TO_THE_REFERENCE_VALUE 3 +#define PACKET3_WAIT_REG_MEM__FUNCTION__NOT_EQUAL_REFERENCE_VALUE 4 +#define PACKET3_WAIT_REG_MEM__FUNCTION__GREATER_THAN_OR_EQUAL_REFERENCE_VALUE 5 +#define PACKET3_WAIT_REG_MEM__FUNCTION__GREATER_THAN_REFERENCE_VALUE 6 +#define PACKET3_WAIT_REG_MEM__MEM_SPACE__REGISTER_SPACE 0 +#define PACKET3_WAIT_REG_MEM__MEM_SPACE__MEMORY_SPACE 1 +#define PACKET3_WAIT_REG_MEM__OPERATION__WAIT_REG_MEM 0 +#define PACKET3_WAIT_REG_MEM__OPERATION__WR_WAIT_WR_REG 1 +#define PACKET3_WAIT_REG_MEM__OPERATION__WAIT_MEM_PREEMPTABLE 3 #define PACKET3_INDIRECT_BUFFER 0x3F #define INDIRECT_BUFFER_VALID (1 << 23) #define INDIRECT_BUFFER_CACHE_POLICY(x) ((x) << 28) @@ -169,7 +232,63 @@ */ #define INDIRECT_BUFFER_PRE_ENB(x) ((x) << 21) #define INDIRECT_BUFFER_PRE_RESUME(x) ((x) << 30) +#define PACKET3_INDIRECT_BUFFER__IB_BASE_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_INDIRECT_BUFFER__IB_BASE_HI(x) ((unsigned)(x)) +#define PACKET3_INDIRECT_BUFFER__IB_SIZE(x) ((((unsigned)(x)) & 0xFFFFF) << 0) +#define PACKET3_INDIRECT_BUFFER__CHAIN(x) ((((unsigned)(x)) & 0x1) << 20) +#define PACKET3_INDIRECT_BUFFER__OFFLOAD_POLLING(x) ((((unsigned)(x)) & 0x1) << 21) +#define PACKET3_INDIRECT_BUFFER__VALID(x) ((((unsigned)(x)) & 0x1) << 23) +#define PACKET3_INDIRECT_BUFFER__VMID(x) ((((unsigned)(x)) & 0xF) << 24) +#define PACKET3_INDIRECT_BUFFER__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 28) +#define PACKET3_INDIRECT_BUFFER__PRIV(x) ((((unsigned)(x)) & 0x1) << 31) +#define PACKET3_INDIRECT_BUFFER__CACHE_POLICY__LRU 0 +#define PACKET3_INDIRECT_BUFFER__CACHE_POLICY__STREAM 1 #define PACKET3_COPY_DATA 0x40 +#define PACKET3_COPY_DATA__SRC_SEL(x) ((((unsigned)(x)) & 0xF) << 0) +#define PACKET3_COPY_DATA__DST_SEL(x) ((((unsigned)(x)) & 0xF) << 8) +#define PACKET3_COPY_DATA__SRC_CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 13) +#define PACKET3_COPY_DATA__COUNT_SEL(x) ((((unsigned)(x)) & 0x1) << 16) +#define PACKET3_COPY_DATA__WR_CONFIRM(x) ((((unsigned)(x)) & 0x1) << 20) +#define PACKET3_COPY_DATA__DST_CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25) +#define PACKET3_COPY_DATA__PQ_EXE_STATUS(x) ((((unsigned)(x)) & 0x1) << 29) +#define PACKET3_COPY_DATA__SRC_REG_OFFSET(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_COPY_DATA__SRC_32B_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_COPY_DATA__SRC_64B_ADDR_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3) +#define PACKET3_COPY_DATA__SRC_GDS_ADDR_LO(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_COPY_DATA__IMM_DATA(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__SRC_MEMTC_ADDR_HI(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__SRC_IMM_DATA(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__DST_REG_OFFSET(x) ((((unsigned)(x)) & 0x3FFFF) << 0) +#define PACKET3_COPY_DATA__DST_32B_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2) +#define PACKET3_COPY_DATA__DST_64B_ADDR_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3) +#define PACKET3_COPY_DATA__DST_GDS_ADDR_LO(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_COPY_DATA__DST_ADDR_HI(x) ((unsigned)(x)) +#define PACKET3_COPY_DATA__SRC_SEL__MEM_MAPPED_REGISTER 0 +#define PACKET3_COPY_DATA__SRC_SEL__MEMORY 1 +#define PACKET3_COPY_DATA__SRC_SEL__TC_L2 2 +#define PACKET3_COPY_DATA__SRC_SEL__GDS 3 +#define PACKET3_COPY_DATA__SRC_SEL__PERFCOUNTERS 4 +#define PACKET3_COPY_DATA__SRC_SEL__IMMEDIATE_DATA 5 +#define PACKET3_COPY_DATA__SRC_SEL__ATOMIC_RETURN_DATA 6 +#define PACKET3_COPY_DATA__SRC_SEL__GDS_ATOMIC_RETURN_DATA0 7 +#define PACKET3_COPY_DATA__SRC_SEL__GDS_ATOMIC_RETURN_DATA1 8 +#define PACKET3_COPY_DATA__SRC_SEL__GPU_CLOCK_COUNT 9 +#define PACKET3_COPY_DATA__DST_SEL__MEM_MAPPED_REGISTER 0 +#define PACKET3_COPY_DATA__DST_SEL__TC_L2 2 +#define PACKET3_COPY_DATA__DST_SEL__GDS 3 +#define PACKET3_COPY_DATA__DST_SEL__PERFCOUNTERS 4 +#define PACKET3_COPY_DATA__DST_SEL__MEMORY 5 +#define PACKET3_COPY_DATA__DST_SEL__MEM_MAPPED_REG_DC 6 +#define PACKET3_COPY_DATA__SRC_CACHE_POLICY__LRU 0 +#define PACKET3_COPY_DATA__SRC_CACHE_POLICY__STREAM 1 +#define PACKET3_COPY_DATA__COUNT_SEL__32_BITS_OF_DATA 0 +#define PACKET3_COPY_DATA__COUNT_SEL__64_BITS_OF_DATA 1 +#define PACKET3_COPY_DATA__WR_CONFIRM__DO_NOT_WAIT_FOR_CONFIRMATION 0 +#define PACKET3_COPY_DATA__WR_CONFIRM__WAIT_FOR_CONFIRMATION 1 +#define PACKET3_COPY_DATA__DST_CACHE_POLICY__LRU 0 +#define PACKET3_COPY_DATA__DST_CACHE_POLICY__STREAM 1 +#define PACKET3_COPY_DATA__PQ_EXE_STATUS__DEFAULT 0 +#define PACKET3_COPY_DATA__PQ_EXE_STATUS__PHASE_UPDATE 1 #define PACKET3_PFP_SYNC_ME 0x42 #define PACKET3_COND_WRITE 0x45 #define PACKET3_EVENT_WRITE 0x46 @@ -181,6 +300,15 @@ * 3 - SAMPLE_STREAMOUTSTAT* * 4 - *S_PARTIAL_FLUSH */ +#define PACKET3_EVENT_WRITE__EVENT_TYPE(x) ((((unsigned)(x)) & 0x3F) << 0) +#define PACKET3_EVENT_WRITE__EVENT_INDEX(x) ((((unsigned)(x)) & 0xF) << 8) +#define PACKET3_EVENT_WRITE__OFFLOAD_ENABLE(x) ((((unsigned)(x)) & 0x1) << 31) +#define PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE(x) ((((unsigned)(x)) & 0x3) << 29) +#define PACKET3_EVENT_WRITE__ADDRESS_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3) +#define PACKET3_EVENT_WRITE__ADDRESS_HI(x) (((unsigned)(x)) << 0) +#define PACKET3_EVENT_WRITE__EVENT_INDEX__OTHER 0 +#define PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_PIPELINESTATS 2 +#define PACKET3_EVENT_WRITE__EVENT_INDEX__CS_PARTIAL_FLUSH 4 #define PACKET3_RELEASE_MEM 0x49 #define EVENT_TYPE(x) ((x) << 0) #define EVENT_INDEX(x) ((x) << 8) @@ -286,6 +414,13 @@ #define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_ICACHE_ACTION_ENA(x) ((x) << 29) #define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_KCACHE_WB_ACTION_ENA(x) ((x) << 30) #define PACKET3_REWIND 0x59 +#define PACKET3_ACQUIRE_MEM__COHER_SIZE(x) ((unsigned)(x)) +#define PACKET3_ACQUIRE_MEM__COHER_SIZE_HI(x) ((((unsigned)(x)) & 0xFF) << 0) +#define PACKET3_ACQUIRE_MEM__COHER_SIZE_HI_VG10(x) ((((unsigned)(x)) & 0xFFFFFF) << 0) +#define PACKET3_ACQUIRE_MEM__COHER_BASE_LO(x) ((unsigned)(x)) +#define PACKET3_ACQUIRE_MEM__COHER_BASE_HI(x) ((((unsigned)(x)) & 0xFFFFFF) << 0) +#define PACKET3_ACQUIRE_MEM__POLL_INTERVAL(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_ACQUIRE_MEM__GCR_CNTL(x) ((((unsigned)(x)) & 0x7FF) << 0) #define PACKET3_LOAD_UCONFIG_REG 0x5E #define PACKET3_LOAD_SH_REG 0x5F #define PACKET3_LOAD_CONFIG_REG 0x60 @@ -300,12 +435,16 @@ #define PACKET3_SET_SH_REG 0x76 #define PACKET3_SET_SH_REG_START 0x00002c00 #define PACKET3_SET_SH_REG_END 0x00003000 +#define PACKET3_SET_SH_REG__REG_OFFSET(x) ((((unsigned)(x)) & 0xFFFF) << 0) +#define PACKET3_SET_SH_REG__VMID_SHIFT(x) ((((unsigned)(x)) & 0x1F) << 23) +#define PACKET3_SET_SH_REG__INDEX(x) ((((unsigned)(x)) & 0xF) << 28) #define PACKET3_SET_SH_REG_OFFSET 0x77 #define PACKET3_SET_QUEUE_REG 0x78 #define PACKET3_SET_UCONFIG_REG 0x79 #define PACKET3_SET_UCONFIG_REG_START 0x0000c000 #define PACKET3_SET_UCONFIG_REG_END 0x0000c400 #define PACKET3_SET_UCONFIG_REG_INDEX_TYPE (2 << 28) +#define PACKET3_SET_UCONFIG_REG__REG_OFFSET(x) ((((unsigned)(x)) & 0xFFFF) << 0) #define PACKET3_SCRATCH_RAM_WRITE 0x7D #define PACKET3_SCRATCH_RAM_READ 0x7E #define PACKET3_LOAD_CONST_RAM 0x80 diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c index 62ad67d0b598..7925cbb61d0d 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc21.c +++ b/drivers/gpu/drm/amd/amdgpu/soc21.c @@ -390,6 +390,7 @@ soc21_asic_reset_method(struct amdgpu_device *adev) case IP_VERSION(14, 0, 0): case IP_VERSION(14, 0, 1): case IP_VERSION(14, 0, 4): + case IP_VERSION(14, 0, 5): return AMD_RESET_METHOD_MODE2; default: if (amdgpu_dpm_is_baco_supported(adev)) @@ -781,6 +782,34 @@ static int soc21_common_early_init(struct amdgpu_ip_block *ip_block) AMD_PG_SUPPORT_GFX_PG; adev->external_rev_id = adev->rev_id + 0x40; break; + case IP_VERSION(11, 5, 3): + adev->cg_flags = AMD_CG_SUPPORT_VCN_MGCG | + AMD_CG_SUPPORT_JPEG_MGCG | + AMD_CG_SUPPORT_GFX_CGCG | + AMD_CG_SUPPORT_GFX_CGLS | + AMD_CG_SUPPORT_GFX_MGCG | + AMD_CG_SUPPORT_GFX_FGCG | + AMD_CG_SUPPORT_REPEATER_FGCG | + AMD_CG_SUPPORT_GFX_PERF_CLK | + AMD_CG_SUPPORT_GFX_3D_CGCG | + AMD_CG_SUPPORT_GFX_3D_CGLS | + AMD_CG_SUPPORT_MC_MGCG | + AMD_CG_SUPPORT_MC_LS | + AMD_CG_SUPPORT_HDP_LS | + AMD_CG_SUPPORT_HDP_DS | + AMD_CG_SUPPORT_HDP_SD | + AMD_CG_SUPPORT_ATHUB_MGCG | + AMD_CG_SUPPORT_ATHUB_LS | + AMD_CG_SUPPORT_IH_CG | + AMD_CG_SUPPORT_BIF_MGCG | + AMD_CG_SUPPORT_BIF_LS; + adev->pg_flags = AMD_PG_SUPPORT_VCN_DPG | + AMD_PG_SUPPORT_VCN | + AMD_PG_SUPPORT_JPEG_DPG | + AMD_PG_SUPPORT_JPEG | + AMD_PG_SUPPORT_GFX_PG; + adev->external_rev_id = adev->rev_id + 0x50; + break; default: /* FIXME: not supported yet */ return -EINVAL; @@ -940,6 +969,7 @@ static int soc21_common_set_clockgating_state(struct amdgpu_ip_block *ip_block, case IP_VERSION(7, 7, 1): case IP_VERSION(7, 11, 0): case IP_VERSION(7, 11, 1): + case IP_VERSION(7, 11, 2): case IP_VERSION(7, 11, 3): adev->nbio.funcs->update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); @@ -972,9 +1002,9 @@ static int soc21_common_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void soc21_common_get_clockgating_state(void *handle, u64 *flags) +static void soc21_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; adev->nbio.funcs->get_clockgating_state(adev, flags); diff --git a/drivers/gpu/drm/amd/amdgpu/soc24.c b/drivers/gpu/drm/amd/amdgpu/soc24.c index 6b8e078ee7c7..4e506c91e978 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc24.c +++ b/drivers/gpu/drm/amd/amdgpu/soc24.c @@ -26,7 +26,6 @@ #include <linux/pci.h> #include "amdgpu.h" -#include "amdgpu_atombios.h" #include "amdgpu_ih.h" #include "amdgpu_uvd.h" #include "amdgpu_vce.h" @@ -575,9 +574,9 @@ static int soc24_common_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void soc24_common_get_clockgating_state(void *handle, u64 *flags) +static void soc24_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; adev->nbio.funcs->get_clockgating_state(adev, flags); diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c index a7b9c358a2d4..74f57b2d30a5 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c @@ -415,6 +415,7 @@ static int umc_v12_0_aca_bank_parser(struct aca_handle *handle, struct aca_bank err_type = ACA_ERROR_TYPE_CE; else return 0; + bank->aca_err_type = err_type; ret = aca_bank_info_decode(bank, &info); if (ret) diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c index 050a0f309390..2295c8713d61 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c @@ -837,9 +837,9 @@ out: return ret; } -static void uvd_v5_0_get_clockgating_state(void *handle, u64 *flags) +static void uvd_v5_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; mutex_lock(&adev->pm.mutex); diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c index d9d036ee51fb..070a0624c2c5 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c @@ -1498,9 +1498,9 @@ out: return ret; } -static void uvd_v6_0_get_clockgating_state(void *handle, u64 *flags) +static void uvd_v6_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; mutex_lock(&adev->pm.mutex); diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index f8bddcd19b68..01248a3982ba 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -828,9 +828,9 @@ out: return ret; } -static void vce_v3_0_get_clockgating_state(void *handle, u64 *flags) +static void vce_v3_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; mutex_lock(&adev->pm.mutex); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index b518202955ca..105e59f6132b 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -452,64 +452,62 @@ static int vcn_v2_5_resume(struct amdgpu_ip_block *ip_block) * vcn_v2_5_mc_resume - memory controller programming * * @adev: amdgpu_device pointer + * @i: instance to resume * * Let the VCN memory controller know it's offsets */ -static void vcn_v2_5_mc_resume(struct amdgpu_device *adev) +static void vcn_v2_5_mc_resume(struct amdgpu_device *adev, int i) { uint32_t size; uint32_t offset; - int i; - - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4); - /* cache window 0: fw */ - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW, - (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo)); - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH, - (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi)); - WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET0, 0); - offset = 0; - } else { - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW, - lower_32_bits(adev->vcn.inst[i].gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH, - upper_32_bits(adev->vcn.inst[i].gpu_addr)); - offset = size; - WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET0, - AMDGPU_UVD_FIRMWARE_OFFSET >> 3); - } - WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE0, size); + if (adev->vcn.harvest_config & (1 << i)) + return; - /* cache window 1: stack */ - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW, - lower_32_bits(adev->vcn.inst[i].gpu_addr + offset)); - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH, - upper_32_bits(adev->vcn.inst[i].gpu_addr + offset)); - WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET1, 0); - WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE); - - /* cache window 2: context */ - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW, - lower_32_bits(adev->vcn.inst[i].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE)); - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH, - upper_32_bits(adev->vcn.inst[i].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE)); - WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET2, 0); - WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE); - - /* non-cache window */ - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW, - lower_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH, - upper_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_VCPU_NONCACHE_OFFSET0, 0); - WREG32_SOC15(VCN, i, mmUVD_VCPU_NONCACHE_SIZE0, - AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared))); + size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4); + /* cache window 0: fw */ + if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW, + (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo)); + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH, + (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi)); + WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET0, 0); + offset = 0; + } else { + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.inst[i].gpu_addr)); + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.inst[i].gpu_addr)); + offset = size; + WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET0, + AMDGPU_UVD_FIRMWARE_OFFSET >> 3); } + WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE0, size); + + /* cache window 1: stack */ + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.inst[i].gpu_addr + offset)); + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.inst[i].gpu_addr + offset)); + WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET1, 0); + WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE); + + /* cache window 2: context */ + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.inst[i].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE)); + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.inst[i].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE)); + WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET2, 0); + WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE); + + /* non-cache window */ + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr)); + WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr)); + WREG32_SOC15(VCN, i, mmUVD_VCPU_NONCACHE_OFFSET0, 0); + WREG32_SOC15(VCN, i, mmUVD_VCPU_NONCACHE_SIZE0, + AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared))); } static void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect) @@ -612,117 +610,115 @@ static void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx * vcn_v2_5_disable_clock_gating - disable VCN clock gating * * @adev: amdgpu_device pointer + * @i: instance to disable clockgating on * * Disable clock gating for VCN block */ -static void vcn_v2_5_disable_clock_gating(struct amdgpu_device *adev) +static void vcn_v2_5_disable_clock_gating(struct amdgpu_device *adev, int i) { uint32_t data; - int i; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - /* UVD disable CGC */ - data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL); - if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG) - data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; - else - data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK; - data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; - data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT; - WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data); - - data = RREG32_SOC15(VCN, i, mmUVD_CGC_GATE); - data &= ~(UVD_CGC_GATE__SYS_MASK - | UVD_CGC_GATE__UDEC_MASK - | UVD_CGC_GATE__MPEG2_MASK - | UVD_CGC_GATE__REGS_MASK - | UVD_CGC_GATE__RBC_MASK - | UVD_CGC_GATE__LMI_MC_MASK - | UVD_CGC_GATE__LMI_UMC_MASK - | UVD_CGC_GATE__IDCT_MASK - | UVD_CGC_GATE__MPRD_MASK - | UVD_CGC_GATE__MPC_MASK - | UVD_CGC_GATE__LBSI_MASK - | UVD_CGC_GATE__LRBBM_MASK - | UVD_CGC_GATE__UDEC_RE_MASK - | UVD_CGC_GATE__UDEC_CM_MASK - | UVD_CGC_GATE__UDEC_IT_MASK - | UVD_CGC_GATE__UDEC_DB_MASK - | UVD_CGC_GATE__UDEC_MP_MASK - | UVD_CGC_GATE__WCB_MASK - | UVD_CGC_GATE__VCPU_MASK - | UVD_CGC_GATE__MMSCH_MASK); - - WREG32_SOC15(VCN, i, mmUVD_CGC_GATE, data); - - SOC15_WAIT_ON_RREG(VCN, i, mmUVD_CGC_GATE, 0, 0xFFFFFFFF); - - data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL); - data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK - | UVD_CGC_CTRL__UDEC_CM_MODE_MASK - | UVD_CGC_CTRL__UDEC_IT_MODE_MASK - | UVD_CGC_CTRL__UDEC_DB_MODE_MASK - | UVD_CGC_CTRL__UDEC_MP_MODE_MASK - | UVD_CGC_CTRL__SYS_MODE_MASK - | UVD_CGC_CTRL__UDEC_MODE_MASK - | UVD_CGC_CTRL__MPEG2_MODE_MASK - | UVD_CGC_CTRL__REGS_MODE_MASK - | UVD_CGC_CTRL__RBC_MODE_MASK - | UVD_CGC_CTRL__LMI_MC_MODE_MASK - | UVD_CGC_CTRL__LMI_UMC_MODE_MASK - | UVD_CGC_CTRL__IDCT_MODE_MASK - | UVD_CGC_CTRL__MPRD_MODE_MASK - | UVD_CGC_CTRL__MPC_MODE_MASK - | UVD_CGC_CTRL__LBSI_MODE_MASK - | UVD_CGC_CTRL__LRBBM_MODE_MASK - | UVD_CGC_CTRL__WCB_MODE_MASK - | UVD_CGC_CTRL__VCPU_MODE_MASK - | UVD_CGC_CTRL__MMSCH_MODE_MASK); - WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data); - - /* turn on */ - data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_GATE); - data |= (UVD_SUVD_CGC_GATE__SRE_MASK - | UVD_SUVD_CGC_GATE__SIT_MASK - | UVD_SUVD_CGC_GATE__SMP_MASK - | UVD_SUVD_CGC_GATE__SCM_MASK - | UVD_SUVD_CGC_GATE__SDB_MASK - | UVD_SUVD_CGC_GATE__SRE_H264_MASK - | UVD_SUVD_CGC_GATE__SRE_HEVC_MASK - | UVD_SUVD_CGC_GATE__SIT_H264_MASK - | UVD_SUVD_CGC_GATE__SIT_HEVC_MASK - | UVD_SUVD_CGC_GATE__SCM_H264_MASK - | UVD_SUVD_CGC_GATE__SCM_HEVC_MASK - | UVD_SUVD_CGC_GATE__SDB_H264_MASK - | UVD_SUVD_CGC_GATE__SDB_HEVC_MASK - | UVD_SUVD_CGC_GATE__SCLR_MASK - | UVD_SUVD_CGC_GATE__UVD_SC_MASK - | UVD_SUVD_CGC_GATE__ENT_MASK - | UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK - | UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK - | UVD_SUVD_CGC_GATE__SITE_MASK - | UVD_SUVD_CGC_GATE__SRE_VP9_MASK - | UVD_SUVD_CGC_GATE__SCM_VP9_MASK - | UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK - | UVD_SUVD_CGC_GATE__SDB_VP9_MASK - | UVD_SUVD_CGC_GATE__IME_HEVC_MASK); - WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_GATE, data); - - data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL); - data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK - | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK - | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK - | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK - | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK - | UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK - | UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK - | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK + if (adev->vcn.harvest_config & (1 << i)) + return; + /* UVD disable CGC */ + data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL); + if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG) + data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + else + data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK; + data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; + data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT; + WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data); + + data = RREG32_SOC15(VCN, i, mmUVD_CGC_GATE); + data &= ~(UVD_CGC_GATE__SYS_MASK + | UVD_CGC_GATE__UDEC_MASK + | UVD_CGC_GATE__MPEG2_MASK + | UVD_CGC_GATE__REGS_MASK + | UVD_CGC_GATE__RBC_MASK + | UVD_CGC_GATE__LMI_MC_MASK + | UVD_CGC_GATE__LMI_UMC_MASK + | UVD_CGC_GATE__IDCT_MASK + | UVD_CGC_GATE__MPRD_MASK + | UVD_CGC_GATE__MPC_MASK + | UVD_CGC_GATE__LBSI_MASK + | UVD_CGC_GATE__LRBBM_MASK + | UVD_CGC_GATE__UDEC_RE_MASK + | UVD_CGC_GATE__UDEC_CM_MASK + | UVD_CGC_GATE__UDEC_IT_MASK + | UVD_CGC_GATE__UDEC_DB_MASK + | UVD_CGC_GATE__UDEC_MP_MASK + | UVD_CGC_GATE__WCB_MASK + | UVD_CGC_GATE__VCPU_MASK + | UVD_CGC_GATE__MMSCH_MASK); + + WREG32_SOC15(VCN, i, mmUVD_CGC_GATE, data); + + SOC15_WAIT_ON_RREG(VCN, i, mmUVD_CGC_GATE, 0, 0xFFFFFFFF); + + data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL); + data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK + | UVD_CGC_CTRL__UDEC_CM_MODE_MASK + | UVD_CGC_CTRL__UDEC_IT_MODE_MASK + | UVD_CGC_CTRL__UDEC_DB_MODE_MASK + | UVD_CGC_CTRL__UDEC_MP_MODE_MASK + | UVD_CGC_CTRL__SYS_MODE_MASK + | UVD_CGC_CTRL__UDEC_MODE_MASK + | UVD_CGC_CTRL__MPEG2_MODE_MASK + | UVD_CGC_CTRL__REGS_MODE_MASK + | UVD_CGC_CTRL__RBC_MODE_MASK + | UVD_CGC_CTRL__LMI_MC_MODE_MASK + | UVD_CGC_CTRL__LMI_UMC_MODE_MASK + | UVD_CGC_CTRL__IDCT_MODE_MASK + | UVD_CGC_CTRL__MPRD_MODE_MASK + | UVD_CGC_CTRL__MPC_MODE_MASK + | UVD_CGC_CTRL__LBSI_MODE_MASK + | UVD_CGC_CTRL__LRBBM_MODE_MASK + | UVD_CGC_CTRL__WCB_MODE_MASK + | UVD_CGC_CTRL__VCPU_MODE_MASK + | UVD_CGC_CTRL__MMSCH_MODE_MASK); + WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data); + + /* turn on */ + data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_GATE); + data |= (UVD_SUVD_CGC_GATE__SRE_MASK + | UVD_SUVD_CGC_GATE__SIT_MASK + | UVD_SUVD_CGC_GATE__SMP_MASK + | UVD_SUVD_CGC_GATE__SCM_MASK + | UVD_SUVD_CGC_GATE__SDB_MASK + | UVD_SUVD_CGC_GATE__SRE_H264_MASK + | UVD_SUVD_CGC_GATE__SRE_HEVC_MASK + | UVD_SUVD_CGC_GATE__SIT_H264_MASK + | UVD_SUVD_CGC_GATE__SIT_HEVC_MASK + | UVD_SUVD_CGC_GATE__SCM_H264_MASK + | UVD_SUVD_CGC_GATE__SCM_HEVC_MASK + | UVD_SUVD_CGC_GATE__SDB_H264_MASK + | UVD_SUVD_CGC_GATE__SDB_HEVC_MASK + | UVD_SUVD_CGC_GATE__SCLR_MASK + | UVD_SUVD_CGC_GATE__UVD_SC_MASK + | UVD_SUVD_CGC_GATE__ENT_MASK + | UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK + | UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK + | UVD_SUVD_CGC_GATE__SITE_MASK + | UVD_SUVD_CGC_GATE__SRE_VP9_MASK + | UVD_SUVD_CGC_GATE__SCM_VP9_MASK + | UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK + | UVD_SUVD_CGC_GATE__SDB_VP9_MASK + | UVD_SUVD_CGC_GATE__IME_HEVC_MASK); + WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_GATE, data); + + data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL); + data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK + | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK + | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK + | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK + | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK + | UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK + | UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK + | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK | UVD_SUVD_CGC_CTRL__IME_MODE_MASK - | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK); - WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL, data); - } + | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK); + WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL, data); } static void vcn_v2_5_clock_gating_dpg_mode(struct amdgpu_device *adev, @@ -777,62 +773,60 @@ static void vcn_v2_5_clock_gating_dpg_mode(struct amdgpu_device *adev, * vcn_v2_5_enable_clock_gating - enable VCN clock gating * * @adev: amdgpu_device pointer + * @i: instance to enable clockgating on * * Enable clock gating for VCN block */ -static void vcn_v2_5_enable_clock_gating(struct amdgpu_device *adev) +static void vcn_v2_5_enable_clock_gating(struct amdgpu_device *adev, int i) { uint32_t data = 0; - int i; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - /* enable UVD CGC */ - data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL); - if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG) - data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; - else - data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; - data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; - data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT; - WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data); - - data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL); - data |= (UVD_CGC_CTRL__UDEC_RE_MODE_MASK - | UVD_CGC_CTRL__UDEC_CM_MODE_MASK - | UVD_CGC_CTRL__UDEC_IT_MODE_MASK - | UVD_CGC_CTRL__UDEC_DB_MODE_MASK - | UVD_CGC_CTRL__UDEC_MP_MODE_MASK - | UVD_CGC_CTRL__SYS_MODE_MASK - | UVD_CGC_CTRL__UDEC_MODE_MASK - | UVD_CGC_CTRL__MPEG2_MODE_MASK - | UVD_CGC_CTRL__REGS_MODE_MASK - | UVD_CGC_CTRL__RBC_MODE_MASK - | UVD_CGC_CTRL__LMI_MC_MODE_MASK - | UVD_CGC_CTRL__LMI_UMC_MODE_MASK - | UVD_CGC_CTRL__IDCT_MODE_MASK - | UVD_CGC_CTRL__MPRD_MODE_MASK - | UVD_CGC_CTRL__MPC_MODE_MASK - | UVD_CGC_CTRL__LBSI_MODE_MASK - | UVD_CGC_CTRL__LRBBM_MODE_MASK - | UVD_CGC_CTRL__WCB_MODE_MASK - | UVD_CGC_CTRL__VCPU_MODE_MASK); - WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data); - - data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL); - data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK - | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK - | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK - | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK - | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK - | UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK - | UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK - | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK - | UVD_SUVD_CGC_CTRL__IME_MODE_MASK - | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK); - WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL, data); - } + if (adev->vcn.harvest_config & (1 << i)) + return; + /* enable UVD CGC */ + data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL); + if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG) + data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + else + data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; + data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT; + WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data); + + data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL); + data |= (UVD_CGC_CTRL__UDEC_RE_MODE_MASK + | UVD_CGC_CTRL__UDEC_CM_MODE_MASK + | UVD_CGC_CTRL__UDEC_IT_MODE_MASK + | UVD_CGC_CTRL__UDEC_DB_MODE_MASK + | UVD_CGC_CTRL__UDEC_MP_MODE_MASK + | UVD_CGC_CTRL__SYS_MODE_MASK + | UVD_CGC_CTRL__UDEC_MODE_MASK + | UVD_CGC_CTRL__MPEG2_MODE_MASK + | UVD_CGC_CTRL__REGS_MODE_MASK + | UVD_CGC_CTRL__RBC_MODE_MASK + | UVD_CGC_CTRL__LMI_MC_MODE_MASK + | UVD_CGC_CTRL__LMI_UMC_MODE_MASK + | UVD_CGC_CTRL__IDCT_MODE_MASK + | UVD_CGC_CTRL__MPRD_MODE_MASK + | UVD_CGC_CTRL__MPC_MODE_MASK + | UVD_CGC_CTRL__LBSI_MODE_MASK + | UVD_CGC_CTRL__LRBBM_MODE_MASK + | UVD_CGC_CTRL__WCB_MODE_MASK + | UVD_CGC_CTRL__VCPU_MODE_MASK); + WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data); + + data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL); + data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK + | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK + | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK + | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK + | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK + | UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK + | UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK + | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK + | UVD_SUVD_CGC_CTRL__IME_MODE_MASK + | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK); + WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL, data); } static void vcn_v2_6_enable_ras(struct amdgpu_device *adev, int inst_idx, @@ -1006,197 +1000,185 @@ static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo return 0; } -static int vcn_v2_5_start(struct amdgpu_device *adev) +static int vcn_v2_5_start(struct amdgpu_device *adev, int i) { + volatile struct amdgpu_fw_shared *fw_shared = + adev->vcn.inst[i].fw_shared.cpu_addr; struct amdgpu_ring *ring; uint32_t rb_bufsz, tmp; - int i, j, k, r; + int j, k, r; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, true, i); - } + if (adev->vcn.harvest_config & (1 << i)) + return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v2_5_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - continue; - } + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, true, i); - /* disable register anti-hang mechanism */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_POWER_STATUS), 0, - ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK); + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + return vcn_v2_5_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - /* set uvd status busy */ - tmp = RREG32_SOC15(VCN, i, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY; - WREG32_SOC15(VCN, i, mmUVD_STATUS, tmp); - } + /* disable register anti-hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_POWER_STATUS), 0, + ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK); + + /* set uvd status busy */ + tmp = RREG32_SOC15(VCN, i, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY; + WREG32_SOC15(VCN, i, mmUVD_STATUS, tmp); if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) return 0; - /*SW clock gating */ - vcn_v2_5_disable_clock_gating(adev); + /* SW clock gating */ + vcn_v2_5_disable_clock_gating(adev, i); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - /* enable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), - UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); - - /* disable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), 0, - ~UVD_MASTINT_EN__VCPU_EN_MASK); - - /* setup mmUVD_LMI_CTRL */ - tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL); - tmp &= ~0xff; - WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL, tmp | 0x8| - UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | - UVD_LMI_CTRL__MASK_MC_URGENT_MASK | - UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | - UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); - - /* setup mmUVD_MPC_CNTL */ - tmp = RREG32_SOC15(VCN, i, mmUVD_MPC_CNTL); - tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; - tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; - WREG32_SOC15(VCN, i, mmUVD_MPC_CNTL, tmp); - - /* setup UVD_MPC_SET_MUXA0 */ - WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXA0, - ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); - - /* setup UVD_MPC_SET_MUXB0 */ - WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXB0, - ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); - - /* setup mmUVD_MPC_SET_MUX */ - WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUX, - ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | - (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); - } + /* enable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), + UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); - vcn_v2_5_mc_resume(adev); + /* disable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), 0, + ~UVD_MASTINT_EN__VCPU_EN_MASK); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - if (adev->vcn.harvest_config & (1 << i)) - continue; - /* VCN global tiling registers */ - WREG32_SOC15(VCN, i, mmUVD_GFX8_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - WREG32_SOC15(VCN, i, mmUVD_GFX8_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); + /* setup mmUVD_LMI_CTRL */ + tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL); + tmp &= ~0xff; + WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL, tmp | 0x8| + UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); + + /* setup mmUVD_MPC_CNTL */ + tmp = RREG32_SOC15(VCN, i, mmUVD_MPC_CNTL); + tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; + tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; + WREG32_SOC15(VCN, i, mmUVD_MPC_CNTL, tmp); + + /* setup UVD_MPC_SET_MUXA0 */ + WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXA0, + ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); + + /* setup UVD_MPC_SET_MUXB0 */ + WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXB0, + ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); + + /* setup mmUVD_MPC_SET_MUX */ + WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUX, + ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | + (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); + + vcn_v2_5_mc_resume(adev, i); - /* enable LMI MC and UMC channels */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); + /* VCN global tiling registers */ + WREG32_SOC15(VCN, i, mmUVD_GFX8_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(VCN, i, mmUVD_GFX8_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); - /* unblock VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), 0, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + /* enable LMI MC and UMC channels */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0, + ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - for (k = 0; k < 10; ++k) { - uint32_t status; - - for (j = 0; j < 100; ++j) { - status = RREG32_SOC15(VCN, i, mmUVD_STATUS); - if (status & 2) - break; - if (amdgpu_emu_mode == 1) - msleep(500); - else - mdelay(10); - } - r = 0; + /* unblock VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), 0, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + for (k = 0; k < 10; ++k) { + uint32_t status; + + for (j = 0; j < 100; ++j) { + status = RREG32_SOC15(VCN, i, mmUVD_STATUS); if (status & 2) break; + if (amdgpu_emu_mode == 1) + msleep(500); + else + mdelay(10); + } + r = 0; + if (status & 2) + break; - DRM_ERROR("VCN decode not responding, trying to reset the VCPU!!!\n"); - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - mdelay(10); - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); + DRM_ERROR("VCN decode not responding, trying to reset the VCPU!!!\n"); + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + mdelay(10); + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); - mdelay(10); - r = -1; - } + mdelay(10); + r = -1; + } - if (r) { - DRM_ERROR("VCN decode not responding, giving up!!!\n"); - return r; - } + if (r) { + DRM_ERROR("VCN decode not responding, giving up!!!\n"); + return r; + } - /* enable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), - UVD_MASTINT_EN__VCPU_EN_MASK, - ~UVD_MASTINT_EN__VCPU_EN_MASK); + /* enable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), + UVD_MASTINT_EN__VCPU_EN_MASK, + ~UVD_MASTINT_EN__VCPU_EN_MASK); - /* clear the busy bit of VCN_STATUS */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS), 0, - ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); + /* clear the busy bit of VCN_STATUS */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS), 0, + ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); - WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_VMID, 0); + WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_VMID, 0); - ring = &adev->vcn.inst[i].ring_dec; - /* force RBC into idle state */ - rb_bufsz = order_base_2(ring->ring_size); - tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz); - tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1); - tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1); - tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1); - tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1); - WREG32_SOC15(VCN, i, mmUVD_RBC_RB_CNTL, tmp); + ring = &adev->vcn.inst[i].ring_dec; + /* force RBC into idle state */ + rb_bufsz = order_base_2(ring->ring_size); + tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1); + WREG32_SOC15(VCN, i, mmUVD_RBC_RB_CNTL, tmp); - fw_shared->multi_queue.decode_queue_mode |= FW_QUEUE_RING_RESET; - /* program the RB_BASE for ring buffer */ - WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW, - lower_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH, - upper_32_bits(ring->gpu_addr)); + fw_shared->multi_queue.decode_queue_mode |= FW_QUEUE_RING_RESET; + /* program the RB_BASE for ring buffer */ + WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW, + lower_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH, + upper_32_bits(ring->gpu_addr)); - /* Initialize the ring buffer's read and write pointers */ - WREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR, 0); + /* Initialize the ring buffer's read and write pointers */ + WREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR, 0); - ring->wptr = RREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR); - WREG32_SOC15(VCN, i, mmUVD_RBC_RB_WPTR, - lower_32_bits(ring->wptr)); - fw_shared->multi_queue.decode_queue_mode &= ~FW_QUEUE_RING_RESET; + ring->wptr = RREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR); + WREG32_SOC15(VCN, i, mmUVD_RBC_RB_WPTR, + lower_32_bits(ring->wptr)); + fw_shared->multi_queue.decode_queue_mode &= ~FW_QUEUE_RING_RESET; - fw_shared->multi_queue.encode_generalpurpose_queue_mode |= FW_QUEUE_RING_RESET; - ring = &adev->vcn.inst[i].ring_enc[0]; - WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr)); - WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr)); - WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr); - WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4); - fw_shared->multi_queue.encode_generalpurpose_queue_mode &= ~FW_QUEUE_RING_RESET; - - fw_shared->multi_queue.encode_lowlatency_queue_mode |= FW_QUEUE_RING_RESET; - ring = &adev->vcn.inst[i].ring_enc[1]; - WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr)); - WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr)); - WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr); - WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4); - fw_shared->multi_queue.encode_lowlatency_queue_mode &= ~FW_QUEUE_RING_RESET; - } + fw_shared->multi_queue.encode_generalpurpose_queue_mode |= FW_QUEUE_RING_RESET; + ring = &adev->vcn.inst[i].ring_enc[0]; + WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr)); + WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr)); + WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr); + WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4); + fw_shared->multi_queue.encode_generalpurpose_queue_mode &= ~FW_QUEUE_RING_RESET; + + fw_shared->multi_queue.encode_lowlatency_queue_mode |= FW_QUEUE_RING_RESET; + ring = &adev->vcn.inst[i].ring_enc[1]; + WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr)); + WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr)); + WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr); + WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4); + fw_shared->multi_queue.encode_lowlatency_queue_mode &= ~FW_QUEUE_RING_RESET; return 0; } @@ -1425,72 +1407,66 @@ static int vcn_v2_5_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) return 0; } -static int vcn_v2_5_stop(struct amdgpu_device *adev) +static int vcn_v2_5_stop(struct amdgpu_device *adev, int i) { uint32_t tmp; - int i, r = 0; + int r; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v2_5_stop_dpg_mode(adev, i); - continue; - } + if (adev->vcn.harvest_config & (1 << i)) + return 0; + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + return vcn_v2_5_stop_dpg_mode(adev, i); - /* wait for vcn idle */ - r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7); - if (r) - return r; + /* wait for vcn idle */ + r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7); + if (r) + return r; - tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | - UVD_LMI_STATUS__READ_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | + UVD_LMI_STATUS__READ_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); + if (r) + return r; - /* block LMI UMC channel */ - tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2); - tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; - WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2, tmp); + /* block LMI UMC channel */ + tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2); + tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; + WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2, tmp); - tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK| - UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK| + UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); + if (r) + return r; - /* block VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), - UVD_RB_ARB_CTRL__VCPU_DIS_MASK, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + /* block VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), + UVD_RB_ARB_CTRL__VCPU_DIS_MASK, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - /* reset VCPU */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); + /* reset VCPU */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); - /* disable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, - ~(UVD_VCPU_CNTL__CLK_EN_MASK)); + /* disable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, + ~(UVD_VCPU_CNTL__CLK_EN_MASK)); - /* clear status */ - WREG32_SOC15(VCN, i, mmUVD_STATUS, 0); + /* clear status */ + WREG32_SOC15(VCN, i, mmUVD_STATUS, 0); - vcn_v2_5_enable_clock_gating(adev); + vcn_v2_5_enable_clock_gating(adev, i); - /* enable register anti-hang mechanism */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_POWER_STATUS), - UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, - ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK); - } + /* enable register anti-hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_POWER_STATUS), + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, + ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, false, i); - } + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, false, i); return 0; } @@ -1811,16 +1787,19 @@ static int vcn_v2_5_set_clockgating_state(struct amdgpu_ip_block *ip_block, { struct amdgpu_device *adev = ip_block->adev; bool enable = (state == AMD_CG_STATE_GATE); + int i; if (amdgpu_sriov_vf(adev)) return 0; - if (enable) { - if (!vcn_v2_5_is_idle(adev)) - return -EBUSY; - vcn_v2_5_enable_clock_gating(adev); - } else { - vcn_v2_5_disable_clock_gating(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (enable) { + if (!vcn_v2_5_is_idle(adev)) + return -EBUSY; + vcn_v2_5_enable_clock_gating(adev, i); + } else { + vcn_v2_5_disable_clock_gating(adev, i); + } } return 0; @@ -1830,20 +1809,22 @@ static int vcn_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state) { struct amdgpu_device *adev = ip_block->adev; - int ret; + int ret = 0, i; if (amdgpu_sriov_vf(adev)) return 0; - if(state == adev->vcn.cur_state) + if (state == adev->vcn.cur_state) return 0; - if (state == AMD_PG_STATE_GATE) - ret = vcn_v2_5_stop(adev); - else - ret = vcn_v2_5_start(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (state == AMD_PG_STATE_GATE) + ret |= vcn_v2_5_stop(adev, i); + else + ret |= vcn_v2_5_start(adev, i); + } - if(!ret) + if (!ret) adev->vcn.cur_state = state; return ret; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c index ecdc027f8220..75211366f8f6 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c @@ -98,8 +98,7 @@ static void vcn_v4_0_3_enable_ras(struct amdgpu_device *adev, static inline bool vcn_v4_0_3_normalizn_reqd(struct amdgpu_device *adev) { - return (amdgpu_sriov_vf(adev) || - (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4))); + return (adev->vcn.caps & AMDGPU_VCN_CAPS(RRMT_ENABLED)) == 0; } /** @@ -295,6 +294,11 @@ static int vcn_v4_0_3_hw_init(struct amdgpu_ip_block *ip_block) ring->sched.ready = true; } } else { + /* This flag is not set for VF, assumed to be disabled always */ + if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) & + 0x100) + adev->vcn.caps |= AMDGPU_VCN_CAPS(RRMT_ENABLED); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { struct amdgpu_vcn4_fw_shared *fw_shared; @@ -1124,11 +1128,6 @@ static int vcn_v4_0_3_start(struct amdgpu_device *adev) uint32_t tmp; for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, true, i); - } - - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { r = vcn_v4_0_3_start_dpg_mode(adev, i, adev->vcn.indirect_sram); continue; @@ -1399,11 +1398,6 @@ static int vcn_v4_0_3_stop(struct amdgpu_device *adev) vcn_v4_0_3_enable_clock_gating(adev, i); } Done: - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, false, i); - } - return 0; } @@ -1931,10 +1925,12 @@ static int vcn_v4_0_3_aca_bank_parser(struct aca_handle *handle, struct aca_bank misc0 = bank->regs[ACA_REG_IDX_MISC0]; switch (type) { case ACA_SMU_TYPE_UE: + bank->aca_err_type = ACA_ERROR_TYPE_UE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE, 1ULL); break; case ACA_SMU_TYPE_CE: + bank->aca_err_type = ACA_ERROR_TYPE_CE; ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_CE, ACA_REG__MISC0__ERRCNT(misc0)); break; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c index 23d3c16c9d9f..5d757e7de9db 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c @@ -1397,7 +1397,7 @@ static void vcn_v4_0_5_unified_ring_set_wptr(struct amdgpu_ring *ring) } } -static const struct amdgpu_ring_funcs vcn_v4_0_5_unified_ring_vm_funcs = { +static struct amdgpu_ring_funcs vcn_v4_0_5_unified_ring_vm_funcs = { .type = AMDGPU_RING_TYPE_VCN_ENC, .align_mask = 0x3f, .nop = VCN_ENC_CMD_NO_OP, @@ -1441,6 +1441,9 @@ static void vcn_v4_0_5_set_unified_ring_funcs(struct amdgpu_device *adev) if (adev->vcn.harvest_config & (1 << i)) continue; + if (amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(4, 0, 5)) + vcn_v4_0_5_unified_ring_vm_funcs.secure_submission_supported = true; + adev->vcn.inst[i].ring_enc[0].funcs = &vcn_v4_0_5_unified_ring_vm_funcs; adev->vcn.inst[i].ring_enc[0].me = i; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c index 8b463c977d08..288a77179036 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c @@ -575,9 +575,6 @@ static int vcn_v5_0_1_start(struct amdgpu_device *adev) uint32_t tmp; int i, j, k, r, vcn_inst; - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_uvd(adev, true); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; @@ -816,9 +813,6 @@ static int vcn_v5_0_1_stop(struct amdgpu_device *adev) WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, 0); } - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_uvd(adev, false); - return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index 06615f160331..12fe571787f4 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -1994,9 +1994,9 @@ static int vi_common_set_powergating_state(struct amdgpu_ip_block *ip_block, return 0; } -static void vi_common_get_clockgating_state(void *handle, u64 *flags) +static void vi_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) |