diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu')
161 files changed, 6440 insertions, 5040 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index 8d0748184a14..7fedbb725e17 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/drivers/gpu/drm/amd/amdgpu/Makefile @@ -45,7 +45,7 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \ amdgpu_atombios.o atombios_crtc.o amdgpu_connectors.o \ atom.o amdgpu_fence.o amdgpu_ttm.o amdgpu_object.o amdgpu_gart.o \ amdgpu_encoders.o amdgpu_display.o amdgpu_i2c.o \ - amdgpu_fb.o amdgpu_gem.o amdgpu_ring.o \ + amdgpu_gem.o amdgpu_ring.o \ amdgpu_cs.o amdgpu_bios.o amdgpu_benchmark.o amdgpu_test.o \ atombios_dp.o amdgpu_afmt.o amdgpu_trace_points.o \ atombios_encoders.o amdgpu_sa.o atombios_i2c.o \ @@ -73,10 +73,8 @@ amdgpu-$(CONFIG_DRM_AMDGPU_SI)+= si.o gmc_v6_0.o gfx_v6_0.o si_ih.o si_dma.o dce amdgpu-y += \ vi.o mxgpu_vi.o nbio_v6_1.o soc15.o emu_soc.o mxgpu_ai.o nbio_v7_0.o vega10_reg_init.o \ - vega20_reg_init.o nbio_v7_4.o nbio_v2_3.o nv.o navi10_reg_init.o navi14_reg_init.o \ - arct_reg_init.o navi12_reg_init.o mxgpu_nv.o sienna_cichlid_reg_init.o vangogh_reg_init.o \ - nbio_v7_2.o dimgrey_cavefish_reg_init.o hdp_v4_0.o hdp_v5_0.o aldebaran_reg_init.o aldebaran.o \ - beige_goby_reg_init.o yellow_carp_reg_init.o cyan_skillfish_reg_init.o + vega20_reg_init.o nbio_v7_4.o nbio_v2_3.o nv.o arct_reg_init.o mxgpu_nv.o \ + nbio_v7_2.o hdp_v4_0.o hdp_v5_0.o aldebaran_reg_init.o aldebaran.o # add DF block amdgpu-y += \ diff --git a/drivers/gpu/drm/amd/amdgpu/aldebaran.c b/drivers/gpu/drm/amd/amdgpu/aldebaran.c index 148f6c3343ab..bcfdb63b1d42 100644 --- a/drivers/gpu/drm/amd/amdgpu/aldebaran.c +++ b/drivers/gpu/drm/amd/amdgpu/aldebaran.c @@ -307,6 +307,8 @@ static int aldebaran_mode2_restore_ip(struct amdgpu_device *adev) adev->ip_blocks[i].status.late_initialized = true; } + amdgpu_ras_set_error_query_ready(adev, true); + amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE); amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 269437b01328..d8b854fcbffa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -205,6 +205,7 @@ extern struct amdgpu_mgpu_info mgpu_info; extern int amdgpu_ras_enable; extern uint amdgpu_ras_mask; extern int amdgpu_bad_page_threshold; +extern bool amdgpu_ignore_bad_page_threshold; extern struct amdgpu_watchdog_timer amdgpu_watchdog_timer; extern int amdgpu_async_gfx_ring; extern int amdgpu_mcbp; @@ -457,7 +458,6 @@ struct amdgpu_flip_work { uint64_t base; struct drm_pending_vblank_event *event; struct amdgpu_bo *old_abo; - struct dma_fence *excl; unsigned shared_count; struct dma_fence **shared; struct dma_fence_cb cb; @@ -744,6 +744,7 @@ enum amd_hw_ip_block_type { UVD_HWIP, VCN_HWIP = UVD_HWIP, JPEG_HWIP = VCN_HWIP, + VCN1_HWIP, VCE_HWIP, DF_HWIP, DCE_HWIP, @@ -755,11 +756,16 @@ enum amd_hw_ip_block_type { CLK_HWIP, UMC_HWIP, RSMU_HWIP, + XGMI_HWIP, + DCI_HWIP, MAX_HWIP }; #define HWIP_MAX_INSTANCE 10 +#define HW_ID_MAX 300 +#define IP_VERSION(mj, mn, rv) (((mj) << 16) | ((mn) << 8) | (rv)) + struct amd_powerplay { void *pp_handle; const struct amd_pm_funcs *pp_funcs; @@ -806,6 +812,7 @@ struct amd_powerplay { #define AMDGPU_RESET_MAGIC_NUM 64 #define AMDGPU_MAX_DF_PERFMONS 4 +#define AMDGPU_PRODUCT_NAME_LEN 64 struct amdgpu_device { struct device *dev; struct pci_dev *pdev; @@ -830,6 +837,7 @@ struct amdgpu_device { struct notifier_block acpi_nb; struct amdgpu_i2c_chan *i2c_bus[AMDGPU_MAX_I2C_BUS]; struct debugfs_blob_wrapper debugfs_vbios_blob; + struct debugfs_blob_wrapper debugfs_discovery_blob; struct mutex srbm_mutex; /* GRBM index mutex. Protects concurrent access to GRBM index */ struct mutex grbm_idx_mutex; @@ -1069,17 +1077,16 @@ struct amdgpu_device { bool runpm; bool in_runpm; bool has_pr3; + bool is_fw_fb; bool pm_sysfs_en; bool ucode_sysfs_en; /* Chip product information */ char product_number[16]; - char product_name[32]; + char product_name[AMDGPU_PRODUCT_NAME_LEN]; char serial[20]; - struct amdgpu_autodump autodump; - atomic_t throttling_logging_enabled; struct ratelimit_state throttling_logging_rs; uint32_t ras_hw_enabled; @@ -1090,6 +1097,9 @@ struct amdgpu_device { pci_channel_state_t pci_channel_state; struct amdgpu_reset_control *reset_cntl; + uint32_t ip_versions[MAX_HWIP][HWIP_MAX_INSTANCE]; + + bool ram_is_direct_mapped; }; static inline struct amdgpu_device *drm_to_adev(struct drm_device *ddev) @@ -1310,6 +1320,8 @@ void amdgpu_device_flush_hdp(struct amdgpu_device *adev, void amdgpu_device_invalidate_hdp(struct amdgpu_device *adev, struct amdgpu_ring *ring); +void amdgpu_device_halt(struct amdgpu_device *adev); + /* atpx handler */ #if defined(CONFIG_VGA_SWITCHEROO) void amdgpu_register_atpx_handler(void); @@ -1353,8 +1365,6 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon); u32 amdgpu_get_vblank_counter_kms(struct drm_crtc *crtc); int amdgpu_enable_vblank_kms(struct drm_crtc *crtc); void amdgpu_disable_vblank_kms(struct drm_crtc *crtc); -long amdgpu_kms_compat_ioctl(struct file *filp, unsigned int cmd, - unsigned long arg); int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 1d41c2c00623..776a947b45df 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -31,6 +31,8 @@ #include <linux/dma-buf.h> #include "amdgpu_xgmi.h" #include <uapi/linux/kfd_ioctl.h> +#include "amdgpu_ras.h" +#include "amdgpu_umc.h" /* Total memory size in system memory and all GPU VRAM. Used to * estimate worst case amount of memory to reserve for page tables @@ -70,8 +72,7 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev) if (!kfd_initialized) return; - adev->kfd.dev = kgd2kfd_probe((struct kgd_dev *)adev, - adev->pdev, adev->asic_type, vf); + adev->kfd.dev = kgd2kfd_probe(adev, vf); if (adev->kfd.dev) amdgpu_amdkfd_total_mem_size += adev->gmc.real_vram_size; @@ -232,19 +233,16 @@ int amdgpu_amdkfd_post_reset(struct amdgpu_device *adev) return r; } -void amdgpu_amdkfd_gpu_reset(struct kgd_dev *kgd) +void amdgpu_amdkfd_gpu_reset(struct amdgpu_device *adev) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - if (amdgpu_device_should_recover_gpu(adev)) amdgpu_device_gpu_recover(adev, NULL); } -int amdgpu_amdkfd_alloc_gtt_mem(struct kgd_dev *kgd, size_t size, +int amdgpu_amdkfd_alloc_gtt_mem(struct amdgpu_device *adev, size_t size, void **mem_obj, uint64_t *gpu_addr, void **cpu_ptr, bool cp_mqd_gfx9) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; struct amdgpu_bo *bo = NULL; struct amdgpu_bo_param bp; int r; @@ -313,7 +311,7 @@ allocate_mem_reserve_bo_failed: return r; } -void amdgpu_amdkfd_free_gtt_mem(struct kgd_dev *kgd, void *mem_obj) +void amdgpu_amdkfd_free_gtt_mem(struct amdgpu_device *adev, void *mem_obj) { struct amdgpu_bo *bo = (struct amdgpu_bo *) mem_obj; @@ -324,10 +322,9 @@ void amdgpu_amdkfd_free_gtt_mem(struct kgd_dev *kgd, void *mem_obj) amdgpu_bo_unref(&(bo)); } -int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size, +int amdgpu_amdkfd_alloc_gws(struct amdgpu_device *adev, size_t size, void **mem_obj) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; struct amdgpu_bo *bo = NULL; struct amdgpu_bo_user *ubo; struct amdgpu_bo_param bp; @@ -354,18 +351,16 @@ int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size, return 0; } -void amdgpu_amdkfd_free_gws(struct kgd_dev *kgd, void *mem_obj) +void amdgpu_amdkfd_free_gws(struct amdgpu_device *adev, void *mem_obj) { struct amdgpu_bo *bo = (struct amdgpu_bo *)mem_obj; amdgpu_bo_unref(&bo); } -uint32_t amdgpu_amdkfd_get_fw_version(struct kgd_dev *kgd, +uint32_t amdgpu_amdkfd_get_fw_version(struct amdgpu_device *adev, enum kgd_engine_type type) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - switch (type) { case KGD_ENGINE_PFP: return adev->gfx.pfp_fw_version; @@ -398,11 +393,9 @@ uint32_t amdgpu_amdkfd_get_fw_version(struct kgd_dev *kgd, return 0; } -void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd, +void amdgpu_amdkfd_get_local_mem_info(struct amdgpu_device *adev, struct kfd_local_mem_info *mem_info) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - memset(mem_info, 0, sizeof(*mem_info)); mem_info->local_mem_size_public = adev->gmc.visible_vram_size; @@ -427,19 +420,15 @@ void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd, mem_info->mem_clk_max = 100; } -uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct kgd_dev *kgd) +uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct amdgpu_device *adev) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - if (adev->gfx.funcs->get_gpu_clock_counter) return adev->gfx.funcs->get_gpu_clock_counter(adev); return 0; } -uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct kgd_dev *kgd) +uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct amdgpu_device *adev) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - /* the sclk is in quantas of 10kHz */ if (amdgpu_sriov_vf(adev)) return adev->clock.default_sclk / 100; @@ -449,9 +438,8 @@ uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct kgd_dev *kgd) return 100; } -void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info) +void amdgpu_amdkfd_get_cu_info(struct amdgpu_device *adev, struct kfd_cu_info *cu_info) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; struct amdgpu_cu_info acu_info = adev->gfx.cu_info; memset(cu_info, 0, sizeof(*cu_info)); @@ -472,13 +460,12 @@ void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info) cu_info->lds_size = acu_info.lds_size; } -int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd, - struct kgd_dev **dma_buf_kgd, +int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd, + struct amdgpu_device **dmabuf_adev, uint64_t *bo_size, void *metadata_buffer, size_t buffer_size, uint32_t *metadata_size, uint32_t *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; struct dma_buf *dma_buf; struct drm_gem_object *obj; struct amdgpu_bo *bo; @@ -506,8 +493,8 @@ int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd, goto out_put; r = 0; - if (dma_buf_kgd) - *dma_buf_kgd = (struct kgd_dev *)adev; + if (dmabuf_adev) + *dmabuf_adev = adev; if (bo_size) *bo_size = amdgpu_bo_size(bo); if (metadata_buffer) @@ -527,32 +514,18 @@ out_put: return r; } -uint64_t amdgpu_amdkfd_get_vram_usage(struct kgd_dev *kgd) +uint64_t amdgpu_amdkfd_get_vram_usage(struct amdgpu_device *adev) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; struct ttm_resource_manager *vram_man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM); return amdgpu_vram_mgr_usage(vram_man); } -uint64_t amdgpu_amdkfd_get_hive_id(struct kgd_dev *kgd) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - - return adev->gmc.xgmi.hive_id; -} - -uint64_t amdgpu_amdkfd_get_unique_id(struct kgd_dev *kgd) +uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct amdgpu_device *dst, + struct amdgpu_device *src) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - - return adev->unique_id; -} - -uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct kgd_dev *dst, struct kgd_dev *src) -{ - struct amdgpu_device *peer_adev = (struct amdgpu_device *)src; - struct amdgpu_device *adev = (struct amdgpu_device *)dst; + struct amdgpu_device *peer_adev = src; + struct amdgpu_device *adev = dst; int ret = amdgpu_xgmi_get_hops_count(adev, peer_adev); if (ret < 0) { @@ -564,16 +537,18 @@ uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct kgd_dev *dst, struct kgd_dev *s return (uint8_t)ret; } -int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct kgd_dev *dst, struct kgd_dev *src, bool is_min) +int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct amdgpu_device *dst, + struct amdgpu_device *src, + bool is_min) { - struct amdgpu_device *adev = (struct amdgpu_device *)dst, *peer_adev; + struct amdgpu_device *adev = dst, *peer_adev; int num_links; if (adev->asic_type != CHIP_ALDEBARAN) return 0; if (src) - peer_adev = (struct amdgpu_device *)src; + peer_adev = src; /* num links returns 0 for indirect peers since indirect route is unknown. */ num_links = is_min ? 1 : amdgpu_xgmi_get_num_links(adev, peer_adev); @@ -588,9 +563,8 @@ int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct kgd_dev *dst, struct kgd_dev return (num_links * 16 * 25000)/BITS_PER_BYTE; } -int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min) +int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct amdgpu_device *adev, bool is_min) { - struct amdgpu_device *adev = (struct amdgpu_device *)dev; int num_lanes_shift = (is_min ? ffs(adev->pm.pcie_mlw_mask) : fls(adev->pm.pcie_mlw_mask)) - 1; int gen_speed_shift = (is_min ? ffs(adev->pm.pcie_gen_mask & @@ -646,39 +620,11 @@ int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min) return (num_lanes_factor * gen_speed_mbits_factor)/BITS_PER_BYTE; } -uint64_t amdgpu_amdkfd_get_mmio_remap_phys_addr(struct kgd_dev *kgd) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - - return adev->rmmio_remap.bus_addr; -} - -uint32_t amdgpu_amdkfd_get_num_gws(struct kgd_dev *kgd) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - - return adev->gds.gws_size; -} - -uint32_t amdgpu_amdkfd_get_asic_rev_id(struct kgd_dev *kgd) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - - return adev->rev_id; -} - -int amdgpu_amdkfd_get_noretry(struct kgd_dev *kgd) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - - return adev->gmc.noretry; -} - -int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine, +int amdgpu_amdkfd_submit_ib(struct amdgpu_device *adev, + enum kgd_engine_type engine, uint32_t vmid, uint64_t gpu_addr, uint32_t *ib_cmd, uint32_t ib_len) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; struct amdgpu_job *job; struct amdgpu_ib *ib; struct amdgpu_ring *ring; @@ -729,10 +675,8 @@ err: return ret; } -void amdgpu_amdkfd_set_compute_idle(struct kgd_dev *kgd, bool idle) +void amdgpu_amdkfd_set_compute_idle(struct amdgpu_device *adev, bool idle) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_COMPUTE, !idle); @@ -746,10 +690,9 @@ bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid) return false; } -int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid) +int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct amdgpu_device *adev, + uint16_t vmid) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - if (adev->family == AMDGPU_FAMILY_AI) { int i; @@ -762,10 +705,9 @@ int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid) return 0; } -int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid, - enum TLB_FLUSH_TYPE flush_type) +int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct amdgpu_device *adev, + uint16_t pasid, enum TLB_FLUSH_TYPE flush_type) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; bool all_hub = false; if (adev->family == AMDGPU_FAMILY_AI) @@ -774,9 +716,18 @@ int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid, return amdgpu_gmc_flush_gpu_tlb_pasid(adev, pasid, flush_type, all_hub); } -bool amdgpu_amdkfd_have_atomics_support(struct kgd_dev *kgd) +bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - return adev->have_atomics_support; } + +void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev, bool reset) +{ + struct ras_err_data err_data = {0, 0, 0, NULL}; + + /* CPU MCA will handle page retirement if connected_to_cpu is 1 */ + if (!adev->gmc.xgmi.connected_to_cpu) + amdgpu_umc_poison_handler(adev, &err_data, reset); + else if (reset) + amdgpu_amdkfd_gpu_reset(adev); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index 3bc52b2c604f..61f899e54fd5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -144,14 +144,16 @@ void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev, void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev); void amdgpu_amdkfd_device_init(struct amdgpu_device *adev); void amdgpu_amdkfd_device_fini_sw(struct amdgpu_device *adev); -int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine, +int amdgpu_amdkfd_submit_ib(struct amdgpu_device *adev, + enum kgd_engine_type engine, uint32_t vmid, uint64_t gpu_addr, uint32_t *ib_cmd, uint32_t ib_len); -void amdgpu_amdkfd_set_compute_idle(struct kgd_dev *kgd, bool idle); -bool amdgpu_amdkfd_have_atomics_support(struct kgd_dev *kgd); -int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid); -int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid, - enum TLB_FLUSH_TYPE flush_type); +void amdgpu_amdkfd_set_compute_idle(struct amdgpu_device *adev, bool idle); +bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev); +int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct amdgpu_device *adev, + uint16_t vmid); +int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct amdgpu_device *adev, + uint16_t pasid, enum TLB_FLUSH_TYPE flush_type); bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid); @@ -159,7 +161,7 @@ int amdgpu_amdkfd_pre_reset(struct amdgpu_device *adev); int amdgpu_amdkfd_post_reset(struct amdgpu_device *adev); -void amdgpu_amdkfd_gpu_reset(struct kgd_dev *kgd); +void amdgpu_amdkfd_gpu_reset(struct amdgpu_device *adev); int amdgpu_queue_mask_bit_to_set_resource_bit(struct amdgpu_device *adev, int queue_bit); @@ -198,37 +200,36 @@ int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm) } #endif /* Shared API */ -int amdgpu_amdkfd_alloc_gtt_mem(struct kgd_dev *kgd, size_t size, +int amdgpu_amdkfd_alloc_gtt_mem(struct amdgpu_device *adev, size_t size, void **mem_obj, uint64_t *gpu_addr, void **cpu_ptr, bool mqd_gfx9); -void amdgpu_amdkfd_free_gtt_mem(struct kgd_dev *kgd, void *mem_obj); -int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size, void **mem_obj); -void amdgpu_amdkfd_free_gws(struct kgd_dev *kgd, void *mem_obj); +void amdgpu_amdkfd_free_gtt_mem(struct amdgpu_device *adev, void *mem_obj); +int amdgpu_amdkfd_alloc_gws(struct amdgpu_device *adev, size_t size, + void **mem_obj); +void amdgpu_amdkfd_free_gws(struct amdgpu_device *adev, void *mem_obj); int amdgpu_amdkfd_add_gws_to_process(void *info, void *gws, struct kgd_mem **mem); int amdgpu_amdkfd_remove_gws_from_process(void *info, void *mem); -uint32_t amdgpu_amdkfd_get_fw_version(struct kgd_dev *kgd, +uint32_t amdgpu_amdkfd_get_fw_version(struct amdgpu_device *adev, enum kgd_engine_type type); -void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd, +void amdgpu_amdkfd_get_local_mem_info(struct amdgpu_device *adev, struct kfd_local_mem_info *mem_info); -uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct kgd_dev *kgd); +uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct amdgpu_device *adev); -uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct kgd_dev *kgd); -void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info); -int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd, - struct kgd_dev **dmabuf_kgd, +uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct amdgpu_device *adev); +void amdgpu_amdkfd_get_cu_info(struct amdgpu_device *adev, + struct kfd_cu_info *cu_info); +int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd, + struct amdgpu_device **dmabuf_adev, uint64_t *bo_size, void *metadata_buffer, size_t buffer_size, uint32_t *metadata_size, uint32_t *flags); -uint64_t amdgpu_amdkfd_get_vram_usage(struct kgd_dev *kgd); -uint64_t amdgpu_amdkfd_get_hive_id(struct kgd_dev *kgd); -uint64_t amdgpu_amdkfd_get_unique_id(struct kgd_dev *kgd); -uint64_t amdgpu_amdkfd_get_mmio_remap_phys_addr(struct kgd_dev *kgd); -uint32_t amdgpu_amdkfd_get_num_gws(struct kgd_dev *kgd); -uint32_t amdgpu_amdkfd_get_asic_rev_id(struct kgd_dev *kgd); -int amdgpu_amdkfd_get_noretry(struct kgd_dev *kgd); -uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct kgd_dev *dst, struct kgd_dev *src); -int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct kgd_dev *dst, struct kgd_dev *src, bool is_min); -int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min); +uint64_t amdgpu_amdkfd_get_vram_usage(struct amdgpu_device *adev); +uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct amdgpu_device *dst, + struct amdgpu_device *src); +int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct amdgpu_device *dst, + struct amdgpu_device *src, + bool is_min); +int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct amdgpu_device *adev, bool is_min); /* Read user wptr from a specified user address space with page fault * disabled. The memory must be pinned and mapped to the hardware when @@ -258,43 +259,56 @@ int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min); (&((struct amdgpu_fpriv *) \ ((struct drm_file *)(drm_priv))->driver_priv)->vm) -int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct kgd_dev *kgd, +int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev, struct file *filp, u32 pasid, void **process_info, struct dma_fence **ef); -void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *drm_priv); +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); int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( - struct kgd_dev *kgd, uint64_t va, uint64_t size, + struct amdgpu_device *adev, uint64_t va, uint64_t size, void *drm_priv, struct kgd_mem **mem, uint64_t *offset, uint32_t flags); int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( - struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv, + struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv, uint64_t *size); int amdgpu_amdkfd_gpuvm_map_memory_to_gpu( - struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv, bool *table_freed); + struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv, + bool *table_freed); int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu( - struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv); + struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv); int amdgpu_amdkfd_gpuvm_sync_memory( - struct kgd_dev *kgd, struct kgd_mem *mem, bool intr); -int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd, + struct amdgpu_device *adev, struct kgd_mem *mem, bool intr); +int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct amdgpu_device *adev, struct kgd_mem *mem, void **kptr, uint64_t *size); +void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct amdgpu_device *adev, + struct kgd_mem *mem); + int amdgpu_amdkfd_gpuvm_restore_process_bos(void *process_info, struct dma_fence **ef); -int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd, +int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct amdgpu_device *adev, struct kfd_vm_fault_info *info); -int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd, +int amdgpu_amdkfd_gpuvm_import_dmabuf(struct amdgpu_device *adev, struct dma_buf *dmabuf, uint64_t va, void *drm_priv, struct kgd_mem **mem, uint64_t *size, uint64_t *mmap_offset); -int amdgpu_amdkfd_get_tile_config(struct kgd_dev *kgd, +int amdgpu_amdkfd_get_tile_config(struct amdgpu_device *adev, struct tile_config *config); +void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev, + bool reset); #if IS_ENABLED(CONFIG_HSA_AMD) void amdgpu_amdkfd_gpuvm_init_mem_limits(void); void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev, struct amdgpu_vm *vm); -void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo); + +/** + * @amdgpu_amdkfd_release_notify() - Notify KFD when GEM object is released + * + * Allows KFD to release its resources associated with the GEM object. + */ +void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo); void amdgpu_amdkfd_reserve_system_mem(uint64_t size); #else static inline @@ -309,7 +323,7 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev, } static inline -void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo) +void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo) { } #endif @@ -321,8 +335,7 @@ int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm, #if IS_ENABLED(CONFIG_HSA_AMD) int kgd2kfd_init(void); void kgd2kfd_exit(void); -struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, struct pci_dev *pdev, - unsigned int asic_type, bool vf); +struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf); bool kgd2kfd_device_init(struct kfd_dev *kfd, struct drm_device *ddev, const struct kgd2kfd_shared_resources *gpu_resources); @@ -346,8 +359,7 @@ static inline void kgd2kfd_exit(void) } static inline -struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, struct pci_dev *pdev, - unsigned int asic_type, bool vf) +struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf) { return NULL; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c index 5a7f680bcb3f..abe93b3ff765 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c @@ -57,11 +57,6 @@ (*dump)[i++][1] = RREG32(addr); \ } while (0) -static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd) -{ - return (struct amdgpu_device *)kgd; -} - static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd) { return (struct v9_sdma_mqd *)mqd; @@ -123,10 +118,9 @@ static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev, return sdma_rlc_reg_offset; } -int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, +int kgd_arcturus_hqd_sdma_load(struct amdgpu_device *adev, void *mqd, uint32_t __user *wptr, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; unsigned long end_jiffies; @@ -193,11 +187,10 @@ int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, return 0; } -int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd, +int kgd_arcturus_hqd_sdma_dump(struct amdgpu_device *adev, uint32_t engine_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, engine_id, queue_id); uint32_t i = 0, reg; @@ -225,9 +218,9 @@ int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd, return 0; } -bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) +bool kgd_arcturus_hqd_sdma_is_occupied(struct amdgpu_device *adev, + void *mqd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t sdma_rlc_rb_cntl; @@ -244,10 +237,9 @@ bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) return false; } -int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, +int kgd_arcturus_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, unsigned int utimeout) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t temp; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h index ce08131b7b5f..756c1a5679c0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h @@ -20,11 +20,12 @@ * OTHER DEALINGS IN THE SOFTWARE. */ -int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, +int kgd_arcturus_hqd_sdma_load(struct amdgpu_device *adev, void *mqd, uint32_t __user *wptr, struct mm_struct *mm); -int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd, +int kgd_arcturus_hqd_sdma_dump(struct amdgpu_device *adev, uint32_t engine_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs); -bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd); -int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, +bool kgd_arcturus_hqd_sdma_is_occupied(struct amdgpu_device *adev, + void *mqd); +int kgd_arcturus_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, unsigned int utimeout); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c index 960acf68150a..7b7f4b2764c1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c @@ -39,37 +39,26 @@ enum hqd_dequeue_request_type { SAVE_WAVES }; -static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd) -{ - return (struct amdgpu_device *)kgd; -} - -static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe, +static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe, uint32_t queue, uint32_t vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - mutex_lock(&adev->srbm_mutex); nv_grbm_select(adev, mec, pipe, queue, vmid); } -static void unlock_srbm(struct kgd_dev *kgd) +static void unlock_srbm(struct amdgpu_device *adev) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - nv_grbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); } -static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id, +static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, queue_id, 0); + lock_srbm(adev, mec, pipe, queue_id, 0); } static uint64_t get_queue_mask(struct amdgpu_device *adev, @@ -81,33 +70,29 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev, return 1ull << bit; } -static void release_queue(struct kgd_dev *kgd) +static void release_queue(struct amdgpu_device *adev) { - unlock_srbm(kgd); + unlock_srbm(adev); } -static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid, +static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid, uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, sh_mem_config); WREG32_SOC15(GC, 0, mmSH_MEM_BASES, sh_mem_bases); /* APE1 no longer exists on GFX9 */ - unlock_srbm(kgd); + unlock_srbm(adev); } -static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid, +static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, unsigned int vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - /* * We have to assume that there is no outstanding mapping. * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because @@ -150,22 +135,21 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid, * but still works */ -static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id) +static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t mec; uint32_t pipe; mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, 0, 0); + lock_srbm(adev, mec, pipe, 0, 0); WREG32_SOC15(GC, 0, mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK | CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK); - unlock_srbm(kgd); + unlock_srbm(adev); return 0; } @@ -218,12 +202,11 @@ static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd) return (struct v10_sdma_mqd *)mqd; } -static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, - uint32_t queue_id, uint32_t __user *wptr, - uint32_t wptr_shift, uint32_t wptr_mask, - struct mm_struct *mm) +static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd, + uint32_t pipe_id, uint32_t queue_id, + uint32_t __user *wptr, uint32_t wptr_shift, + uint32_t wptr_mask, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v10_compute_mqd *m; uint32_t *mqd_hqd; uint32_t reg, hqd_base, data; @@ -231,7 +214,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, m = get_mqd(mqd); pr_debug("Load hqd of pipe %d queue %d\n", pipe_id, queue_id); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); /* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */ mqd_hqd = &m->cp_mqd_base_addr_lo; @@ -296,16 +279,15 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1); WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, data); - release_queue(kgd); + release_queue(adev); return 0; } -static int kgd_hiq_mqd_load(struct kgd_dev *kgd, void *mqd, +static int kgd_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t doorbell_off) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; struct v10_compute_mqd *m; uint32_t mec, pipe; @@ -313,7 +295,7 @@ static int kgd_hiq_mqd_load(struct kgd_dev *kgd, void *mqd, m = get_mqd(mqd); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); @@ -349,16 +331,15 @@ static int kgd_hiq_mqd_load(struct kgd_dev *kgd, void *mqd, out_unlock: spin_unlock(&adev->gfx.kiq.ring_lock); - release_queue(kgd); + release_queue(adev); return r; } -static int kgd_hqd_dump(struct kgd_dev *kgd, +static int kgd_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t i = 0, reg; #define HQD_N_REGS 56 #define DUMP_REG(addr) do { \ @@ -372,13 +353,13 @@ static int kgd_hqd_dump(struct kgd_dev *kgd, if (*dump == NULL) return -ENOMEM; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR); reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++) DUMP_REG(reg); - release_queue(kgd); + release_queue(adev); WARN_ON_ONCE(i != HQD_N_REGS); *n_regs = i; @@ -386,10 +367,9 @@ static int kgd_hqd_dump(struct kgd_dev *kgd, return 0; } -static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd, uint32_t __user *wptr, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v10_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; unsigned long end_jiffies; @@ -456,11 +436,10 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, return 0; } -static int kgd_hqd_sdma_dump(struct kgd_dev *kgd, +static int kgd_hqd_sdma_dump(struct amdgpu_device *adev, uint32_t engine_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, engine_id, queue_id); uint32_t i = 0, reg; @@ -488,15 +467,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd, return 0; } -static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address, - uint32_t pipe_id, uint32_t queue_id) +static bool kgd_hqd_is_occupied(struct amdgpu_device *adev, + uint64_t queue_address, uint32_t pipe_id, + uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t act; bool retval = false; uint32_t low, high; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); act = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE); if (act) { low = lower_32_bits(queue_address >> 8); @@ -506,13 +485,12 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address, high == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI)) retval = true; } - release_queue(kgd); + release_queue(adev); return retval; } -static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) +static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v10_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t sdma_rlc_rb_cntl; @@ -529,12 +507,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) return false; } -static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); enum hqd_dequeue_request_type type; unsigned long end_jiffies; uint32_t temp; @@ -548,7 +525,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd, int retry; #endif - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); if (m->cp_hqd_vmid == 0) WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0); @@ -633,20 +610,19 @@ loop: break; if (time_after(jiffies, end_jiffies)) { pr_err("cp queue preemption time out.\n"); - release_queue(kgd); + release_queue(adev); return -ETIME; } usleep_range(500, 1000); } - release_queue(kgd); + release_queue(adev); return 0; } -static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, unsigned int utimeout) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v10_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t temp; @@ -683,11 +659,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, return 0; } -static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd, +static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev, uint8_t vmid, uint16_t *p_pasid) { uint32_t value; - struct amdgpu_device *adev = (struct amdgpu_device *) kgd; value = RREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING) + vmid); @@ -696,12 +671,12 @@ static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd, return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK); } -static int kgd_address_watch_disable(struct kgd_dev *kgd) +static int kgd_address_watch_disable(struct amdgpu_device *adev) { return 0; } -static int kgd_address_watch_execute(struct kgd_dev *kgd, +static int kgd_address_watch_execute(struct amdgpu_device *adev, unsigned int watch_point_id, uint32_t cntl_val, uint32_t addr_hi, @@ -710,11 +685,10 @@ static int kgd_address_watch_execute(struct kgd_dev *kgd, return 0; } -static int kgd_wave_control_execute(struct kgd_dev *kgd, +static int kgd_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, uint32_t sq_cmd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t data = 0; mutex_lock(&adev->grbm_idx_mutex); @@ -735,18 +709,16 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd, return 0; } -static uint32_t kgd_address_watch_get_offset(struct kgd_dev *kgd, +static uint32_t kgd_address_watch_get_offset(struct amdgpu_device *adev, unsigned int watch_point_id, unsigned int reg_offset) { return 0; } -static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, - uint64_t page_table_base) +static void set_vm_context_page_table_base(struct amdgpu_device *adev, + uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) { pr_err("trying to set page table base for wrong VMID %u\n", vmid); @@ -757,12 +729,10 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base); } -static void program_trap_handler_settings(struct kgd_dev *kgd, +static void program_trap_handler_settings(struct amdgpu_device *adev, uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); /* * Program TBA registers @@ -781,7 +751,7 @@ static void program_trap_handler_settings(struct kgd_dev *kgd, WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI), upper_32_bits(tma_addr >> 8)); - unlock_srbm(kgd); + unlock_srbm(adev); } const struct kfd2kgd_calls gfx_v10_kfd2kgd = { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c index dac0d751d5af..1f37d3574001 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c @@ -38,37 +38,26 @@ enum hqd_dequeue_request_type { SAVE_WAVES }; -static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd) -{ - return (struct amdgpu_device *)kgd; -} - -static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe, +static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe, uint32_t queue, uint32_t vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - mutex_lock(&adev->srbm_mutex); nv_grbm_select(adev, mec, pipe, queue, vmid); } -static void unlock_srbm(struct kgd_dev *kgd) +static void unlock_srbm(struct amdgpu_device *adev) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - nv_grbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); } -static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id, +static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, queue_id, 0); + lock_srbm(adev, mec, pipe, queue_id, 0); } static uint64_t get_queue_mask(struct amdgpu_device *adev, @@ -80,34 +69,30 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev, return 1ull << bit; } -static void release_queue(struct kgd_dev *kgd) +static void release_queue(struct amdgpu_device *adev) { - unlock_srbm(kgd); + unlock_srbm(adev); } -static void program_sh_mem_settings_v10_3(struct kgd_dev *kgd, uint32_t vmid, +static void program_sh_mem_settings_v10_3(struct amdgpu_device *adev, uint32_t vmid, uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, sh_mem_config); WREG32_SOC15(GC, 0, mmSH_MEM_BASES, sh_mem_bases); /* APE1 no longer exists on GFX9 */ - unlock_srbm(kgd); + unlock_srbm(adev); } /* ATC is defeatured on Sienna_Cichlid */ -static int set_pasid_vmid_mapping_v10_3(struct kgd_dev *kgd, unsigned int pasid, +static int set_pasid_vmid_mapping_v10_3(struct amdgpu_device *adev, unsigned int pasid, unsigned int vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - uint32_t value = pasid << IH_VMID_0_LUT__PASID__SHIFT; /* Mapping vmid to pasid also for IH block */ @@ -118,22 +103,21 @@ static int set_pasid_vmid_mapping_v10_3(struct kgd_dev *kgd, unsigned int pasid, return 0; } -static int init_interrupts_v10_3(struct kgd_dev *kgd, uint32_t pipe_id) +static int init_interrupts_v10_3(struct amdgpu_device *adev, uint32_t pipe_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t mec; uint32_t pipe; mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, 0, 0); + lock_srbm(adev, mec, pipe, 0, 0); WREG32_SOC15(GC, 0, mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK | CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK); - unlock_srbm(kgd); + unlock_srbm(adev); return 0; } @@ -188,12 +172,11 @@ static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd) return (struct v10_sdma_mqd *)mqd; } -static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, - uint32_t queue_id, uint32_t __user *wptr, - uint32_t wptr_shift, uint32_t wptr_mask, - struct mm_struct *mm) +static int hqd_load_v10_3(struct amdgpu_device *adev, void *mqd, + uint32_t pipe_id, uint32_t queue_id, + uint32_t __user *wptr, uint32_t wptr_shift, + uint32_t wptr_mask, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v10_compute_mqd *m; uint32_t *mqd_hqd; uint32_t reg, hqd_base, data; @@ -201,7 +184,7 @@ static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, m = get_mqd(mqd); pr_debug("Load hqd of pipe %d queue %d\n", pipe_id, queue_id); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); /* HIQ is set during driver init period with vmid set to 0*/ if (m->cp_hqd_vmid == 0) { @@ -281,16 +264,15 @@ static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1); WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, data); - release_queue(kgd); + release_queue(adev); return 0; } -static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd, +static int hiq_mqd_load_v10_3(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t doorbell_off) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; struct v10_compute_mqd *m; uint32_t mec, pipe; @@ -298,7 +280,7 @@ static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd, m = get_mqd(mqd); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); @@ -334,16 +316,15 @@ static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd, out_unlock: spin_unlock(&adev->gfx.kiq.ring_lock); - release_queue(kgd); + release_queue(adev); return r; } -static int hqd_dump_v10_3(struct kgd_dev *kgd, +static int hqd_dump_v10_3(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t i = 0, reg; #define HQD_N_REGS 56 #define DUMP_REG(addr) do { \ @@ -357,13 +338,13 @@ static int hqd_dump_v10_3(struct kgd_dev *kgd, if (*dump == NULL) return -ENOMEM; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR); reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++) DUMP_REG(reg); - release_queue(kgd); + release_queue(adev); WARN_ON_ONCE(i != HQD_N_REGS); *n_regs = i; @@ -371,10 +352,9 @@ static int hqd_dump_v10_3(struct kgd_dev *kgd, return 0; } -static int hqd_sdma_load_v10_3(struct kgd_dev *kgd, void *mqd, +static int hqd_sdma_load_v10_3(struct amdgpu_device *adev, void *mqd, uint32_t __user *wptr, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v10_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; unsigned long end_jiffies; @@ -441,11 +421,10 @@ static int hqd_sdma_load_v10_3(struct kgd_dev *kgd, void *mqd, return 0; } -static int hqd_sdma_dump_v10_3(struct kgd_dev *kgd, +static int hqd_sdma_dump_v10_3(struct amdgpu_device *adev, uint32_t engine_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, engine_id, queue_id); uint32_t i = 0, reg; @@ -473,15 +452,15 @@ static int hqd_sdma_dump_v10_3(struct kgd_dev *kgd, return 0; } -static bool hqd_is_occupied_v10_3(struct kgd_dev *kgd, uint64_t queue_address, - uint32_t pipe_id, uint32_t queue_id) +static bool hqd_is_occupied_v10_3(struct amdgpu_device *adev, + uint64_t queue_address, uint32_t pipe_id, + uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t act; bool retval = false; uint32_t low, high; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); act = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE); if (act) { low = lower_32_bits(queue_address >> 8); @@ -491,13 +470,13 @@ static bool hqd_is_occupied_v10_3(struct kgd_dev *kgd, uint64_t queue_address, high == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI)) retval = true; } - release_queue(kgd); + release_queue(adev); return retval; } -static bool hqd_sdma_is_occupied_v10_3(struct kgd_dev *kgd, void *mqd) +static bool hqd_sdma_is_occupied_v10_3(struct amdgpu_device *adev, + void *mqd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v10_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t sdma_rlc_rb_cntl; @@ -514,18 +493,17 @@ static bool hqd_sdma_is_occupied_v10_3(struct kgd_dev *kgd, void *mqd) return false; } -static int hqd_destroy_v10_3(struct kgd_dev *kgd, void *mqd, +static int hqd_destroy_v10_3(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); enum hqd_dequeue_request_type type; unsigned long end_jiffies; uint32_t temp; struct v10_compute_mqd *m = get_mqd(mqd); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); if (m->cp_hqd_vmid == 0) WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0); @@ -555,20 +533,19 @@ static int hqd_destroy_v10_3(struct kgd_dev *kgd, void *mqd, if (time_after(jiffies, end_jiffies)) { pr_err("cp queue pipe %d queue %d preemption failed\n", pipe_id, queue_id); - release_queue(kgd); + release_queue(adev); return -ETIME; } usleep_range(500, 1000); } - release_queue(kgd); + release_queue(adev); return 0; } -static int hqd_sdma_destroy_v10_3(struct kgd_dev *kgd, void *mqd, +static int hqd_sdma_destroy_v10_3(struct amdgpu_device *adev, void *mqd, unsigned int utimeout) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v10_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t temp; @@ -606,12 +583,12 @@ static int hqd_sdma_destroy_v10_3(struct kgd_dev *kgd, void *mqd, } -static int address_watch_disable_v10_3(struct kgd_dev *kgd) +static int address_watch_disable_v10_3(struct amdgpu_device *adev) { return 0; } -static int address_watch_execute_v10_3(struct kgd_dev *kgd, +static int address_watch_execute_v10_3(struct amdgpu_device *adev, unsigned int watch_point_id, uint32_t cntl_val, uint32_t addr_hi, @@ -620,11 +597,10 @@ static int address_watch_execute_v10_3(struct kgd_dev *kgd, return 0; } -static int wave_control_execute_v10_3(struct kgd_dev *kgd, +static int wave_control_execute_v10_3(struct amdgpu_device *adev, uint32_t gfx_index_val, uint32_t sq_cmd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t data = 0; mutex_lock(&adev->grbm_idx_mutex); @@ -645,28 +621,24 @@ static int wave_control_execute_v10_3(struct kgd_dev *kgd, return 0; } -static uint32_t address_watch_get_offset_v10_3(struct kgd_dev *kgd, +static uint32_t address_watch_get_offset_v10_3(struct amdgpu_device *adev, unsigned int watch_point_id, unsigned int reg_offset) { return 0; } -static void set_vm_context_page_table_base_v10_3(struct kgd_dev *kgd, uint32_t vmid, - uint64_t page_table_base) +static void set_vm_context_page_table_base_v10_3(struct amdgpu_device *adev, + uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - /* SDMA is on gfxhub as well for Navi1* series */ adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base); } -static void program_trap_handler_settings_v10_3(struct kgd_dev *kgd, +static void program_trap_handler_settings_v10_3(struct amdgpu_device *adev, uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); /* * Program TBA registers @@ -685,15 +657,14 @@ static void program_trap_handler_settings_v10_3(struct kgd_dev *kgd, WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI), upper_32_bits(tma_addr >> 8)); - unlock_srbm(kgd); + unlock_srbm(adev); } #if 0 -uint32_t enable_debug_trap_v10_3(struct kgd_dev *kgd, +uint32_t enable_debug_trap_v10_3(struct amdgpu_device *adev, uint32_t trap_debug_wave_launch_mode, uint32_t vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t data = 0; uint32_t orig_wave_cntl_value; uint32_t orig_stall_vmid; @@ -720,10 +691,8 @@ uint32_t enable_debug_trap_v10_3(struct kgd_dev *kgd, return 0; } -uint32_t disable_debug_trap_v10_3(struct kgd_dev *kgd) +uint32_t disable_debug_trap_v10_3(struct amdgpu_device *adev) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - mutex_lock(&adev->grbm_idx_mutex); WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); @@ -733,11 +702,10 @@ uint32_t disable_debug_trap_v10_3(struct kgd_dev *kgd) return 0; } -uint32_t set_wave_launch_trap_override_v10_3(struct kgd_dev *kgd, +uint32_t set_wave_launch_trap_override_v10_3(struct amdgpu_device *adev, uint32_t trap_override, uint32_t trap_mask) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t data = 0; mutex_lock(&adev->grbm_idx_mutex); @@ -762,11 +730,10 @@ uint32_t set_wave_launch_trap_override_v10_3(struct kgd_dev *kgd, return 0; } -uint32_t set_wave_launch_mode_v10_3(struct kgd_dev *kgd, +uint32_t set_wave_launch_mode_v10_3(struct amdgpu_device *adev, uint8_t wave_launch_mode, uint32_t vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t data = 0; bool is_stall_mode; bool is_mode_set; @@ -805,16 +772,14 @@ uint32_t set_wave_launch_mode_v10_3(struct kgd_dev *kgd, * sem_rearm_wait_time -- Wait Count for Semaphore re-arm. * deq_retry_wait_time -- Wait Count for Global Wave Syncs. */ -void get_iq_wait_times_v10_3(struct kgd_dev *kgd, +void get_iq_wait_times_v10_3(struct amdgpu_device *adev, uint32_t *wait_times) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - *wait_times = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2)); } -void build_grace_period_packet_info_v10_3(struct kgd_dev *kgd, +void build_grace_period_packet_info_v10_3(struct amdgpu_device *adev, uint32_t wait_times, uint32_t grace_period, uint32_t *reg_offset, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c index b91d27e39bad..36528dad7684 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c @@ -82,68 +82,54 @@ union TCP_WATCH_CNTL_BITS { float f32All; }; -static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd) -{ - return (struct amdgpu_device *)kgd; -} - -static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe, +static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe, uint32_t queue, uint32_t vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t value = PIPEID(pipe) | MEID(mec) | VMID(vmid) | QUEUEID(queue); mutex_lock(&adev->srbm_mutex); WREG32(mmSRBM_GFX_CNTL, value); } -static void unlock_srbm(struct kgd_dev *kgd) +static void unlock_srbm(struct amdgpu_device *adev) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - WREG32(mmSRBM_GFX_CNTL, 0); mutex_unlock(&adev->srbm_mutex); } -static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id, +static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, queue_id, 0); + lock_srbm(adev, mec, pipe, queue_id, 0); } -static void release_queue(struct kgd_dev *kgd) +static void release_queue(struct amdgpu_device *adev) { - unlock_srbm(kgd); + unlock_srbm(adev); } -static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid, +static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid, uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); WREG32(mmSH_MEM_CONFIG, sh_mem_config); WREG32(mmSH_MEM_APE1_BASE, sh_mem_ape1_base); WREG32(mmSH_MEM_APE1_LIMIT, sh_mem_ape1_limit); WREG32(mmSH_MEM_BASES, sh_mem_bases); - unlock_srbm(kgd); + unlock_srbm(adev); } -static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid, +static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, unsigned int vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - /* * We have to assume that there is no outstanding mapping. * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because @@ -165,21 +151,20 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid, return 0; } -static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id) +static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t mec; uint32_t pipe; mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, 0, 0); + lock_srbm(adev, mec, pipe, 0, 0); WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK | CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK); - unlock_srbm(kgd); + unlock_srbm(adev); return 0; } @@ -207,12 +192,11 @@ static inline struct cik_sdma_rlc_registers *get_sdma_mqd(void *mqd) return (struct cik_sdma_rlc_registers *)mqd; } -static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, - uint32_t queue_id, uint32_t __user *wptr, - uint32_t wptr_shift, uint32_t wptr_mask, - struct mm_struct *mm) +static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd, + uint32_t pipe_id, uint32_t queue_id, + uint32_t __user *wptr, uint32_t wptr_shift, + uint32_t wptr_mask, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct cik_mqd *m; uint32_t *mqd_hqd; uint32_t reg, wptr_val, data; @@ -220,7 +204,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, m = get_mqd(mqd); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); /* HQD registers extend from CP_MQD_BASE_ADDR to CP_MQD_CONTROL. */ mqd_hqd = &m->cp_mqd_base_addr_lo; @@ -239,25 +223,24 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, * release srbm_mutex to avoid circular dependency between * srbm_mutex->mm_sem->reservation_ww_class_mutex->srbm_mutex. */ - release_queue(kgd); + release_queue(adev); valid_wptr = read_user_wptr(mm, wptr, wptr_val); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); if (valid_wptr) WREG32(mmCP_HQD_PQ_WPTR, (wptr_val << wptr_shift) & wptr_mask); data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1); WREG32(mmCP_HQD_ACTIVE, data); - release_queue(kgd); + release_queue(adev); return 0; } -static int kgd_hqd_dump(struct kgd_dev *kgd, +static int kgd_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t i = 0, reg; #define HQD_N_REGS (35+4) #define DUMP_REG(addr) do { \ @@ -271,7 +254,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd, if (*dump == NULL) return -ENOMEM; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE0); DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE1); @@ -281,7 +264,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd, for (reg = mmCP_MQD_BASE_ADDR; reg <= mmCP_MQD_CONTROL; reg++) DUMP_REG(reg); - release_queue(kgd); + release_queue(adev); WARN_ON_ONCE(i != HQD_N_REGS); *n_regs = i; @@ -289,10 +272,9 @@ static int kgd_hqd_dump(struct kgd_dev *kgd, return 0; } -static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd, uint32_t __user *wptr, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct cik_sdma_rlc_registers *m; unsigned long end_jiffies; uint32_t sdma_rlc_reg_offset; @@ -345,11 +327,10 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, return 0; } -static int kgd_hqd_sdma_dump(struct kgd_dev *kgd, +static int kgd_hqd_sdma_dump(struct amdgpu_device *adev, uint32_t engine_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t sdma_offset = engine_id * SDMA1_REGISTER_OFFSET + queue_id * KFD_CIK_SDMA_QUEUE_OFFSET; uint32_t i = 0, reg; @@ -372,15 +353,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd, return 0; } -static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address, - uint32_t pipe_id, uint32_t queue_id) +static bool kgd_hqd_is_occupied(struct amdgpu_device *adev, + uint64_t queue_address, uint32_t pipe_id, + uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t act; bool retval = false; uint32_t low, high; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); act = RREG32(mmCP_HQD_ACTIVE); if (act) { low = lower_32_bits(queue_address >> 8); @@ -390,13 +371,12 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address, high == RREG32(mmCP_HQD_PQ_BASE_HI)) retval = true; } - release_queue(kgd); + release_queue(adev); return retval; } -static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) +static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct cik_sdma_rlc_registers *m; uint32_t sdma_rlc_reg_offset; uint32_t sdma_rlc_rb_cntl; @@ -412,12 +392,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) return false; } -static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t temp; enum hqd_dequeue_request_type type; unsigned long flags, end_jiffies; @@ -426,7 +405,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd, if (amdgpu_in_reset(adev)) return -EIO; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); WREG32(mmCP_HQD_PQ_DOORBELL_CONTROL, 0); switch (reset_type) { @@ -504,20 +483,19 @@ loop: break; if (time_after(jiffies, end_jiffies)) { pr_err("cp queue preemption time out\n"); - release_queue(kgd); + release_queue(adev); return -ETIME; } usleep_range(500, 1000); } - release_queue(kgd); + release_queue(adev); return 0; } -static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, unsigned int utimeout) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct cik_sdma_rlc_registers *m; uint32_t sdma_rlc_reg_offset; uint32_t temp; @@ -551,9 +529,8 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, return 0; } -static int kgd_address_watch_disable(struct kgd_dev *kgd) +static int kgd_address_watch_disable(struct amdgpu_device *adev) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); union TCP_WATCH_CNTL_BITS cntl; unsigned int i; @@ -571,13 +548,12 @@ static int kgd_address_watch_disable(struct kgd_dev *kgd) return 0; } -static int kgd_address_watch_execute(struct kgd_dev *kgd, +static int kgd_address_watch_execute(struct amdgpu_device *adev, unsigned int watch_point_id, uint32_t cntl_val, uint32_t addr_hi, uint32_t addr_lo) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); union TCP_WATCH_CNTL_BITS cntl; cntl.u32All = cntl_val; @@ -602,11 +578,10 @@ static int kgd_address_watch_execute(struct kgd_dev *kgd, return 0; } -static int kgd_wave_control_execute(struct kgd_dev *kgd, +static int kgd_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, uint32_t sq_cmd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t data; mutex_lock(&adev->grbm_idx_mutex); @@ -627,18 +602,17 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd, return 0; } -static uint32_t kgd_address_watch_get_offset(struct kgd_dev *kgd, +static uint32_t kgd_address_watch_get_offset(struct amdgpu_device *adev, unsigned int watch_point_id, unsigned int reg_offset) { return watchRegs[watch_point_id * ADDRESS_WATCH_REG_MAX + reg_offset]; } -static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd, +static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev, uint8_t vmid, uint16_t *p_pasid) { uint32_t value; - struct amdgpu_device *adev = (struct amdgpu_device *) kgd; value = RREG32(mmATC_VMID0_PASID_MAPPING + vmid); *p_pasid = value & ATC_VMID0_PASID_MAPPING__PASID_MASK; @@ -646,21 +620,17 @@ static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd, return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK); } -static void set_scratch_backing_va(struct kgd_dev *kgd, +static void set_scratch_backing_va(struct amdgpu_device *adev, uint64_t va, uint32_t vmid) { - struct amdgpu_device *adev = (struct amdgpu_device *) kgd; - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); WREG32(mmSH_HIDDEN_PRIVATE_BASE_VMID, va); - unlock_srbm(kgd); + unlock_srbm(adev); } -static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, - uint64_t page_table_base) +static void set_vm_context_page_table_base(struct amdgpu_device *adev, + uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) { pr_err("trying to set page table base for wrong VMID\n"); return; @@ -676,10 +646,8 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, * @vmid: vmid pointer * read vmid from register (CIK). */ -static uint32_t read_vmid_from_vmfault_reg(struct kgd_dev *kgd) +static uint32_t read_vmid_from_vmfault_reg(struct amdgpu_device *adev) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - uint32_t status = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS); return REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c index 5ce0ce704a21..52832cd69a93 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c @@ -39,68 +39,54 @@ enum hqd_dequeue_request_type { RESET_WAVES }; -static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd) -{ - return (struct amdgpu_device *)kgd; -} - -static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe, +static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe, uint32_t queue, uint32_t vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t value = PIPEID(pipe) | MEID(mec) | VMID(vmid) | QUEUEID(queue); mutex_lock(&adev->srbm_mutex); WREG32(mmSRBM_GFX_CNTL, value); } -static void unlock_srbm(struct kgd_dev *kgd) +static void unlock_srbm(struct amdgpu_device *adev) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - WREG32(mmSRBM_GFX_CNTL, 0); mutex_unlock(&adev->srbm_mutex); } -static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id, +static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, queue_id, 0); + lock_srbm(adev, mec, pipe, queue_id, 0); } -static void release_queue(struct kgd_dev *kgd) +static void release_queue(struct amdgpu_device *adev) { - unlock_srbm(kgd); + unlock_srbm(adev); } -static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid, +static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid, uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); WREG32(mmSH_MEM_CONFIG, sh_mem_config); WREG32(mmSH_MEM_APE1_BASE, sh_mem_ape1_base); WREG32(mmSH_MEM_APE1_LIMIT, sh_mem_ape1_limit); WREG32(mmSH_MEM_BASES, sh_mem_bases); - unlock_srbm(kgd); + unlock_srbm(adev); } -static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid, +static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, unsigned int vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - /* * We have to assume that there is no outstanding mapping. * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because @@ -123,21 +109,20 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid, return 0; } -static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id) +static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t mec; uint32_t pipe; mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, 0, 0); + lock_srbm(adev, mec, pipe, 0, 0); WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK | CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK); - unlock_srbm(kgd); + unlock_srbm(adev); return 0; } @@ -165,12 +150,11 @@ static inline struct vi_sdma_mqd *get_sdma_mqd(void *mqd) return (struct vi_sdma_mqd *)mqd; } -static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, - uint32_t queue_id, uint32_t __user *wptr, - uint32_t wptr_shift, uint32_t wptr_mask, - struct mm_struct *mm) +static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd, + uint32_t pipe_id, uint32_t queue_id, + uint32_t __user *wptr, uint32_t wptr_shift, + uint32_t wptr_mask, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct vi_mqd *m; uint32_t *mqd_hqd; uint32_t reg, wptr_val, data; @@ -178,7 +162,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, m = get_mqd(mqd); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); /* HIQ is set during driver init period with vmid set to 0*/ if (m->cp_hqd_vmid == 0) { @@ -206,7 +190,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, * on ASICs that do not support context-save. * EOP writes/reads can start anywhere in the ring. */ - if (get_amdgpu_device(kgd)->asic_type != CHIP_TONGA) { + if (adev->asic_type != CHIP_TONGA) { WREG32(mmCP_HQD_EOP_RPTR, m->cp_hqd_eop_rptr); WREG32(mmCP_HQD_EOP_WPTR, m->cp_hqd_eop_wptr); WREG32(mmCP_HQD_EOP_WPTR_MEM, m->cp_hqd_eop_wptr_mem); @@ -226,25 +210,24 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, * release srbm_mutex to avoid circular dependency between * srbm_mutex->mm_sem->reservation_ww_class_mutex->srbm_mutex. */ - release_queue(kgd); + release_queue(adev); valid_wptr = read_user_wptr(mm, wptr, wptr_val); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); if (valid_wptr) WREG32(mmCP_HQD_PQ_WPTR, (wptr_val << wptr_shift) & wptr_mask); data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1); WREG32(mmCP_HQD_ACTIVE, data); - release_queue(kgd); + release_queue(adev); return 0; } -static int kgd_hqd_dump(struct kgd_dev *kgd, +static int kgd_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t i = 0, reg; #define HQD_N_REGS (54+4) #define DUMP_REG(addr) do { \ @@ -258,7 +241,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd, if (*dump == NULL) return -ENOMEM; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE0); DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE1); @@ -268,7 +251,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd, for (reg = mmCP_MQD_BASE_ADDR; reg <= mmCP_HQD_EOP_DONES; reg++) DUMP_REG(reg); - release_queue(kgd); + release_queue(adev); WARN_ON_ONCE(i != HQD_N_REGS); *n_regs = i; @@ -276,10 +259,9 @@ static int kgd_hqd_dump(struct kgd_dev *kgd, return 0; } -static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd, uint32_t __user *wptr, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct vi_sdma_mqd *m; unsigned long end_jiffies; uint32_t sdma_rlc_reg_offset; @@ -331,11 +313,10 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, return 0; } -static int kgd_hqd_sdma_dump(struct kgd_dev *kgd, +static int kgd_hqd_sdma_dump(struct amdgpu_device *adev, uint32_t engine_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t sdma_offset = engine_id * SDMA1_REGISTER_OFFSET + queue_id * KFD_VI_SDMA_QUEUE_OFFSET; uint32_t i = 0, reg; @@ -367,15 +348,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd, return 0; } -static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address, - uint32_t pipe_id, uint32_t queue_id) +static bool kgd_hqd_is_occupied(struct amdgpu_device *adev, + uint64_t queue_address, uint32_t pipe_id, + uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t act; bool retval = false; uint32_t low, high; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); act = RREG32(mmCP_HQD_ACTIVE); if (act) { low = lower_32_bits(queue_address >> 8); @@ -385,13 +366,12 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address, high == RREG32(mmCP_HQD_PQ_BASE_HI)) retval = true; } - release_queue(kgd); + release_queue(adev); return retval; } -static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) +static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct vi_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t sdma_rlc_rb_cntl; @@ -407,12 +387,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) return false; } -static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t temp; enum hqd_dequeue_request_type type; unsigned long flags, end_jiffies; @@ -422,7 +401,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd, if (amdgpu_in_reset(adev)) return -EIO; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); if (m->cp_hqd_vmid == 0) WREG32_FIELD(RLC_CP_SCHEDULERS, scheduler1, 0); @@ -502,20 +481,19 @@ loop: break; if (time_after(jiffies, end_jiffies)) { pr_err("cp queue preemption time out.\n"); - release_queue(kgd); + release_queue(adev); return -ETIME; } usleep_range(500, 1000); } - release_queue(kgd); + release_queue(adev); return 0; } -static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, unsigned int utimeout) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct vi_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t temp; @@ -549,11 +527,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, return 0; } -static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd, +static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev, uint8_t vmid, uint16_t *p_pasid) { uint32_t value; - struct amdgpu_device *adev = (struct amdgpu_device *) kgd; value = RREG32(mmATC_VMID0_PASID_MAPPING + vmid); *p_pasid = value & ATC_VMID0_PASID_MAPPING__PASID_MASK; @@ -561,12 +538,12 @@ static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd, return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK); } -static int kgd_address_watch_disable(struct kgd_dev *kgd) +static int kgd_address_watch_disable(struct amdgpu_device *adev) { return 0; } -static int kgd_address_watch_execute(struct kgd_dev *kgd, +static int kgd_address_watch_execute(struct amdgpu_device *adev, unsigned int watch_point_id, uint32_t cntl_val, uint32_t addr_hi, @@ -575,11 +552,10 @@ static int kgd_address_watch_execute(struct kgd_dev *kgd, return 0; } -static int kgd_wave_control_execute(struct kgd_dev *kgd, +static int kgd_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, uint32_t sq_cmd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t data = 0; mutex_lock(&adev->grbm_idx_mutex); @@ -600,28 +576,24 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd, return 0; } -static uint32_t kgd_address_watch_get_offset(struct kgd_dev *kgd, +static uint32_t kgd_address_watch_get_offset(struct amdgpu_device *adev, unsigned int watch_point_id, unsigned int reg_offset) { return 0; } -static void set_scratch_backing_va(struct kgd_dev *kgd, +static void set_scratch_backing_va(struct amdgpu_device *adev, uint64_t va, uint32_t vmid) { - struct amdgpu_device *adev = (struct amdgpu_device *) kgd; - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); WREG32(mmSH_HIDDEN_PRIVATE_BASE_VMID, va); - unlock_srbm(kgd); + unlock_srbm(adev); } -static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, - uint64_t page_table_base) +static void set_vm_context_page_table_base(struct amdgpu_device *adev, + uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) { pr_err("trying to set page table base for wrong VMID\n"); return; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c index bcc1cbeb8799..1abf662a0e91 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c @@ -46,37 +46,26 @@ enum hqd_dequeue_request_type { SAVE_WAVES }; -static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd) -{ - return (struct amdgpu_device *)kgd; -} - -static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe, +static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe, uint32_t queue, uint32_t vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - mutex_lock(&adev->srbm_mutex); soc15_grbm_select(adev, mec, pipe, queue, vmid); } -static void unlock_srbm(struct kgd_dev *kgd) +static void unlock_srbm(struct amdgpu_device *adev) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - soc15_grbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); } -static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id, +static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, queue_id, 0); + lock_srbm(adev, mec, pipe, queue_id, 0); } static uint64_t get_queue_mask(struct amdgpu_device *adev, @@ -88,33 +77,29 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev, return 1ull << bit; } -static void release_queue(struct kgd_dev *kgd) +static void release_queue(struct amdgpu_device *adev) { - unlock_srbm(kgd); + unlock_srbm(adev); } -void kgd_gfx_v9_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid, +void kgd_gfx_v9_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid, uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG), sh_mem_config); WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_BASES), sh_mem_bases); /* APE1 no longer exists on GFX9 */ - unlock_srbm(kgd); + unlock_srbm(adev); } -int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid, +int kgd_gfx_v9_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, unsigned int vmid) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - /* * We have to assume that there is no outstanding mapping. * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because @@ -171,22 +156,21 @@ int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid, * but still works */ -int kgd_gfx_v9_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id) +int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t mec; uint32_t pipe; mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(kgd, mec, pipe, 0, 0); + lock_srbm(adev, mec, pipe, 0, 0); - WREG32(SOC15_REG_OFFSET(GC, 0, mmCPC_INT_CNTL), + WREG32_SOC15(GC, 0, mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK | CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK); - unlock_srbm(kgd); + unlock_srbm(adev); return 0; } @@ -233,19 +217,18 @@ static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd) return (struct v9_sdma_mqd *)mqd; } -int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, - uint32_t queue_id, uint32_t __user *wptr, - uint32_t wptr_shift, uint32_t wptr_mask, - struct mm_struct *mm) +int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd, + uint32_t pipe_id, uint32_t queue_id, + uint32_t __user *wptr, uint32_t wptr_shift, + uint32_t wptr_mask, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_mqd *m; uint32_t *mqd_hqd; uint32_t reg, hqd_base, data; m = get_mqd(mqd); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); /* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */ mqd_hqd = &m->cp_mqd_base_addr_lo; @@ -296,7 +279,7 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, lower_32_bits((uintptr_t)wptr)); WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI), upper_32_bits((uintptr_t)wptr)); - WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_PQ_WPTR_POLL_CNTL1), + WREG32_SOC15(GC, 0, mmCP_PQ_WPTR_POLL_CNTL1, (uint32_t)get_queue_mask(adev, pipe_id, queue_id)); } @@ -308,16 +291,15 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1); WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data); - release_queue(kgd); + release_queue(adev); return 0; } -int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd, +int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t doorbell_off) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; struct v9_mqd *m; uint32_t mec, pipe; @@ -325,7 +307,7 @@ int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd, m = get_mqd(mqd); - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); @@ -361,16 +343,15 @@ int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd, out_unlock: spin_unlock(&adev->gfx.kiq.ring_lock); - release_queue(kgd); + release_queue(adev); return r; } -int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd, +int kgd_gfx_v9_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t i = 0, reg; #define HQD_N_REGS 56 #define DUMP_REG(addr) do { \ @@ -384,13 +365,13 @@ int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd, if (*dump == NULL) return -ENOMEM; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR); reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++) DUMP_REG(reg); - release_queue(kgd); + release_queue(adev); WARN_ON_ONCE(i != HQD_N_REGS); *n_regs = i; @@ -398,10 +379,9 @@ int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd, return 0; } -static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd, uint32_t __user *wptr, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; unsigned long end_jiffies; @@ -468,11 +448,10 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, return 0; } -static int kgd_hqd_sdma_dump(struct kgd_dev *kgd, +static int kgd_hqd_sdma_dump(struct amdgpu_device *adev, uint32_t engine_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, engine_id, queue_id); uint32_t i = 0, reg; @@ -500,31 +479,30 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd, return 0; } -bool kgd_gfx_v9_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address, - uint32_t pipe_id, uint32_t queue_id) +bool kgd_gfx_v9_hqd_is_occupied(struct amdgpu_device *adev, + uint64_t queue_address, uint32_t pipe_id, + uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t act; bool retval = false; uint32_t low, high; - acquire_queue(kgd, pipe_id, queue_id); - act = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE)); + acquire_queue(adev, pipe_id, queue_id); + act = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE); if (act) { low = lower_32_bits(queue_address >> 8); high = upper_32_bits(queue_address >> 8); - if (low == RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_BASE)) && - high == RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_BASE_HI))) + if (low == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE) && + high == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI)) retval = true; } - release_queue(kgd); + release_queue(adev); return retval; } -static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) +static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t sdma_rlc_rb_cntl; @@ -541,12 +519,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) return false; } -int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd, +int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, uint32_t queue_id) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); enum hqd_dequeue_request_type type; unsigned long end_jiffies; uint32_t temp; @@ -555,7 +532,7 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd, if (amdgpu_in_reset(adev)) return -EIO; - acquire_queue(kgd, pipe_id, queue_id); + acquire_queue(adev, pipe_id, queue_id); if (m->cp_hqd_vmid == 0) WREG32_FIELD15_RLC(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0); @@ -579,25 +556,24 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd, end_jiffies = (utimeout * HZ / 1000) + jiffies; while (true) { - temp = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE)); + temp = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE); if (!(temp & CP_HQD_ACTIVE__ACTIVE_MASK)) break; if (time_after(jiffies, end_jiffies)) { pr_err("cp queue preemption time out.\n"); - release_queue(kgd); + release_queue(adev); return -ETIME; } usleep_range(500, 1000); } - release_queue(kgd); + release_queue(adev); return 0; } -static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, +static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, unsigned int utimeout) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t temp; @@ -634,11 +610,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, return 0; } -bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd, +bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev, uint8_t vmid, uint16_t *p_pasid) { uint32_t value; - struct amdgpu_device *adev = (struct amdgpu_device *) kgd; value = RREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING) + vmid); @@ -647,12 +622,12 @@ bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd, return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK); } -int kgd_gfx_v9_address_watch_disable(struct kgd_dev *kgd) +int kgd_gfx_v9_address_watch_disable(struct amdgpu_device *adev) { return 0; } -int kgd_gfx_v9_address_watch_execute(struct kgd_dev *kgd, +int kgd_gfx_v9_address_watch_execute(struct amdgpu_device *adev, unsigned int watch_point_id, uint32_t cntl_val, uint32_t addr_hi, @@ -661,17 +636,16 @@ int kgd_gfx_v9_address_watch_execute(struct kgd_dev *kgd, return 0; } -int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd, +int kgd_gfx_v9_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, uint32_t sq_cmd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t data = 0; mutex_lock(&adev->grbm_idx_mutex); WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, gfx_index_val); - WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CMD), sq_cmd); + WREG32_SOC15(GC, 0, mmSQ_CMD, sq_cmd); data = REG_SET_FIELD(data, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES, 1); @@ -686,18 +660,16 @@ int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd, return 0; } -uint32_t kgd_gfx_v9_address_watch_get_offset(struct kgd_dev *kgd, +uint32_t kgd_gfx_v9_address_watch_get_offset(struct amdgpu_device *adev, unsigned int watch_point_id, unsigned int reg_offset) { return 0; } -void kgd_gfx_v9_set_vm_context_page_table_base(struct kgd_dev *kgd, +void kgd_gfx_v9_set_vm_context_page_table_base(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) { pr_err("trying to set page table base for wrong VMID %u\n", vmid); @@ -750,7 +722,7 @@ static void get_wave_count(struct amdgpu_device *adev, int queue_idx, pipe_idx = queue_idx / adev->gfx.mec.num_queue_per_pipe; queue_slot = queue_idx % adev->gfx.mec.num_queue_per_pipe; soc15_grbm_select(adev, 1, pipe_idx, queue_slot, 0); - reg_val = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_CSQ_WF_ACTIVE_COUNT_0) + + reg_val = RREG32_SOC15_IP(GC, SOC15_REG_OFFSET(GC, 0, mmSPI_CSQ_WF_ACTIVE_COUNT_0) + queue_slot); *wave_cnt = reg_val & SPI_CSQ_WF_ACTIVE_COUNT_0__COUNT_MASK; if (*wave_cnt != 0) @@ -804,7 +776,7 @@ static void get_wave_count(struct amdgpu_device *adev, int queue_idx, * * Reading registers referenced above involves programming GRBM appropriately */ -void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid, +void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid, int *pasid_wave_cnt, int *max_waves_per_cu) { int qidx; @@ -818,10 +790,8 @@ void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid, int pasid_tmp; int max_queue_cnt; int vmid_wave_cnt = 0; - struct amdgpu_device *adev; DECLARE_BITMAP(cp_queue_bitmap, KGD_MAX_QUEUES); - adev = get_amdgpu_device(kgd); lock_spi_csq_mutexes(adev); soc15_grbm_select(adev, 1, 0, 0, 0); @@ -839,8 +809,7 @@ void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid, for (sh_idx = 0; sh_idx < sh_cnt; sh_idx++) { gfx_v9_0_select_se_sh(adev, se_idx, sh_idx, 0xffffffff); - queue_map = RREG32(SOC15_REG_OFFSET(GC, 0, - mmSPI_CSQ_WF_ACTIVE_STATUS)); + queue_map = RREG32_SOC15(GC, 0, mmSPI_CSQ_WF_ACTIVE_STATUS); /* * Assumption: queue map encodes following schema: four @@ -882,30 +851,28 @@ void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid, adev->gfx.cu_info.max_waves_per_simd; } -void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd, +void kgd_gfx_v9_program_trap_handler_settings(struct amdgpu_device *adev, uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); - - lock_srbm(kgd, 0, 0, 0, vmid); + lock_srbm(adev, 0, 0, 0, vmid); /* * Program TBA registers */ - WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TBA_LO), + WREG32_SOC15(GC, 0, mmSQ_SHADER_TBA_LO, lower_32_bits(tba_addr >> 8)); - WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TBA_HI), + WREG32_SOC15(GC, 0, mmSQ_SHADER_TBA_HI, upper_32_bits(tba_addr >> 8)); /* * Program TMA registers */ - WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_LO), + WREG32_SOC15(GC, 0, mmSQ_SHADER_TMA_LO, lower_32_bits(tma_addr >> 8)); - WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI), + WREG32_SOC15(GC, 0, mmSQ_SHADER_TMA_HI, upper_32_bits(tma_addr >> 8)); - unlock_srbm(kgd); + unlock_srbm(adev); } const struct kfd2kgd_calls gfx_v9_kfd2kgd = { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h index c63591106879..24be49df26fd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h @@ -22,48 +22,49 @@ -void kgd_gfx_v9_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid, +void kgd_gfx_v9_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid, uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases); -int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid, +int kgd_gfx_v9_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, unsigned int vmid); -int kgd_gfx_v9_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id); -int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, +int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id); +int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t __user *wptr, uint32_t wptr_shift, uint32_t wptr_mask, struct mm_struct *mm); -int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd, +int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t doorbell_off); -int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd, +int kgd_gfx_v9_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs); -bool kgd_gfx_v9_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address, - uint32_t pipe_id, uint32_t queue_id); -int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd, +bool kgd_gfx_v9_hqd_is_occupied(struct amdgpu_device *adev, + uint64_t queue_address, uint32_t pipe_id, + uint32_t queue_id); +int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, uint32_t queue_id); -int kgd_gfx_v9_address_watch_disable(struct kgd_dev *kgd); -int kgd_gfx_v9_address_watch_execute(struct kgd_dev *kgd, +int kgd_gfx_v9_address_watch_disable(struct amdgpu_device *adev); +int kgd_gfx_v9_address_watch_execute(struct amdgpu_device *adev, unsigned int watch_point_id, uint32_t cntl_val, uint32_t addr_hi, uint32_t addr_lo); -int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd, +int kgd_gfx_v9_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, uint32_t sq_cmd); -uint32_t kgd_gfx_v9_address_watch_get_offset(struct kgd_dev *kgd, +uint32_t kgd_gfx_v9_address_watch_get_offset(struct amdgpu_device *adev, unsigned int watch_point_id, unsigned int reg_offset); -bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd, +bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev, uint8_t vmid, uint16_t *p_pasid); -void kgd_gfx_v9_set_vm_context_page_table_base(struct kgd_dev *kgd, +void kgd_gfx_v9_set_vm_context_page_table_base(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base); -void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid, +void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid, int *pasid_wave_cnt, int *max_waves_per_cu); -void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd, +void kgd_gfx_v9_program_trap_handler_settings(struct amdgpu_device *adev, uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 054c1a224def..f9bab963a948 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -60,12 +60,6 @@ static const char * const domain_bit_to_string[] = { static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work); - -static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd) -{ - return (struct amdgpu_device *)kgd; -} - static bool kfd_mem_is_attached(struct amdgpu_vm *avm, struct kgd_mem *mem) { @@ -126,8 +120,19 @@ static size_t amdgpu_amdkfd_acc_size(uint64_t size) PAGE_ALIGN(size); } +/** + * @amdgpu_amdkfd_reserve_mem_limit() - Decrease available memory by size + * of buffer including any reserved for control structures + * + * @adev: Device to which allocated BO belongs to + * @size: Size of buffer, in bytes, encapsulated by B0. This should be + * equivalent to amdgpu_bo_size(BO) + * @alloc_flag: Flag used in allocating a BO as noted above + * + * Return: returns -ENOMEM in case of error, ZERO otherwise + */ static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, - uint64_t size, u32 domain, bool sg) + uint64_t size, u32 alloc_flag) { uint64_t reserved_for_pt = ESTIMATE_PT_SIZE(amdgpu_amdkfd_total_mem_size); @@ -137,20 +142,24 @@ static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, acc_size = amdgpu_amdkfd_acc_size(size); vram_needed = 0; - if (domain == AMDGPU_GEM_DOMAIN_GTT) { - /* TTM GTT memory */ + if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_GTT) { system_mem_needed = acc_size + size; ttm_mem_needed = acc_size + size; - } else if (domain == AMDGPU_GEM_DOMAIN_CPU && !sg) { - /* Userptr */ + } else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) { + system_mem_needed = acc_size; + ttm_mem_needed = acc_size; + vram_needed = size; + } else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) { system_mem_needed = acc_size + size; ttm_mem_needed = acc_size; - } else { - /* VRAM and SG */ + } else if (alloc_flag & + (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL | + KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) { system_mem_needed = acc_size; ttm_mem_needed = acc_size; - if (domain == AMDGPU_GEM_DOMAIN_VRAM) - vram_needed = size; + } else { + pr_err("%s: Invalid BO type %#x\n", __func__, alloc_flag); + return -ENOMEM; } spin_lock(&kfd_mem_limit.mem_limit_lock); @@ -166,62 +175,72 @@ static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, (adev->kfd.vram_used + vram_needed > adev->gmc.real_vram_size - reserved_for_pt)) { ret = -ENOMEM; - } else { - kfd_mem_limit.system_mem_used += system_mem_needed; - kfd_mem_limit.ttm_mem_used += ttm_mem_needed; - adev->kfd.vram_used += vram_needed; + goto release; } + /* Update memory accounting by decreasing available system + * memory, TTM memory and GPU memory as computed above + */ + adev->kfd.vram_used += vram_needed; + kfd_mem_limit.system_mem_used += system_mem_needed; + kfd_mem_limit.ttm_mem_used += ttm_mem_needed; + +release: spin_unlock(&kfd_mem_limit.mem_limit_lock); return ret; } static void unreserve_mem_limit(struct amdgpu_device *adev, - uint64_t size, u32 domain, bool sg) + uint64_t size, u32 alloc_flag) { size_t acc_size; acc_size = amdgpu_amdkfd_acc_size(size); spin_lock(&kfd_mem_limit.mem_limit_lock); - if (domain == AMDGPU_GEM_DOMAIN_GTT) { + + if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_GTT) { kfd_mem_limit.system_mem_used -= (acc_size + size); kfd_mem_limit.ttm_mem_used -= (acc_size + size); - } else if (domain == AMDGPU_GEM_DOMAIN_CPU && !sg) { + } else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) { + kfd_mem_limit.system_mem_used -= acc_size; + kfd_mem_limit.ttm_mem_used -= acc_size; + adev->kfd.vram_used -= size; + } else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) { kfd_mem_limit.system_mem_used -= (acc_size + size); kfd_mem_limit.ttm_mem_used -= acc_size; - } else { + } else if (alloc_flag & + (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL | + KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) { kfd_mem_limit.system_mem_used -= acc_size; kfd_mem_limit.ttm_mem_used -= acc_size; - if (domain == AMDGPU_GEM_DOMAIN_VRAM) { - adev->kfd.vram_used -= size; - WARN_ONCE(adev->kfd.vram_used < 0, - "kfd VRAM memory accounting unbalanced"); - } + } else { + pr_err("%s: Invalid BO type %#x\n", __func__, alloc_flag); + goto release; } - WARN_ONCE(kfd_mem_limit.system_mem_used < 0, - "kfd system memory accounting unbalanced"); + + WARN_ONCE(adev->kfd.vram_used < 0, + "KFD VRAM memory accounting unbalanced"); WARN_ONCE(kfd_mem_limit.ttm_mem_used < 0, - "kfd TTM memory accounting unbalanced"); + "KFD TTM memory accounting unbalanced"); + WARN_ONCE(kfd_mem_limit.system_mem_used < 0, + "KFD system memory accounting unbalanced"); +release: spin_unlock(&kfd_mem_limit.mem_limit_lock); } -void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo) +void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); - u32 domain = bo->preferred_domains; - bool sg = (bo->preferred_domains == AMDGPU_GEM_DOMAIN_CPU); + u32 alloc_flags = bo->kfd_bo->alloc_flags; + u64 size = amdgpu_bo_size(bo); - if (bo->flags & AMDGPU_AMDKFD_CREATE_USERPTR_BO) { - domain = AMDGPU_GEM_DOMAIN_CPU; - sg = false; - } + unreserve_mem_limit(adev, size, alloc_flags); - unreserve_mem_limit(adev, amdgpu_bo_size(bo), domain, sg); + kfree(bo->kfd_bo); } - /* amdgpu_amdkfd_remove_eviction_fence - Removes eviction fence from BO's * reservation object. * @@ -644,12 +663,6 @@ kfd_mem_attach_dmabuf(struct amdgpu_device *adev, struct kgd_mem *mem, if (IS_ERR(gobj)) return PTR_ERR(gobj); - /* Import takes an extra reference on the dmabuf. Drop it now to - * avoid leaking it. We only need the one reference in - * kgd_mem->dmabuf. - */ - dma_buf_put(mem->dmabuf); - *bo = gem_to_amdgpu_bo(gobj); (*bo)->flags |= AMDGPU_GEM_CREATE_PREEMPTIBLE; (*bo)->parent = amdgpu_bo_ref(mem->bo); @@ -695,10 +708,12 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, pr_debug("\t add VA 0x%llx - 0x%llx to vm %p\n", va, va + bo_size, vm); - if (adev == bo_adev || (mem->domain == AMDGPU_GEM_DOMAIN_VRAM && - amdgpu_xgmi_same_hive(adev, bo_adev))) { - /* Mappings on the local GPU and VRAM mappings in the - * local hive share the original BO + if (adev == bo_adev || + (amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm) && adev->ram_is_direct_mapped) || + (mem->domain == AMDGPU_GEM_DOMAIN_VRAM && amdgpu_xgmi_same_hive(adev, bo_adev))) { + /* Mappings on the local GPU, or VRAM mappings in the + * local hive, or userptr mapping IOMMU direct map mode + * share the original BO */ attachment[i]->type = KFD_MEM_ATT_SHARED; bo[i] = mem->bo; @@ -734,14 +749,19 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, } /* Add BO to VM internal data structures */ + ret = amdgpu_bo_reserve(bo[i], false); + if (ret) { + pr_debug("Unable to reserve BO during memory attach"); + goto unwind; + } attachment[i]->bo_va = amdgpu_vm_bo_add(adev, vm, bo[i]); + amdgpu_bo_unreserve(bo[i]); if (unlikely(!attachment[i]->bo_va)) { ret = -ENOMEM; pr_err("Failed to add BO object to VM. ret == %d\n", ret); goto unwind; } - attachment[i]->va = va; attachment[i]->pte_flags = get_pte_flags(adev, mem); attachment[i]->adev = adev; @@ -757,7 +777,9 @@ unwind: if (!attachment[i]) continue; if (attachment[i]->bo_va) { + amdgpu_bo_reserve(bo[i], true); amdgpu_vm_bo_rmv(adev, attachment[i]->bo_va); + amdgpu_bo_unreserve(bo[i]); list_del(&attachment[i]->list); } if (bo[i]) @@ -1269,12 +1291,60 @@ create_evict_fence_fail: return ret; } -int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct kgd_dev *kgd, +/** + * amdgpu_amdkfd_gpuvm_pin_bo() - Pins a BO using following criteria + * @bo: Handle of buffer object being pinned + * @domain: Domain into which BO should be pinned + * + * - USERPTR BOs are UNPINNABLE and will return error + * - All other BO types (GTT, VRAM, MMIO and DOORBELL) will have their + * PIN count incremented. It is valid to PIN a BO multiple times + * + * Return: ZERO if successful in pinning, Non-Zero in case of error. + */ +static int amdgpu_amdkfd_gpuvm_pin_bo(struct amdgpu_bo *bo, u32 domain) +{ + int ret = 0; + + ret = amdgpu_bo_reserve(bo, false); + if (unlikely(ret)) + return ret; + + ret = amdgpu_bo_pin_restricted(bo, domain, 0, 0); + if (ret) + pr_err("Error in Pinning BO to domain: %d\n", domain); + + amdgpu_bo_sync_wait(bo, AMDGPU_FENCE_OWNER_KFD, false); + amdgpu_bo_unreserve(bo); + + return ret; +} + +/** + * amdgpu_amdkfd_gpuvm_unpin_bo() - Unpins BO using following criteria + * @bo: Handle of buffer object being unpinned + * + * - Is a illegal request for USERPTR BOs and is ignored + * - All other BO types (GTT, VRAM, MMIO and DOORBELL) will have their + * PIN count decremented. Calls to UNPIN must balance calls to PIN + */ +static void amdgpu_amdkfd_gpuvm_unpin_bo(struct amdgpu_bo *bo) +{ + int ret = 0; + + ret = amdgpu_bo_reserve(bo, false); + if (unlikely(ret)) + return; + + amdgpu_bo_unpin(bo); + amdgpu_bo_unreserve(bo); +} + +int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev, struct file *filp, u32 pasid, void **process_info, struct dma_fence **ef) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_fpriv *drv_priv; struct amdgpu_vm *avm; int ret; @@ -1350,12 +1420,12 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev, } } -void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *drm_priv) +void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev, + void *drm_priv) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_vm *avm; - if (WARN_ON(!kgd || !drm_priv)) + if (WARN_ON(!adev || !drm_priv)) return; avm = drm_priv_to_vm(drm_priv); @@ -1383,17 +1453,16 @@ uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv) } int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( - struct kgd_dev *kgd, uint64_t va, uint64_t size, + struct amdgpu_device *adev, uint64_t va, uint64_t size, void *drm_priv, struct kgd_mem **mem, uint64_t *offset, uint32_t flags) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv); enum ttm_bo_type bo_type = ttm_bo_type_device; struct sg_table *sg = NULL; uint64_t user_addr = 0; struct amdgpu_bo *bo; - struct drm_gem_object *gobj; + struct drm_gem_object *gobj = NULL; u32 domain, alloc_domain; u64 alloc_flags; int ret; @@ -1451,7 +1520,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( amdgpu_sync_create(&(*mem)->sync); - ret = amdgpu_amdkfd_reserve_mem_limit(adev, size, alloc_domain, !!sg); + ret = amdgpu_amdkfd_reserve_mem_limit(adev, size, flags); if (ret) { pr_debug("Insufficient memory\n"); goto err_reserve_limit; @@ -1492,6 +1561,15 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( ret = init_user_pages(*mem, user_addr); if (ret) goto allocate_init_user_pages_failed; + } else if (flags & (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL | + KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) { + ret = amdgpu_amdkfd_gpuvm_pin_bo(bo, AMDGPU_GEM_DOMAIN_GTT); + if (ret) { + pr_err("Pinning MMIO/DOORBELL BO during ALLOC FAILED\n"); + goto err_pin_bo; + } + bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT; + bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT; } if (offset) @@ -1500,17 +1578,20 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( return 0; allocate_init_user_pages_failed: +err_pin_bo: remove_kgd_mem_from_kfd_bo_list(*mem, avm->process_info); drm_vma_node_revoke(&gobj->vma_node, drm_priv); err_node_allow: - amdgpu_bo_unref(&bo); /* Don't unreserve system mem limit twice */ goto err_reserve_limit; err_bo_create: - unreserve_mem_limit(adev, size, alloc_domain, !!sg); + unreserve_mem_limit(adev, size, flags); err_reserve_limit: mutex_destroy(&(*mem)->lock); - kfree(*mem); + if (gobj) + drm_gem_object_put(gobj); + else + kfree(*mem); err: if (sg) { sg_free_table(sg); @@ -1520,7 +1601,7 @@ err: } int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( - struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv, + struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv, uint64_t *size) { struct amdkfd_process_info *process_info = mem->process_info; @@ -1533,6 +1614,14 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( bool is_imported = false; mutex_lock(&mem->lock); + + /* Unpin MMIO/DOORBELL BO's that were pinnned during allocation */ + if (mem->alloc_flags & + (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL | + KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) { + amdgpu_amdkfd_gpuvm_unpin_bo(mem->bo); + } + mapped_to_gpu_memory = mem->mapped_to_gpu_memory; is_imported = mem->is_imported; mutex_unlock(&mem->lock); @@ -1568,12 +1657,12 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( pr_debug("Release VA 0x%llx - 0x%llx\n", mem->va, mem->va + bo_size * (1 + mem->aql_queue)); - ret = unreserve_bo_and_vms(&ctx, false, false); - /* Remove from VM internal data structures */ list_for_each_entry_safe(entry, tmp, &mem->attachments, list) kfd_mem_detach(entry); + ret = unreserve_bo_and_vms(&ctx, false, false); + /* Free the sync object */ amdgpu_sync_free(&mem->sync); @@ -1600,18 +1689,21 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( drm_vma_node_revoke(&mem->bo->tbo.base.vma_node, drm_priv); if (mem->dmabuf) dma_buf_put(mem->dmabuf); - drm_gem_object_put(&mem->bo->tbo.base); mutex_destroy(&mem->lock); - kfree(mem); + + /* If this releases the last reference, it will end up calling + * amdgpu_amdkfd_release_notify and kfree the mem struct. That's why + * this needs to be the last call here. + */ + drm_gem_object_put(&mem->bo->tbo.base); return ret; } int amdgpu_amdkfd_gpuvm_map_memory_to_gpu( - struct kgd_dev *kgd, struct kgd_mem *mem, + struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv, bool *table_freed) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv); int ret; struct amdgpu_bo *bo; @@ -1738,7 +1830,7 @@ out: } int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu( - struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv) + struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv) { struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv); struct amdkfd_process_info *process_info = avm->process_info; @@ -1799,7 +1891,7 @@ out: } int amdgpu_amdkfd_gpuvm_sync_memory( - struct kgd_dev *kgd, struct kgd_mem *mem, bool intr) + struct amdgpu_device *adev, struct kgd_mem *mem, bool intr) { struct amdgpu_sync sync; int ret; @@ -1815,7 +1907,7 @@ int amdgpu_amdkfd_gpuvm_sync_memory( return ret; } -int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd, +int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct amdgpu_device *adev, struct kgd_mem *mem, void **kptr, uint64_t *size) { int ret; @@ -1871,12 +1963,20 @@ bo_reserve_failed: return ret; } -int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd, - struct kfd_vm_fault_info *mem) +void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct amdgpu_device *adev, + struct kgd_mem *mem) { - struct amdgpu_device *adev; + struct amdgpu_bo *bo = mem->bo; - adev = (struct amdgpu_device *)kgd; + amdgpu_bo_reserve(bo, true); + amdgpu_bo_kunmap(bo); + amdgpu_bo_unpin(bo); + amdgpu_bo_unreserve(bo); +} + +int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct amdgpu_device *adev, + struct kfd_vm_fault_info *mem) +{ if (atomic_read(&adev->gmc.vm_fault_info_updated) == 1) { *mem = *adev->gmc.vm_fault_info; mb(); @@ -1885,13 +1985,12 @@ int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd, return 0; } -int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd, +int amdgpu_amdkfd_gpuvm_import_dmabuf(struct amdgpu_device *adev, struct dma_buf *dma_buf, uint64_t va, void *drm_priv, struct kgd_mem **mem, uint64_t *size, uint64_t *mmap_offset) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv); struct drm_gem_object *obj; struct amdgpu_bo *bo; @@ -2041,19 +2140,26 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info, /* Get updated user pages */ ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages); if (ret) { - pr_debug("%s: Failed to get user pages: %d\n", - __func__, ret); + pr_debug("Failed %d to get user pages\n", ret); + + /* Return -EFAULT bad address error as success. It will + * fail later with a VM fault if the GPU tries to access + * it. Better than hanging indefinitely with stalled + * user mode queues. + * + * Return other error -EBUSY or -ENOMEM to retry restore + */ + if (ret != -EFAULT) + return ret; + } else { - /* Return error -EBUSY or -ENOMEM, retry restore */ - return ret; + /* + * FIXME: Cannot ignore the return code, must hold + * notifier_lock + */ + amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm); } - /* - * FIXME: Cannot ignore the return code, must hold - * notifier_lock - */ - amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm); - /* Mark the BO as valid unless it was invalidated * again concurrently. */ @@ -2511,11 +2617,9 @@ int amdgpu_amdkfd_remove_gws_from_process(void *info, void *mem) } /* Returns GPU-specific tiling mode information */ -int amdgpu_amdkfd_get_tile_config(struct kgd_dev *kgd, +int amdgpu_amdkfd_get_tile_config(struct amdgpu_device *adev, struct tile_config *config) { - struct amdgpu_device *adev = (struct amdgpu_device *)kgd; - config->gb_addr_config = adev->gfx.config.gb_addr_config; config->tile_config_ptr = adev->gfx.config.tile_mode_array; config->num_tile_configs = diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c index 96b7bb13a2dd..12a6b1c99c93 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c @@ -1569,6 +1569,18 @@ void amdgpu_atombios_scratch_regs_engine_hung(struct amdgpu_device *adev, WREG32(adev->bios_scratch_reg_offset + 3, tmp); } +void amdgpu_atombios_scratch_regs_set_backlight_level(struct amdgpu_device *adev, + u32 backlight_level) +{ + u32 tmp = RREG32(adev->bios_scratch_reg_offset + 2); + + tmp &= ~ATOM_S2_CURRENT_BL_LEVEL_MASK; + tmp |= (backlight_level << ATOM_S2_CURRENT_BL_LEVEL_SHIFT) & + ATOM_S2_CURRENT_BL_LEVEL_MASK; + + WREG32(adev->bios_scratch_reg_offset + 2, tmp); +} + bool amdgpu_atombios_scratch_need_asic_init(struct amdgpu_device *adev) { u32 tmp = RREG32(adev->bios_scratch_reg_offset + 7); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h index 8cc0222dba19..27e74b1fc260 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h @@ -185,6 +185,8 @@ bool amdgpu_atombios_has_gpu_virtualization_table(struct amdgpu_device *adev); void amdgpu_atombios_scratch_regs_lock(struct amdgpu_device *adev, bool lock); void amdgpu_atombios_scratch_regs_engine_hung(struct amdgpu_device *adev, bool hung); +void amdgpu_atombios_scratch_regs_set_backlight_level(struct amdgpu_device *adev, + u32 backlight_level); bool amdgpu_atombios_scratch_need_asic_init(struct amdgpu_device *adev); void amdgpu_atombios_copy_swap(u8 *dst, u8 *src, u8 num_bytes, bool to_le); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c index 97178b307ed6..4d4ddf026faf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c @@ -470,8 +470,8 @@ bool amdgpu_atomfirmware_dynamic_boot_config_supported(struct amdgpu_device *ade /** * amdgpu_atomfirmware_ras_rom_addr -- Get the RAS EEPROM addr from VBIOS - * adev: amdgpu_device pointer - * i2c_address: pointer to u8; if not NULL, will contain + * @adev: amdgpu_device pointer + * @i2c_address: pointer to u8; if not NULL, will contain * the RAS EEPROM address if the function returns true * * Return true if VBIOS supports RAS EEPROM address reporting, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c index 7abe9500c0c6..d6d986be906a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c @@ -11,6 +11,7 @@ #include <linux/pci.h> #include <linux/delay.h> +#include "amdgpu.h" #include "amd_acpi.h" #define AMDGPU_PX_QUIRK_FORCE_ATPX (1 << 0) @@ -165,7 +166,7 @@ static void amdgpu_atpx_parse_functions(struct amdgpu_atpx_functions *f, u32 mas } /** - * amdgpu_atpx_validate_functions - validate ATPX functions + * amdgpu_atpx_validate - validate ATPX functions * * @atpx: amdgpu atpx struct * diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c index 15c45b2a3983..714178f1b6c6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c @@ -61,7 +61,7 @@ static void amdgpu_bo_list_free(struct kref *ref) int amdgpu_bo_list_create(struct amdgpu_device *adev, struct drm_file *filp, struct drm_amdgpu_bo_list_entry *info, - unsigned num_entries, struct amdgpu_bo_list **result) + size_t num_entries, struct amdgpu_bo_list **result) { unsigned last_entry = 0, first_userptr = num_entries; struct amdgpu_bo_list_entry *array; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h index c905a4cfc173..044b41f0bfd9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h @@ -61,7 +61,7 @@ int amdgpu_bo_create_list_entry_array(struct drm_amdgpu_bo_list_in *in, int amdgpu_bo_list_create(struct amdgpu_device *adev, struct drm_file *filp, struct drm_amdgpu_bo_list_entry *info, - unsigned num_entries, + size_t num_entries, struct amdgpu_bo_list **list); static inline struct amdgpu_bo_list_entry * diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c index b9c11c2b2885..c16a2704ced6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c @@ -108,7 +108,7 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector) case DRM_MODE_CONNECTOR_DVII: case DRM_MODE_CONNECTOR_HDMIB: if (amdgpu_connector->use_digital) { - if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) { + if (connector->display_info.is_hdmi) { if (connector->display_info.bpc) bpc = connector->display_info.bpc; } @@ -116,7 +116,7 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector) break; case DRM_MODE_CONNECTOR_DVID: case DRM_MODE_CONNECTOR_HDMIA: - if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) { + if (connector->display_info.is_hdmi) { if (connector->display_info.bpc) bpc = connector->display_info.bpc; } @@ -125,7 +125,7 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector) dig_connector = amdgpu_connector->con_priv; if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) || (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) || - drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) { + connector->display_info.is_hdmi) { if (connector->display_info.bpc) bpc = connector->display_info.bpc; } @@ -149,7 +149,7 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector) break; } - if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) { + if (connector->display_info.is_hdmi) { /* * Pre DCE-8 hw can't handle > 12 bpc, and more than 12 bpc doesn't make * much sense without support for > 12 bpc framebuffers. RGB 4:4:4 at @@ -315,8 +315,10 @@ static void amdgpu_connector_get_edid(struct drm_connector *connector) if (!amdgpu_connector->edid) { /* some laptops provide a hardcoded edid in rom for LCDs */ if (((connector->connector_type == DRM_MODE_CONNECTOR_LVDS) || - (connector->connector_type == DRM_MODE_CONNECTOR_eDP))) + (connector->connector_type == DRM_MODE_CONNECTOR_eDP))) { amdgpu_connector->edid = amdgpu_connector_get_hardcoded_edid(adev); + drm_connector_update_edid_property(connector, amdgpu_connector->edid); + } } } @@ -326,6 +328,7 @@ static void amdgpu_connector_free_edid(struct drm_connector *connector) kfree(amdgpu_connector->edid); amdgpu_connector->edid = NULL; + drm_connector_update_edid_property(connector, NULL); } static int amdgpu_connector_ddc_get_modes(struct drm_connector *connector) @@ -387,6 +390,9 @@ amdgpu_connector_lcd_native_mode(struct drm_encoder *encoder) native_mode->vdisplay != 0 && native_mode->clock != 0) { mode = drm_mode_duplicate(dev, native_mode); + if (!mode) + return NULL; + mode->type = DRM_MODE_TYPE_PREFERRED | DRM_MODE_TYPE_DRIVER; drm_mode_set_name(mode); @@ -401,6 +407,9 @@ amdgpu_connector_lcd_native_mode(struct drm_encoder *encoder) * simpler. */ mode = drm_cvt_mode(dev, native_mode->hdisplay, native_mode->vdisplay, 60, true, false, false); + if (!mode) + return NULL; + mode->type = DRM_MODE_TYPE_PREFERRED | DRM_MODE_TYPE_DRIVER; DRM_DEBUG_KMS("Adding cvt approximation of native panel mode %s\n", mode->name); } @@ -827,6 +836,7 @@ static int amdgpu_connector_vga_get_modes(struct drm_connector *connector) amdgpu_connector_get_edid(connector); ret = amdgpu_connector_ddc_get_modes(connector); + amdgpu_get_native_mode(connector); return ret; } @@ -1170,7 +1180,7 @@ static enum drm_mode_status amdgpu_connector_dvi_mode_valid(struct drm_connector (amdgpu_connector->connector_object_id == CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D) || (amdgpu_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_B)) { return MODE_OK; - } else if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) { + } else if (connector->display_info.is_hdmi) { /* HDMI 1.3+ supports max clock of 340 Mhz */ if (mode->clock > 340000) return MODE_CLOCK_HIGH; @@ -1462,7 +1472,7 @@ static enum drm_mode_status amdgpu_connector_dp_mode_valid(struct drm_connector (amdgpu_dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) { return amdgpu_atombios_dp_mode_valid_helper(connector, mode); } else { - if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) { + if (connector->display_info.is_hdmi) { /* HDMI 1.3+ supports max clock of 340 Mhz */ if (mode->clock > 340000) return MODE_CLOCK_HIGH; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 913f9eaa9cd6..0311d799a010 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -1222,6 +1222,8 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p, if (r) goto error_unlock; + drm_sched_job_arm(&job->base); + /* No memory allocation is allowed while holding the notifier lock. * The lock is held until amdgpu_cs_submit is finished and fence is * added to BOs. @@ -1259,7 +1261,7 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p, trace_amdgpu_cs_ioctl(job); amdgpu_vm_bo_trace_cs(&fpriv->vm, &p->ticket); - drm_sched_entity_push_job(&job->base, entity); + drm_sched_entity_push_job(&job->base); amdgpu_vm_move_to_lru_tail(p->adev, &fpriv->vm); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c index e7a010b7ca1f..468003583b2a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c @@ -43,14 +43,61 @@ const unsigned int amdgpu_ctx_num_entities[AMDGPU_HW_IP_NUM] = { [AMDGPU_HW_IP_VCN_JPEG] = 1, }; +bool amdgpu_ctx_priority_is_valid(int32_t ctx_prio) +{ + switch (ctx_prio) { + case AMDGPU_CTX_PRIORITY_UNSET: + case AMDGPU_CTX_PRIORITY_VERY_LOW: + case AMDGPU_CTX_PRIORITY_LOW: + case AMDGPU_CTX_PRIORITY_NORMAL: + case AMDGPU_CTX_PRIORITY_HIGH: + case AMDGPU_CTX_PRIORITY_VERY_HIGH: + return true; + default: + return false; + } +} + +static enum drm_sched_priority +amdgpu_ctx_to_drm_sched_prio(int32_t ctx_prio) +{ + switch (ctx_prio) { + case AMDGPU_CTX_PRIORITY_UNSET: + return DRM_SCHED_PRIORITY_UNSET; + + case AMDGPU_CTX_PRIORITY_VERY_LOW: + return DRM_SCHED_PRIORITY_MIN; + + case AMDGPU_CTX_PRIORITY_LOW: + return DRM_SCHED_PRIORITY_MIN; + + case AMDGPU_CTX_PRIORITY_NORMAL: + return DRM_SCHED_PRIORITY_NORMAL; + + case AMDGPU_CTX_PRIORITY_HIGH: + return DRM_SCHED_PRIORITY_HIGH; + + case AMDGPU_CTX_PRIORITY_VERY_HIGH: + return DRM_SCHED_PRIORITY_HIGH; + + /* This should not happen as we sanitized userspace provided priority + * already, WARN if this happens. + */ + default: + WARN(1, "Invalid context priority %d\n", ctx_prio); + return DRM_SCHED_PRIORITY_NORMAL; + } + +} + static int amdgpu_ctx_priority_permit(struct drm_file *filp, - enum drm_sched_priority priority) + int32_t priority) { - if (priority < 0 || priority >= DRM_SCHED_PRIORITY_COUNT) + if (!amdgpu_ctx_priority_is_valid(priority)) return -EINVAL; /* NORMAL and below are accessible by everyone */ - if (priority <= DRM_SCHED_PRIORITY_NORMAL) + if (priority <= AMDGPU_CTX_PRIORITY_NORMAL) return 0; if (capable(CAP_SYS_NICE)) @@ -62,26 +109,51 @@ static int amdgpu_ctx_priority_permit(struct drm_file *filp, return -EACCES; } -static enum gfx_pipe_priority amdgpu_ctx_sched_prio_to_compute_prio(enum drm_sched_priority prio) +static enum amdgpu_gfx_pipe_priority amdgpu_ctx_prio_to_compute_prio(int32_t prio) { switch (prio) { - case DRM_SCHED_PRIORITY_HIGH: - case DRM_SCHED_PRIORITY_KERNEL: + case AMDGPU_CTX_PRIORITY_HIGH: + case AMDGPU_CTX_PRIORITY_VERY_HIGH: return AMDGPU_GFX_PIPE_PRIO_HIGH; default: return AMDGPU_GFX_PIPE_PRIO_NORMAL; } } -static unsigned int amdgpu_ctx_prio_sched_to_hw(struct amdgpu_device *adev, - enum drm_sched_priority prio, - u32 hw_ip) +static enum amdgpu_ring_priority_level amdgpu_ctx_sched_prio_to_ring_prio(int32_t prio) { + switch (prio) { + case AMDGPU_CTX_PRIORITY_HIGH: + return AMDGPU_RING_PRIO_1; + case AMDGPU_CTX_PRIORITY_VERY_HIGH: + return AMDGPU_RING_PRIO_2; + default: + return AMDGPU_RING_PRIO_0; + } +} + +static unsigned int amdgpu_ctx_get_hw_prio(struct amdgpu_ctx *ctx, u32 hw_ip) +{ + struct amdgpu_device *adev = ctx->adev; + int32_t ctx_prio; unsigned int hw_prio; - hw_prio = (hw_ip == AMDGPU_HW_IP_COMPUTE) ? - amdgpu_ctx_sched_prio_to_compute_prio(prio) : - AMDGPU_RING_PRIO_DEFAULT; + ctx_prio = (ctx->override_priority == AMDGPU_CTX_PRIORITY_UNSET) ? + ctx->init_priority : ctx->override_priority; + + switch (hw_ip) { + case AMDGPU_HW_IP_COMPUTE: + hw_prio = amdgpu_ctx_prio_to_compute_prio(ctx_prio); + break; + case AMDGPU_HW_IP_VCE: + case AMDGPU_HW_IP_VCN_ENC: + hw_prio = amdgpu_ctx_sched_prio_to_ring_prio(ctx_prio); + break; + default: + hw_prio = AMDGPU_RING_PRIO_DEFAULT; + break; + } + hw_ip = array_index_nospec(hw_ip, AMDGPU_HW_IP_NUM); if (adev->gpu_sched[hw_ip][hw_prio].num_scheds == 0) hw_prio = AMDGPU_RING_PRIO_DEFAULT; @@ -89,15 +161,17 @@ static unsigned int amdgpu_ctx_prio_sched_to_hw(struct amdgpu_device *adev, return hw_prio; } + static int amdgpu_ctx_init_entity(struct amdgpu_ctx *ctx, u32 hw_ip, - const u32 ring) + const u32 ring) { struct amdgpu_device *adev = ctx->adev; struct amdgpu_ctx_entity *entity; struct drm_gpu_scheduler **scheds = NULL, *sched = NULL; unsigned num_scheds = 0; + int32_t ctx_prio; unsigned int hw_prio; - enum drm_sched_priority priority; + enum drm_sched_priority drm_prio; int r; entity = kzalloc(struct_size(entity, fences, amdgpu_sched_jobs), @@ -105,10 +179,11 @@ static int amdgpu_ctx_init_entity(struct amdgpu_ctx *ctx, u32 hw_ip, if (!entity) return -ENOMEM; + ctx_prio = (ctx->override_priority == AMDGPU_CTX_PRIORITY_UNSET) ? + ctx->init_priority : ctx->override_priority; entity->sequence = 1; - priority = (ctx->override_priority == DRM_SCHED_PRIORITY_UNSET) ? - ctx->init_priority : ctx->override_priority; - hw_prio = amdgpu_ctx_prio_sched_to_hw(adev, priority, hw_ip); + hw_prio = amdgpu_ctx_get_hw_prio(ctx, hw_ip); + drm_prio = amdgpu_ctx_to_drm_sched_prio(ctx_prio); hw_ip = array_index_nospec(hw_ip, AMDGPU_HW_IP_NUM); scheds = adev->gpu_sched[hw_ip][hw_prio].sched; @@ -124,7 +199,7 @@ static int amdgpu_ctx_init_entity(struct amdgpu_ctx *ctx, u32 hw_ip, num_scheds = 1; } - r = drm_sched_entity_init(&entity->entity, priority, scheds, num_scheds, + r = drm_sched_entity_init(&entity->entity, drm_prio, scheds, num_scheds, &ctx->guilty); if (r) goto error_free_entity; @@ -139,7 +214,7 @@ error_free_entity: } static int amdgpu_ctx_init(struct amdgpu_device *adev, - enum drm_sched_priority priority, + int32_t priority, struct drm_file *filp, struct amdgpu_ctx *ctx) { @@ -161,7 +236,7 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev, ctx->reset_counter_query = ctx->reset_counter; ctx->vram_lost_counter = atomic_read(&adev->vram_lost_counter); ctx->init_priority = priority; - ctx->override_priority = DRM_SCHED_PRIORITY_UNSET; + ctx->override_priority = AMDGPU_CTX_PRIORITY_UNSET; return 0; } @@ -234,7 +309,7 @@ int amdgpu_ctx_get_entity(struct amdgpu_ctx *ctx, u32 hw_ip, u32 instance, static int amdgpu_ctx_alloc(struct amdgpu_device *adev, struct amdgpu_fpriv *fpriv, struct drm_file *filp, - enum drm_sched_priority priority, + int32_t priority, uint32_t *id) { struct amdgpu_ctx_mgr *mgr = &fpriv->ctx_mgr; @@ -397,19 +472,19 @@ int amdgpu_ctx_ioctl(struct drm_device *dev, void *data, { int r; uint32_t id; - enum drm_sched_priority priority; + int32_t priority; union drm_amdgpu_ctx *args = data; struct amdgpu_device *adev = drm_to_adev(dev); struct amdgpu_fpriv *fpriv = filp->driver_priv; id = args->in.ctx_id; - r = amdgpu_to_sched_priority(args->in.priority, &priority); + priority = args->in.priority; /* For backwards compatibility reasons, we need to accept * ioctls with garbage in the priority field */ - if (r == -EINVAL) - priority = DRM_SCHED_PRIORITY_NORMAL; + if (!amdgpu_ctx_priority_is_valid(priority)) + priority = AMDGPU_CTX_PRIORITY_NORMAL; switch (args->in.op) { case AMDGPU_CTX_OP_ALLOC_CTX: @@ -515,9 +590,9 @@ struct dma_fence *amdgpu_ctx_get_fence(struct amdgpu_ctx *ctx, } static void amdgpu_ctx_set_entity_priority(struct amdgpu_ctx *ctx, - struct amdgpu_ctx_entity *aentity, - int hw_ip, - enum drm_sched_priority priority) + struct amdgpu_ctx_entity *aentity, + int hw_ip, + int32_t priority) { struct amdgpu_device *adev = ctx->adev; unsigned int hw_prio; @@ -525,12 +600,12 @@ static void amdgpu_ctx_set_entity_priority(struct amdgpu_ctx *ctx, unsigned num_scheds; /* set sw priority */ - drm_sched_entity_set_priority(&aentity->entity, priority); + drm_sched_entity_set_priority(&aentity->entity, + amdgpu_ctx_to_drm_sched_prio(priority)); /* set hw priority */ if (hw_ip == AMDGPU_HW_IP_COMPUTE) { - hw_prio = amdgpu_ctx_prio_sched_to_hw(adev, priority, - AMDGPU_HW_IP_COMPUTE); + hw_prio = amdgpu_ctx_get_hw_prio(ctx, hw_ip); hw_prio = array_index_nospec(hw_prio, AMDGPU_RING_PRIO_MAX); scheds = adev->gpu_sched[hw_ip][hw_prio].sched; num_scheds = adev->gpu_sched[hw_ip][hw_prio].num_scheds; @@ -540,14 +615,14 @@ static void amdgpu_ctx_set_entity_priority(struct amdgpu_ctx *ctx, } void amdgpu_ctx_priority_override(struct amdgpu_ctx *ctx, - enum drm_sched_priority priority) + int32_t priority) { - enum drm_sched_priority ctx_prio; + int32_t ctx_prio; unsigned i, j; ctx->override_priority = priority; - ctx_prio = (ctx->override_priority == DRM_SCHED_PRIORITY_UNSET) ? + ctx_prio = (ctx->override_priority == AMDGPU_CTX_PRIORITY_UNSET) ? ctx->init_priority : ctx->override_priority; for (i = 0; i < AMDGPU_HW_IP_NUM; ++i) { for (j = 0; j < amdgpu_ctx_num_entities[i]; ++j) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h index 14db16bc3322..a44b8b8ed39c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h @@ -47,8 +47,8 @@ struct amdgpu_ctx { spinlock_t ring_lock; struct amdgpu_ctx_entity *entities[AMDGPU_HW_IP_NUM][AMDGPU_MAX_ENTITY_NUM]; bool preamble_presented; - enum drm_sched_priority init_priority; - enum drm_sched_priority override_priority; + int32_t init_priority; + int32_t override_priority; struct mutex lock; atomic_t guilty; unsigned long ras_counter_ce; @@ -75,8 +75,8 @@ void amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct dma_fence *amdgpu_ctx_get_fence(struct amdgpu_ctx *ctx, struct drm_sched_entity *entity, uint64_t seq); -void amdgpu_ctx_priority_override(struct amdgpu_ctx *ctx, - enum drm_sched_priority priority); +bool amdgpu_ctx_priority_is_valid(int32_t ctx_prio); +void amdgpu_ctx_priority_override(struct amdgpu_ctx *ctx, int32_t ctx_prio); int amdgpu_ctx_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c index 463b9c0283f7..25e2e5bf90eb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c @@ -27,7 +27,6 @@ #include <linux/pci.h> #include <linux/uaccess.h> #include <linux/pm_runtime.h> -#include <linux/poll.h> #include "amdgpu.h" #include "amdgpu_pm.h" @@ -36,87 +35,10 @@ #include "amdgpu_rap.h" #include "amdgpu_securedisplay.h" #include "amdgpu_fw_attestation.h" - -int amdgpu_debugfs_wait_dump(struct amdgpu_device *adev) -{ -#if defined(CONFIG_DEBUG_FS) - unsigned long timeout = 600 * HZ; - int ret; - - wake_up_interruptible(&adev->autodump.gpu_hang); - - ret = wait_for_completion_interruptible_timeout(&adev->autodump.dumping, timeout); - if (ret == 0) { - pr_err("autodump: timeout, move on to gpu recovery\n"); - return -ETIMEDOUT; - } -#endif - return 0; -} +#include "amdgpu_umr.h" #if defined(CONFIG_DEBUG_FS) -static int amdgpu_debugfs_autodump_open(struct inode *inode, struct file *file) -{ - struct amdgpu_device *adev = inode->i_private; - int ret; - - file->private_data = adev; - - ret = down_read_killable(&adev->reset_sem); - if (ret) - return ret; - - if (adev->autodump.dumping.done) { - reinit_completion(&adev->autodump.dumping); - ret = 0; - } else { - ret = -EBUSY; - } - - up_read(&adev->reset_sem); - - return ret; -} - -static int amdgpu_debugfs_autodump_release(struct inode *inode, struct file *file) -{ - struct amdgpu_device *adev = file->private_data; - - complete_all(&adev->autodump.dumping); - return 0; -} - -static unsigned int amdgpu_debugfs_autodump_poll(struct file *file, struct poll_table_struct *poll_table) -{ - struct amdgpu_device *adev = file->private_data; - - poll_wait(file, &adev->autodump.gpu_hang, poll_table); - - if (amdgpu_in_reset(adev)) - return POLLIN | POLLRDNORM | POLLWRNORM; - - return 0; -} - -static const struct file_operations autodump_debug_fops = { - .owner = THIS_MODULE, - .open = amdgpu_debugfs_autodump_open, - .poll = amdgpu_debugfs_autodump_poll, - .release = amdgpu_debugfs_autodump_release, -}; - -static void amdgpu_debugfs_autodump_init(struct amdgpu_device *adev) -{ - init_completion(&adev->autodump.dumping); - complete_all(&adev->autodump.dumping); - init_waitqueue_head(&adev->autodump.gpu_hang); - - debugfs_create_file("amdgpu_autodump", 0600, - adev_to_drm(adev)->primary->debugfs_root, - adev, &autodump_debug_fops); -} - /** * amdgpu_debugfs_process_reg_op - Handle MMIO register reads/writes * @@ -279,6 +201,145 @@ static ssize_t amdgpu_debugfs_regs_write(struct file *f, const char __user *buf, return amdgpu_debugfs_process_reg_op(false, f, (char __user *)buf, size, pos); } +static int amdgpu_debugfs_regs2_open(struct inode *inode, struct file *file) +{ + struct amdgpu_debugfs_regs2_data *rd; + + rd = kzalloc(sizeof *rd, GFP_KERNEL); + if (!rd) + return -ENOMEM; + rd->adev = file_inode(file)->i_private; + file->private_data = rd; + mutex_init(&rd->lock); + + return 0; +} + +static int amdgpu_debugfs_regs2_release(struct inode *inode, struct file *file) +{ + struct amdgpu_debugfs_regs2_data *rd = file->private_data; + mutex_destroy(&rd->lock); + kfree(file->private_data); + return 0; +} + +static ssize_t amdgpu_debugfs_regs2_op(struct file *f, char __user *buf, u32 offset, size_t size, int write_en) +{ + struct amdgpu_debugfs_regs2_data *rd = f->private_data; + struct amdgpu_device *adev = rd->adev; + ssize_t result = 0; + int r; + uint32_t value; + + if (size & 0x3 || offset & 0x3) + return -EINVAL; + + r = pm_runtime_get_sync(adev_to_drm(adev)->dev); + if (r < 0) { + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + return r; + } + + r = amdgpu_virt_enable_access_debugfs(adev); + if (r < 0) { + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + return r; + } + + mutex_lock(&rd->lock); + + if (rd->id.use_grbm) { + if ((rd->id.grbm.sh != 0xFFFFFFFF && rd->id.grbm.sh >= adev->gfx.config.max_sh_per_se) || + (rd->id.grbm.se != 0xFFFFFFFF && rd->id.grbm.se >= adev->gfx.config.max_shader_engines)) { + pm_runtime_mark_last_busy(adev_to_drm(adev)->dev); + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + amdgpu_virt_disable_access_debugfs(adev); + mutex_unlock(&rd->lock); + return -EINVAL; + } + mutex_lock(&adev->grbm_idx_mutex); + amdgpu_gfx_select_se_sh(adev, rd->id.grbm.se, + rd->id.grbm.sh, + rd->id.grbm.instance); + } + + if (rd->id.use_srbm) { + mutex_lock(&adev->srbm_mutex); + amdgpu_gfx_select_me_pipe_q(adev, rd->id.srbm.me, rd->id.srbm.pipe, + rd->id.srbm.queue, rd->id.srbm.vmid); + } + + if (rd->id.pg_lock) + mutex_lock(&adev->pm.mutex); + + while (size) { + if (!write_en) { + value = RREG32(offset >> 2); + r = put_user(value, (uint32_t *)buf); + } else { + r = get_user(value, (uint32_t *)buf); + if (!r) + amdgpu_mm_wreg_mmio_rlc(adev, offset >> 2, value); + } + if (r) { + result = r; + goto end; + } + offset += 4; + size -= 4; + result += 4; + buf += 4; + } +end: + if (rd->id.use_grbm) { + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + mutex_unlock(&adev->grbm_idx_mutex); + } + + if (rd->id.use_srbm) { + amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + } + + if (rd->id.pg_lock) + mutex_unlock(&adev->pm.mutex); + + mutex_unlock(&rd->lock); + + pm_runtime_mark_last_busy(adev_to_drm(adev)->dev); + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + + amdgpu_virt_disable_access_debugfs(adev); + return result; +} + +static long amdgpu_debugfs_regs2_ioctl(struct file *f, unsigned int cmd, unsigned long data) +{ + struct amdgpu_debugfs_regs2_data *rd = f->private_data; + int r; + + switch (cmd) { + case AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE: + mutex_lock(&rd->lock); + r = copy_from_user(&rd->id, (struct amdgpu_debugfs_regs2_iocdata *)data, sizeof rd->id); + mutex_unlock(&rd->lock); + return r ? -EINVAL : 0; + default: + return -EINVAL; + } + return 0; +} + +static ssize_t amdgpu_debugfs_regs2_read(struct file *f, char __user *buf, size_t size, loff_t *pos) +{ + return amdgpu_debugfs_regs2_op(f, buf, *pos, size, 0); +} + +static ssize_t amdgpu_debugfs_regs2_write(struct file *f, const char __user *buf, size_t size, loff_t *pos) +{ + return amdgpu_debugfs_regs2_op(f, (char __user *)buf, *pos, size, 1); +} + /** * amdgpu_debugfs_regs_pcie_read - Read from a PCIE register @@ -1091,6 +1152,16 @@ static ssize_t amdgpu_debugfs_gfxoff_read(struct file *f, char __user *buf, return result; } +static const struct file_operations amdgpu_debugfs_regs2_fops = { + .owner = THIS_MODULE, + .unlocked_ioctl = amdgpu_debugfs_regs2_ioctl, + .read = amdgpu_debugfs_regs2_read, + .write = amdgpu_debugfs_regs2_write, + .open = amdgpu_debugfs_regs2_open, + .release = amdgpu_debugfs_regs2_release, + .llseek = default_llseek +}; + static const struct file_operations amdgpu_debugfs_regs_fops = { .owner = THIS_MODULE, .read = amdgpu_debugfs_regs_read, @@ -1148,6 +1219,7 @@ static const struct file_operations amdgpu_debugfs_gfxoff_fops = { static const struct file_operations *debugfs_regs[] = { &amdgpu_debugfs_regs_fops, + &amdgpu_debugfs_regs2_fops, &amdgpu_debugfs_regs_didt_fops, &amdgpu_debugfs_regs_pcie_fops, &amdgpu_debugfs_regs_smc_fops, @@ -1160,6 +1232,7 @@ static const struct file_operations *debugfs_regs[] = { static const char *debugfs_regs_names[] = { "amdgpu_regs", + "amdgpu_regs2", "amdgpu_regs_didt", "amdgpu_regs_pcie", "amdgpu_regs_smc", @@ -1206,7 +1279,7 @@ static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused) } /* Avoid accidently unparking the sched thread during GPU reset */ - r = down_read_killable(&adev->reset_sem); + r = down_write_killable(&adev->reset_sem); if (r) return r; @@ -1235,7 +1308,7 @@ static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused) kthread_unpark(ring->sched.thread); } - up_read(&adev->reset_sem); + up_write(&adev->reset_sem); pm_runtime_mark_last_busy(dev->dev); pm_runtime_put_autosuspend(dev->dev); @@ -1255,7 +1328,7 @@ static int amdgpu_debugfs_evict_vram(void *data, u64 *val) return r; } - *val = amdgpu_bo_evict_vram(adev); + *val = amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM); pm_runtime_mark_last_busy(dev->dev); pm_runtime_put_autosuspend(dev->dev); @@ -1268,17 +1341,15 @@ static int amdgpu_debugfs_evict_gtt(void *data, u64 *val) { struct amdgpu_device *adev = (struct amdgpu_device *)data; struct drm_device *dev = adev_to_drm(adev); - struct ttm_resource_manager *man; int r; r = pm_runtime_get_sync(dev->dev); if (r < 0) { - pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + pm_runtime_put_autosuspend(dev->dev); return r; } - man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT); - *val = ttm_resource_manager_evict_all(&adev->mman.bdev, man); + *val = amdgpu_ttm_evict_resources(adev, TTM_PL_TT); pm_runtime_mark_last_busy(dev->dev); pm_runtime_put_autosuspend(dev->dev); @@ -1544,6 +1615,12 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev) struct dentry *ent; int r, i; + if (!debugfs_initialized()) + return 0; + + debugfs_create_x32("amdgpu_smu_debug", 0600, root, + &adev->pm.smu_debug_mask); + ent = debugfs_create_file("amdgpu_preempt_ib", 0600, root, adev, &fops_ib_preempt); if (IS_ERR(ent)) { @@ -1582,13 +1659,10 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev) if (!ring) continue; - if (amdgpu_debugfs_ring_init(adev, ring)) { - DRM_ERROR("Failed to register debugfs file for rings !\n"); - } + amdgpu_debugfs_ring_init(adev, ring); } amdgpu_ras_debugfs_create_all(adev); - amdgpu_debugfs_autodump_init(adev); amdgpu_rap_debugfs_init(adev); amdgpu_securedisplay_debugfs_init(adev); amdgpu_fw_attestation_debugfs_init(adev); @@ -1607,6 +1681,11 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev) debugfs_create_blob("amdgpu_vbios", 0444, root, &adev->debugfs_vbios_blob); + adev->debugfs_discovery_blob.data = adev->mman.discovery_bin; + adev->debugfs_discovery_blob.size = adev->mman.discovery_tmr_size; + debugfs_create_blob("amdgpu_discovery", 0444, root, + &adev->debugfs_discovery_blob); + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h index 141a8474e24f..371a6f0deb29 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h @@ -22,14 +22,9 @@ * OTHER DEALINGS IN THE SOFTWARE. * */ - /* * Debugfs */ -struct amdgpu_autodump { - struct completion dumping; - struct wait_queue_head gpu_hang; -}; int amdgpu_debugfs_regs_init(struct amdgpu_device *adev); int amdgpu_debugfs_init(struct amdgpu_device *adev); @@ -37,4 +32,3 @@ void amdgpu_debugfs_fini(struct amdgpu_device *adev); void amdgpu_debugfs_fence_init(struct amdgpu_device *adev); void amdgpu_debugfs_firmware_init(struct amdgpu_device *adev); void amdgpu_debugfs_gem_init(struct amdgpu_device *adev); -int amdgpu_debugfs_wait_dump(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index af9bdf16eefd..a8b08a72b71b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -30,6 +30,7 @@ #include <linux/module.h> #include <linux/console.h> #include <linux/slab.h> +#include <linux/iommu.h> #include <drm/drm_atomic_helper.h> #include <drm/drm_probe_helper.h> @@ -125,6 +126,7 @@ const char *amdgpu_asic_name[] = { "DIMGREY_CAVEFISH", "BEIGE_GOBY", "YELLOW_CARP", + "IP DISCOVERY", "LAST", }; @@ -305,7 +307,7 @@ void amdgpu_device_mm_access(struct amdgpu_device *adev, loff_t pos, uint64_t last; int idx; - if (!drm_dev_enter(&adev->ddev, &idx)) + if (!drm_dev_enter(adev_to_drm(adev), &idx)) return; BUG_ON(!IS_ALIGNED(pos, 4) || !IS_ALIGNED(size, 4)); @@ -330,7 +332,7 @@ void amdgpu_device_mm_access(struct amdgpu_device *adev, loff_t pos, } /** - * amdgpu_device_vram_access - access vram by vram aperature + * amdgpu_device_aper_access - access vram by vram aperature * * @adev: amdgpu_device pointer * @pos: offset of the buffer in vram @@ -549,11 +551,11 @@ void amdgpu_device_wreg(struct amdgpu_device *adev, trace_amdgpu_device_wreg(adev->pdev->device, reg, v); } -/* +/** * amdgpu_mm_wreg_mmio_rlc - write register either with mmio or with RLC path if in range * * this function is invoked only the debugfs register access - * */ + */ void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev, uint32_t reg, uint32_t v) { @@ -1099,7 +1101,7 @@ static void amdgpu_device_wb_fini(struct amdgpu_device *adev) } /** - * amdgpu_device_wb_init- Init Writeback driver info and allocate memory + * amdgpu_device_wb_init - Init Writeback driver info and allocate memory * * @adev: amdgpu_device pointer * @@ -2126,46 +2128,11 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) if (r) return r; break; - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: - case CHIP_RAVEN: - case CHIP_ARCTURUS: - case CHIP_RENOIR: - case CHIP_ALDEBARAN: - if (adev->flags & AMD_IS_APU) - adev->family = AMDGPU_FAMILY_RV; - else - adev->family = AMDGPU_FAMILY_AI; - - r = soc15_set_ip_blocks(adev); - if (r) - return r; - break; - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_VANGOGH: - case CHIP_YELLOW_CARP: - case CHIP_CYAN_SKILLFISH: - if (adev->asic_type == CHIP_VANGOGH) - adev->family = AMDGPU_FAMILY_VGH; - else if (adev->asic_type == CHIP_YELLOW_CARP) - adev->family = AMDGPU_FAMILY_YC; - else - adev->family = AMDGPU_FAMILY_NV; - - r = nv_set_ip_blocks(adev); + default: + r = amdgpu_discovery_set_ip_blocks(adev); if (r) return r; break; - default: - /* FIXME: not supported yet */ - return -EINVAL; } amdgpu_amdkfd_device_probe(adev); @@ -2350,6 +2317,10 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) /* need to do gmc hw init early so we can allocate gpu mem */ if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC) { + /* Try to reserve bad pages early */ + if (amdgpu_sriov_vf(adev)) + amdgpu_virt_exchange_data(adev); + r = amdgpu_device_vram_scratch_init(adev); if (r) { DRM_ERROR("amdgpu_vram_scratch_init failed %d\n", r); @@ -2381,7 +2352,7 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) } if (amdgpu_sriov_vf(adev)) - amdgpu_virt_init_data_exchange(adev); + amdgpu_virt_exchange_data(adev); r = amdgpu_ib_pool_init(adev); if (r) { @@ -2432,10 +2403,6 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) if (!adev->gmc.xgmi.pending_reset) amdgpu_amdkfd_device_init(adev); - r = amdgpu_amdkfd_resume_iommu(adev); - if (r) - goto init_failed; - amdgpu_fru_get_product_info(adev); init_failed: @@ -2652,11 +2619,10 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev) if (r) DRM_ERROR("enable mgpu fan boost failed (%d).\n", r); - /* For XGMI + passthrough configuration on arcturus, enable light SBR */ - if (adev->asic_type == CHIP_ARCTURUS && - amdgpu_passthrough(adev) && - adev->gmc.xgmi.num_physical_nodes > 1) - smu_set_light_sbr(&adev->smu, true); + /* For passthrough configuration on arcturus and aldebaran, enable special handling SBR */ + if (amdgpu_passthrough(adev) && ((adev->asic_type == CHIP_ARCTURUS && adev->gmc.xgmi.num_physical_nodes > 1)|| + adev->asic_type == CHIP_ALDEBARAN )) + smu_handle_passthrough_sbr(&adev->smu, true); if (adev->gmc.xgmi.num_physical_nodes > 1) { mutex_lock(&mgpu_info.mutex); @@ -2695,6 +2661,36 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev) return 0; } +/** + * amdgpu_device_smu_fini_early - smu hw_fini wrapper + * + * @adev: amdgpu_device pointer + * + * For ASICs need to disable SMC first + */ +static void amdgpu_device_smu_fini_early(struct amdgpu_device *adev) +{ + int i, r; + + if (adev->ip_versions[GC_HWIP][0] > IP_VERSION(9, 0, 0)) + return; + + for (i = 0; i < adev->num_ip_blocks; i++) { + if (!adev->ip_blocks[i].status.hw) + continue; + if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC) { + r = adev->ip_blocks[i].version->funcs->hw_fini((void *)adev); + /* XXX handle errors */ + if (r) { + DRM_DEBUG("hw_fini of IP block <%s> failed %d\n", + adev->ip_blocks[i].version->funcs->name, r); + } + adev->ip_blocks[i].status.hw = false; + break; + } + } +} + static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev) { int i, r; @@ -2715,21 +2711,8 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev) amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE); amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE); - /* need to disable SMC first */ - for (i = 0; i < adev->num_ip_blocks; i++) { - if (!adev->ip_blocks[i].status.hw) - continue; - if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC) { - r = adev->ip_blocks[i].version->funcs->hw_fini((void *)adev); - /* XXX handle errors */ - if (r) { - DRM_DEBUG("hw_fini of IP block <%s> failed %d\n", - adev->ip_blocks[i].version->funcs->name, r); - } - adev->ip_blocks[i].status.hw = false; - break; - } - } + /* Workaroud for ASICs need to disable SMC first */ + amdgpu_device_smu_fini_early(adev); for (i = adev->num_ip_blocks - 1; i >= 0; i--) { if (!adev->ip_blocks[i].status.hw) @@ -2745,6 +2728,11 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev) adev->ip_blocks[i].status.hw = false; } + if (amdgpu_sriov_vf(adev)) { + if (amdgpu_virt_release_full_gpu(adev, false)) + DRM_ERROR("failed to release exclusive mode on fini\n"); + } + return 0; } @@ -2766,8 +2754,6 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev) if (amdgpu_sriov_vf(adev) && adev->virt.ras_init_done) amdgpu_virt_release_ras_err_handler_data(adev); - amdgpu_ras_pre_fini(adev); - if (adev->gmc.xgmi.num_physical_nodes > 1) amdgpu_xgmi_remove_device(adev); @@ -2805,10 +2791,6 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev) amdgpu_ras_fini(adev); - if (amdgpu_sriov_vf(adev)) - if (amdgpu_virt_release_full_gpu(adev, false)) - DRM_ERROR("failed to release exclusive mode on fini\n"); - return 0; } @@ -3203,12 +3185,28 @@ static void amdgpu_device_detect_sriov_bios(struct amdgpu_device *adev) bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type) { switch (asic_type) { +#ifdef CONFIG_DRM_AMDGPU_SI + case CHIP_HAINAN: +#endif + case CHIP_TOPAZ: + /* chips with no display hardware */ + return false; #if defined(CONFIG_DRM_AMD_DC) -#if defined(CONFIG_DRM_AMD_DC_SI) case CHIP_TAHITI: case CHIP_PITCAIRN: case CHIP_VERDE: case CHIP_OLAND: + /* + * We have systems in the wild with these ASICs that require + * LVDS and VGA support which is not supported with DC. + * + * Fallback to the non-DC driver here by default so as not to + * cause regressions. + */ +#if defined(CONFIG_DRM_AMD_DC_SI) + return amdgpu_dc > 0; +#else + return false; #endif case CHIP_BONAIRE: case CHIP_KAVERI: @@ -3240,6 +3238,7 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type) case CHIP_NAVI14: case CHIP_NAVI12: case CHIP_RENOIR: + case CHIP_CYAN_SKILLFISH: case CHIP_SIENNA_CICHLID: case CHIP_NAVY_FLOUNDER: case CHIP_DIMGREY_CAVEFISH: @@ -3247,13 +3246,15 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type) case CHIP_VANGOGH: case CHIP_YELLOW_CARP: #endif + default: return amdgpu_dc != 0; -#endif +#else default: if (amdgpu_dc > 0) DRM_INFO_ONCE("Display Core has been requested via kernel parameter " "but isn't supported by ASIC, ignoring\n"); return false; +#endif } } @@ -3354,6 +3355,8 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev) continue; } else if (timeout < 0) { timeout = MAX_SCHEDULE_TIMEOUT; + dev_warn(adev->dev, "lockup timeout disabled"); + add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK); } else { timeout = msecs_to_jiffies(timeout); } @@ -3389,6 +3392,22 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev) return ret; } +/** + * amdgpu_device_check_iommu_direct_map - check if RAM direct mapped to GPU + * + * @adev: amdgpu_device pointer + * + * RAM direct mapped to GPU if IOMMU is not enabled or is pass through mode + */ +static void amdgpu_device_check_iommu_direct_map(struct amdgpu_device *adev) +{ + struct iommu_domain *domain; + + domain = iommu_get_domain_for_dev(adev->dev); + if (!domain || domain->type == IOMMU_DOMAIN_IDENTITY) + adev->ram_is_direct_mapped = true; +} + static const struct attribute *amdgpu_dev_attributes[] = { &dev_attr_product_name.attr, &dev_attr_product_number.attr, @@ -3531,6 +3550,9 @@ int amdgpu_device_init(struct amdgpu_device *adev, adev->rmmio_size = pci_resource_len(adev->pdev, 2); } + for (i = 0; i < AMD_IP_BLOCK_TYPE_NUM; i++) + atomic_set(&adev->pm.pwr_state[i], POWER_STATE_UNKNOWN); + adev->rmmio = ioremap(adev->rmmio_base, adev->rmmio_size); if (adev->rmmio == NULL) { return -ENOMEM; @@ -3538,17 +3560,6 @@ int amdgpu_device_init(struct amdgpu_device *adev, DRM_INFO("register mmio base: 0x%08X\n", (uint32_t)adev->rmmio_base); DRM_INFO("register mmio size: %u\n", (unsigned)adev->rmmio_size); - /* enable PCIE atomic ops */ - r = pci_enable_atomic_ops_to_root(adev->pdev, - PCI_EXP_DEVCAP2_ATOMIC_COMP32 | - PCI_EXP_DEVCAP2_ATOMIC_COMP64); - if (r) { - adev->have_atomics_support = false; - DRM_INFO("PCIE atomic ops is not supported\n"); - } else { - adev->have_atomics_support = true; - } - amdgpu_device_get_pcie_info(adev); if (amdgpu_mcbp) @@ -3571,6 +3582,26 @@ int amdgpu_device_init(struct amdgpu_device *adev, if (r) return r; + /* Need to get xgmi info early to decide the reset behavior*/ + if (adev->gmc.xgmi.supported) { + r = adev->gfxhub.funcs->get_xgmi_info(adev); + if (r) + return r; + } + + /* enable PCIE atomic ops */ + if (amdgpu_sriov_vf(adev)) + adev->have_atomics_support = ((struct amd_sriov_msg_pf2vf_info *) + adev->virt.fw_reserve.p_pf2vf)->pcie_atomic_ops_enabled_flags == + (PCI_EXP_DEVCAP2_ATOMIC_COMP32 | PCI_EXP_DEVCAP2_ATOMIC_COMP64); + else + adev->have_atomics_support = + !pci_enable_atomic_ops_to_root(adev->pdev, + PCI_EXP_DEVCAP2_ATOMIC_COMP32 | + PCI_EXP_DEVCAP2_ATOMIC_COMP64); + if (!adev->have_atomics_support) + dev_info(adev->dev, "PCIE atomic ops is not supported\n"); + /* doorbell bar mapping and doorbell index init*/ amdgpu_device_doorbell_init(adev); @@ -3704,8 +3735,6 @@ fence_driver_init: /* Get a log2 for easy divisions. */ adev->mm_stats.log2_max_MBps = ilog2(max(1u, max_MBps)); - amdgpu_fbdev_init(adev); - r = amdgpu_pm_sysfs_init(adev); if (r) { adev->pm_sysfs_en = false; @@ -3789,6 +3818,8 @@ fence_driver_init: queue_delayed_work(system_wq, &mgpu_info.delayed_reset_work, msecs_to_jiffies(AMDGPU_RESUME_MS)); + amdgpu_device_check_iommu_direct_map(adev); + return 0; release_ras_con: @@ -3822,7 +3853,7 @@ static void amdgpu_device_unmap_mmio(struct amdgpu_device *adev) } /** - * amdgpu_device_fini - tear down the driver + * amdgpu_device_fini_hw - tear down the driver * * @adev: amdgpu_device pointer * @@ -3850,7 +3881,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev) /* disable all interrupts */ amdgpu_irq_disable_all(adev); if (adev->mode_info.mode_config_initialized){ - if (!amdgpu_device_has_dc_support(adev)) + if (!drm_drv_uses_atomic_modeset(adev_to_drm(adev))) drm_helper_force_disable_all(adev_to_drm(adev)); else drm_atomic_helper_shutdown(adev_to_drm(adev)); @@ -3863,21 +3894,27 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev) amdgpu_ucode_sysfs_fini(adev); sysfs_remove_files(&adev->dev->kobj, amdgpu_dev_attributes); - amdgpu_fbdev_fini(adev); + /* disable ras feature must before hw fini */ + amdgpu_ras_pre_fini(adev); + + amdgpu_device_ip_fini_early(adev); amdgpu_irq_fini_hw(adev); - amdgpu_device_ip_fini_early(adev); + if (adev->mman.initialized) + ttm_device_clear_dma_mappings(&adev->mman.bdev); amdgpu_gart_dummy_page_fini(adev); - amdgpu_device_unmap_mmio(adev); + if (drm_dev_is_unplugged(adev_to_drm(adev))) + amdgpu_device_unmap_mmio(adev); + } void amdgpu_device_fini_sw(struct amdgpu_device *adev) { - amdgpu_device_ip_fini(adev); amdgpu_fence_driver_sw_fini(adev); + amdgpu_device_ip_fini(adev); release_firmware(adev->firmware.gpu_info_fw); adev->firmware.gpu_info_fw = NULL; adev->accel_working = false; @@ -3909,6 +3946,25 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev) } +/** + * amdgpu_device_evict_resources - evict device resources + * @adev: amdgpu device object + * + * Evicts all ttm device resources(vram BOs, gart table) from the lru list + * of the vram memory type. Mainly used for evicting device resources + * at suspend time. + * + */ +static void amdgpu_device_evict_resources(struct amdgpu_device *adev) +{ + /* No need to evict vram on APUs for suspend to ram */ + if (adev->in_s3 && (adev->flags & AMD_IS_APU)) + return; + + if (amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM)) + DRM_WARN("evicting device resources failed\n"); + +} /* * Suspend & resume. @@ -3938,7 +3994,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) drm_kms_helper_poll_disable(dev); if (fbcon) - amdgpu_fbdev_set_suspend(adev, 1); + drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, true); cancel_delayed_work_sync(&adev->delayed_init_work); @@ -3949,17 +4005,16 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) if (!adev->in_s0ix) amdgpu_amdkfd_suspend(adev, adev->in_runpm); - /* evict vram memory */ - amdgpu_bo_evict_vram(adev); + /* First evict vram memory */ + amdgpu_device_evict_resources(adev); amdgpu_fence_driver_hw_fini(adev); amdgpu_device_ip_suspend_phase2(adev); - /* evict remaining vram memory - * This second call to evict vram is to evict the gart page table - * using the CPU. + /* This second call to evict device resources is to evict + * the gart page table using the CPU. */ - amdgpu_bo_evict_vram(adev); + amdgpu_device_evict_resources(adev); return 0; } @@ -4016,7 +4071,7 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon) flush_delayed_work(&adev->delayed_init_work); if (fbcon) - amdgpu_fbdev_set_suspend(adev, 0); + drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, false); drm_kms_helper_poll_enable(dev); @@ -4285,6 +4340,11 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev, bool from_hypervisor) { int r; + struct amdgpu_hive_info *hive = NULL; + + amdgpu_amdkfd_pre_reset(adev); + + amdgpu_amdkfd_pre_reset(adev); if (from_hypervisor) r = amdgpu_virt_request_full_gpu(adev, true); @@ -4293,8 +4353,6 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev, if (r) return r; - amdgpu_amdkfd_pre_reset(adev); - /* Resume IP prior to SMC */ r = amdgpu_device_ip_reinit_early_sriov(adev); if (r) @@ -4313,9 +4371,19 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev, if (r) goto error; - amdgpu_irq_gpu_reset_resume_helper(adev); - r = amdgpu_ib_ring_tests(adev); - amdgpu_amdkfd_post_reset(adev); + hive = amdgpu_get_xgmi_hive(adev); + /* Update PSP FW topology after reset */ + if (hive && adev->gmc.xgmi.num_physical_nodes > 1) + r = amdgpu_xgmi_update_topology(hive, adev); + + if (hive) + amdgpu_put_xgmi_hive(hive); + + if (!r) { + amdgpu_irq_gpu_reset_resume_helper(adev); + r = amdgpu_ib_ring_tests(adev); + amdgpu_amdkfd_post_reset(adev); + } error: if (!r && adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST) { @@ -4458,7 +4526,7 @@ int amdgpu_device_mode1_reset(struct amdgpu_device *adev) int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, struct amdgpu_reset_context *reset_context) { - int i, j, r = 0; + int i, r = 0; struct amdgpu_job *job = NULL; bool need_full_reset = test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); @@ -4466,10 +4534,6 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, if (reset_context->reset_req_dev == adev) job = reset_context->job; - /* no need to dump if device is not in good state during probe period */ - if (!adev->gmc.xgmi.pending_reset) - amdgpu_debugfs_wait_dump(adev); - if (amdgpu_sriov_vf(adev)) { /* stop the data exchange thread */ amdgpu_virt_fini_data_exchange(adev); @@ -4484,15 +4548,8 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, /*clear job fence from fence drv to avoid force_completion *leave NULL and vm flush fence in fence drv */ - for (j = 0; j <= ring->fence_drv.num_fences_mask; j++) { - struct dma_fence *old, **ptr; + amdgpu_fence_driver_clear_job_fences(ring); - ptr = &ring->fence_drv.fences[j]; - old = rcu_dereference_protected(*ptr, 1); - if (old && test_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &old->flags)) { - RCU_INIT_POINTER(*ptr, NULL); - } - } /* after all hw jobs are reset, hw fence is meaningless, so force_completion */ amdgpu_fence_driver_force_completion(ring); } @@ -4652,7 +4709,7 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle, if (r) goto out; - amdgpu_fbdev_set_suspend(tmp_adev, 0); + drm_fb_helper_set_suspend_unlocked(adev_to_drm(tmp_adev)->fb_helper, false); /* * The GPU enters bad state once faulty pages @@ -4751,7 +4808,7 @@ static int amdgpu_device_lock_hive_adev(struct amdgpu_device *adev, struct amdgp { struct amdgpu_device *tmp_adev = NULL; - if (adev->gmc.xgmi.num_physical_nodes > 1) { + if (!amdgpu_sriov_vf(adev) && (adev->gmc.xgmi.num_physical_nodes > 1)) { if (!hive) { dev_err(adev->dev, "Hive is NULL while device has multiple xgmi nodes"); return -ENODEV; @@ -4860,6 +4917,9 @@ static void amdgpu_device_recheck_guilty_jobs( /* clear job's guilty and depend the folowing step to decide the real one */ drm_sched_reset_karma(s_job); + /* for the real bad job, it will be resubmitted twice, adding a dma_fence_get + * to make sure fence is balanced */ + dma_fence_get(s_job->s_fence->parent); drm_sched_resubmit_jobs_ext(&ring->sched, 1); ret = dma_fence_wait_timeout(s_job->s_fence->parent, false, ring->sched.timeout); @@ -4895,6 +4955,7 @@ retry: /* got the hw fence, signal finished fence */ atomic_dec(ring->sched.score); + dma_fence_put(s_job->s_fence->parent); dma_fence_get(&s_job->s_fence->finished); dma_fence_signal(&s_job->s_fence->finished); dma_fence_put(&s_job->s_fence->finished); @@ -4959,7 +5020,8 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, * We always reset all schedulers for device and all devices for XGMI * hive so that should take care of them too. */ - hive = amdgpu_get_xgmi_hive(adev); + if (!amdgpu_sriov_vf(adev)) + hive = amdgpu_get_xgmi_hive(adev); if (hive) { if (atomic_cmpxchg(&hive->in_reset, 0, 1) != 0) { DRM_INFO("Bailing on TDR for s_job:%llx, hive: %llx as another already in progress", @@ -5000,7 +5062,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, * to put adev in the 1st position. */ INIT_LIST_HEAD(&device_list); - if (adev->gmc.xgmi.num_physical_nodes > 1) { + if (!amdgpu_sriov_vf(adev) && (adev->gmc.xgmi.num_physical_nodes > 1)) { list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) list_add_tail(&tmp_adev->reset_list, &device_list); if (!list_is_first(&adev->reset_list, &device_list)) @@ -5039,7 +5101,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, */ amdgpu_unregister_gpu_instance(tmp_adev); - amdgpu_fbdev_set_suspend(tmp_adev, 1); + drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, true); /* disable ras on ALL IPs */ if (!need_emergency_restart && @@ -5089,7 +5151,7 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ tmp_vram_lost_counter = atomic_read(&((adev)->vram_lost_counter)); /* Actual ASIC resets if needed.*/ - /* TODO Implement XGMI hive reset logic for SRIOV */ + /* Host driver will handle XGMI hive reset for SRIOV */ if (amdgpu_sriov_vf(adev)) { r = amdgpu_device_reset_sriov(adev, job ? false : true); if (r) @@ -5130,7 +5192,7 @@ skip_hw_reset: drm_sched_start(&ring->sched, !tmp_adev->asic_reset_res); } - if (!amdgpu_device_has_dc_support(tmp_adev) && !job_signaled) { + if (!drm_drv_uses_atomic_modeset(adev_to_drm(tmp_adev)) && !job_signaled) { drm_helper_resume_force_mode(adev_to_drm(tmp_adev)); } @@ -5151,7 +5213,7 @@ skip_sched_resume: list_for_each_entry(tmp_adev, device_list_handle, reset_list) { /* unlock kfd: SRIOV would do it separately */ if (!need_emergency_restart && !amdgpu_sriov_vf(tmp_adev)) - amdgpu_amdkfd_post_reset(tmp_adev); + amdgpu_amdkfd_post_reset(tmp_adev); /* kfd_post_reset will do nothing if kfd device is not initialized, * need to bring up kfd here if it's not be initialized before @@ -5634,3 +5696,42 @@ void amdgpu_device_invalidate_hdp(struct amdgpu_device *adev, amdgpu_asic_invalidate_hdp(adev, ring); } + +/** + * amdgpu_device_halt() - bring hardware to some kind of halt state + * + * @adev: amdgpu_device pointer + * + * Bring hardware to some kind of halt state so that no one can touch it + * any more. It will help to maintain error context when error occurred. + * Compare to a simple hang, the system will keep stable at least for SSH + * access. Then it should be trivial to inspect the hardware state and + * see what's going on. Implemented as following: + * + * 1. drm_dev_unplug() makes device inaccessible to user space(IOCTLs, etc), + * clears all CPU mappings to device, disallows remappings through page faults + * 2. amdgpu_irq_disable_all() disables all interrupts + * 3. amdgpu_fence_driver_hw_fini() signals all HW fences + * 4. set adev->no_hw_access to avoid potential crashes after setp 5 + * 5. amdgpu_device_unmap_mmio() clears all MMIO mappings + * 6. pci_disable_device() and pci_wait_for_pending_transaction() + * flush any in flight DMA operations + */ +void amdgpu_device_halt(struct amdgpu_device *adev) +{ + struct pci_dev *pdev = adev->pdev; + struct drm_device *ddev = adev_to_drm(adev); + + drm_dev_unplug(ddev); + + amdgpu_irq_disable_all(adev); + + amdgpu_fence_driver_hw_fini(adev); + + adev->no_hw_access = true; + + amdgpu_device_unmap_mmio(adev); + + pci_disable_device(pdev); + pci_wait_for_pending_transaction(pdev); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_df.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_df.h index 52488bb45112..6b25837955c4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_df.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_df.h @@ -52,6 +52,7 @@ struct amdgpu_df_funcs { uint64_t (*get_fica)(struct amdgpu_device *adev, uint32_t ficaa_val); void (*set_fica)(struct amdgpu_device *adev, uint32_t ficaa_val, uint32_t ficadl_val, uint32_t ficadh_val); + bool (*query_ras_poison_mode)(struct amdgpu_device *adev); }; struct amdgpu_df { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c index ada7bc19118a..028190d42bb2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c @@ -21,16 +21,59 @@ * */ +#include <linux/firmware.h> + #include "amdgpu.h" #include "amdgpu_discovery.h" #include "soc15_hw_ip.h" #include "discovery.h" +#include "soc15.h" +#include "gfx_v9_0.h" +#include "gmc_v9_0.h" +#include "df_v1_7.h" +#include "df_v3_6.h" +#include "nbio_v6_1.h" +#include "nbio_v7_0.h" +#include "nbio_v7_4.h" +#include "hdp_v4_0.h" +#include "vega10_ih.h" +#include "vega20_ih.h" +#include "sdma_v4_0.h" +#include "uvd_v7_0.h" +#include "vce_v4_0.h" +#include "vcn_v1_0.h" +#include "vcn_v2_5.h" +#include "jpeg_v2_5.h" +#include "smuio_v9_0.h" +#include "gmc_v10_0.h" +#include "gfxhub_v2_0.h" +#include "mmhub_v2_0.h" +#include "nbio_v2_3.h" +#include "nbio_v7_2.h" +#include "hdp_v5_0.h" +#include "nv.h" +#include "navi10_ih.h" +#include "gfx_v10_0.h" +#include "sdma_v5_0.h" +#include "sdma_v5_2.h" +#include "vcn_v2_0.h" +#include "jpeg_v2_0.h" +#include "vcn_v3_0.h" +#include "jpeg_v3_0.h" +#include "amdgpu_vkms.h" +#include "mes_v10_1.h" +#include "smuio_v11_0.h" +#include "smuio_v11_0_6.h" +#include "smuio_v13_0.h" + +#define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin" +MODULE_FIRMWARE(FIRMWARE_IP_DISCOVERY); + #define mmRCC_CONFIG_MEMSIZE 0xde3 #define mmMM_INDEX 0x0 #define mmMM_INDEX_HI 0x6 #define mmMM_DATA 0x1 -#define HW_ID_MAX 300 static const char *hw_id_names[HW_ID_MAX] = { [MP1_HWID] = "MP1", @@ -66,6 +109,8 @@ static const char *hw_id_names[HW_ID_MAX] = { [HDP_HWID] = "HDP", [SDMA0_HWID] = "SDMA0", [SDMA1_HWID] = "SDMA1", + [SDMA2_HWID] = "SDMA2", + [SDMA3_HWID] = "SDMA3", [ISP_HWID] = "ISP", [DBGU_IO_HWID] = "DBGU_IO", [DF_HWID] = "DF", @@ -113,6 +158,8 @@ static int hw_id_map[MAX_HWIP] = { [HDP_HWIP] = HDP_HWID, [SDMA0_HWIP] = SDMA0_HWID, [SDMA1_HWIP] = SDMA1_HWID, + [SDMA2_HWIP] = SDMA2_HWID, + [SDMA3_HWIP] = SDMA3_HWID, [MMHUB_HWIP] = MMHUB_HWID, [ATHUB_HWIP] = ATHUB_HWID, [NBIO_HWIP] = NBIF_HWID, @@ -129,9 +176,11 @@ static int hw_id_map[MAX_HWIP] = { [THM_HWIP] = THM_HWID, [CLK_HWIP] = CLKA_HWID, [UMC_HWIP] = UMC_HWID, + [XGMI_HWIP] = XGMI_HWID, + [DCI_HWIP] = DCI_HWID, }; -static int amdgpu_discovery_read_binary(struct amdgpu_device *adev, uint8_t *binary) +static int amdgpu_discovery_read_binary_from_vram(struct amdgpu_device *adev, uint8_t *binary) { uint64_t vram_size = (uint64_t)RREG32(mmRCC_CONFIG_MEMSIZE) << 20; uint64_t pos = vram_size - DISCOVERY_TMR_OFFSET; @@ -141,6 +190,34 @@ static int amdgpu_discovery_read_binary(struct amdgpu_device *adev, uint8_t *bin return 0; } +static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev, uint8_t *binary) +{ + const struct firmware *fw; + const char *fw_name; + int r; + + switch (amdgpu_discovery) { + case 2: + fw_name = FIRMWARE_IP_DISCOVERY; + break; + default: + dev_warn(adev->dev, "amdgpu_discovery is not set properly\n"); + return -EINVAL; + } + + r = request_firmware(&fw, fw_name, adev->dev); + if (r) { + dev_err(adev->dev, "can't load firmware \"%s\"\n", + fw_name); + return r; + } + + memcpy((u8 *)binary, (u8 *)fw->data, adev->mman.discovery_tmr_size); + release_firmware(fw); + + return 0; +} + static uint16_t amdgpu_discovery_calculate_checksum(uint8_t *data, uint32_t size) { uint16_t checksum = 0; @@ -158,6 +235,14 @@ static inline bool amdgpu_discovery_verify_checksum(uint8_t *data, uint32_t size return !!(amdgpu_discovery_calculate_checksum(data, size) == expected); } +static inline bool amdgpu_discovery_verify_binary_signature(uint8_t *binary) +{ + struct binary_header *bhdr; + bhdr = (struct binary_header *)binary; + + return (le32_to_cpu(bhdr->binary_signature) == BINARY_SIGNATURE); +} + static int amdgpu_discovery_init(struct amdgpu_device *adev) { struct table_info *info; @@ -174,28 +259,40 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) if (!adev->mman.discovery_bin) return -ENOMEM; - r = amdgpu_discovery_read_binary(adev, adev->mman.discovery_bin); + r = amdgpu_discovery_read_binary_from_vram(adev, adev->mman.discovery_bin); if (r) { - DRM_ERROR("failed to read ip discovery binary\n"); + dev_err(adev->dev, "failed to read ip discovery binary from vram\n"); + r = -EINVAL; goto out; } - bhdr = (struct binary_header *)adev->mman.discovery_bin; - - if (le32_to_cpu(bhdr->binary_signature) != BINARY_SIGNATURE) { - DRM_ERROR("invalid ip discovery binary signature\n"); - r = -EINVAL; - goto out; + if(!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin)) { + dev_warn(adev->dev, "get invalid ip discovery binary signature from vram\n"); + /* retry read ip discovery binary from file */ + r = amdgpu_discovery_read_binary_from_file(adev, adev->mman.discovery_bin); + if (r) { + dev_err(adev->dev, "failed to read ip discovery binary from file\n"); + r = -EINVAL; + goto out; + } + /* check the ip discovery binary signature */ + if(!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin)) { + dev_warn(adev->dev, "get invalid ip discovery binary signature from file\n"); + r = -EINVAL; + goto out; + } } + bhdr = (struct binary_header *)adev->mman.discovery_bin; + offset = offsetof(struct binary_header, binary_checksum) + sizeof(bhdr->binary_checksum); - size = bhdr->binary_size - offset; - checksum = bhdr->binary_checksum; + size = le16_to_cpu(bhdr->binary_size) - offset; + checksum = le16_to_cpu(bhdr->binary_checksum); if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, size, checksum)) { - DRM_ERROR("invalid ip discovery binary checksum\n"); + dev_err(adev->dev, "invalid ip discovery binary checksum\n"); r = -EINVAL; goto out; } @@ -206,14 +303,14 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin + offset); if (le32_to_cpu(ihdr->signature) != DISCOVERY_TABLE_SIGNATURE) { - DRM_ERROR("invalid ip discovery data table signature\n"); + dev_err(adev->dev, "invalid ip discovery data table signature\n"); r = -EINVAL; goto out; } if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, - ihdr->size, checksum)) { - DRM_ERROR("invalid ip discovery data table checksum\n"); + le16_to_cpu(ihdr->size), checksum)) { + dev_err(adev->dev, "invalid ip discovery data table checksum\n"); r = -EINVAL; goto out; } @@ -224,8 +321,8 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) ghdr = (struct gpu_info_header *)(adev->mman.discovery_bin + offset); if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, - ghdr->size, checksum)) { - DRM_ERROR("invalid gc data table checksum\n"); + le32_to_cpu(ghdr->size), checksum)) { + dev_err(adev->dev, "invalid gc data table checksum\n"); r = -EINVAL; goto out; } @@ -245,6 +342,22 @@ void amdgpu_discovery_fini(struct amdgpu_device *adev) adev->mman.discovery_bin = NULL; } +static int amdgpu_discovery_validate_ip(const struct ip *ip) +{ + if (ip->number_instance >= HWIP_MAX_INSTANCE) { + DRM_ERROR("Unexpected number_instance (%d) from ip discovery blob\n", + ip->number_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)); + return -EINVAL; + } + + return 0; +} + int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) { struct binary_header *bhdr; @@ -290,6 +403,10 @@ int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) for (j = 0; j < num_ips; j++) { ip = (struct ip *)(adev->mman.discovery_bin + ip_offset); + + if (amdgpu_discovery_validate_ip(ip)) + goto next_ip; + num_base_address = ip->num_base_address; DRM_DEBUG("%s(%d) #%d v%d.%d.%d:\n", @@ -299,8 +416,23 @@ int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) ip->major, ip->minor, ip->revision); - if (le16_to_cpu(ip->hw_id) == VCN_HWID) + if (le16_to_cpu(ip->hw_id) == VCN_HWID) { + /* Bit [5:0]: original revision value + * Bit [7:6]: en/decode capability: + * 0b00 : VCN function normally + * 0b10 : encode is disabled + * 0b01 : decode is disabled + */ + adev->vcn.vcn_config[adev->vcn.num_vcn_inst] = + ip->revision & 0xc0; + ip->revision &= ~0xc0; adev->vcn.num_vcn_inst++; + } + if (le16_to_cpu(ip->hw_id) == SDMA0_HWID || + le16_to_cpu(ip->hw_id) == SDMA1_HWID || + le16_to_cpu(ip->hw_id) == SDMA2_HWID || + le16_to_cpu(ip->hw_id) == SDMA3_HWID) + adev->sdma.num_instances++; for (k = 0; k < num_base_address; k++) { /* @@ -317,10 +449,21 @@ int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) hw_id_names[le16_to_cpu(ip->hw_id)]); adev->reg_offset[hw_ip][ip->number_instance] = ip->base_address; + /* Instance support is somewhat inconsistent. + * SDMA is a good example. Sienna cichlid has 4 total + * SDMA instances, each enumerated separately (HWIDs + * 42, 43, 68, 69). Arcturus has 8 total SDMA instances, + * but they are enumerated as multiple instances of the + * same HWIDs (4x HWID 42, 4x HWID 43). UMC is another + * example. On most chips there are multiple instances + * with the same HWID. + */ + adev->ip_versions[hw_ip][ip->number_instance] = + IP_VERSION(ip->major, ip->minor, ip->revision); } - } +next_ip: ip_offset += sizeof(*ip) + 4 * (ip->num_base_address - 1); } } @@ -376,14 +519,6 @@ int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int n return -EINVAL; } - -int amdgpu_discovery_get_vcn_version(struct amdgpu_device *adev, int vcn_instance, - int *major, int *minor, int *revision) -{ - return amdgpu_discovery_get_ip_version(adev, VCN_HWID, - vcn_instance, major, minor, revision); -} - void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) { struct binary_header *bhdr; @@ -395,12 +530,16 @@ void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) le16_to_cpu(bhdr->table_list[HARVEST_INFO].offset)); for (i = 0; i < 32; i++) { - if (le32_to_cpu(harvest_info->list[i].hw_id) == 0) + if (le16_to_cpu(harvest_info->list[i].hw_id) == 0) break; - switch (le32_to_cpu(harvest_info->list[i].hw_id)) { + switch (le16_to_cpu(harvest_info->list[i].hw_id)) { case VCN_HWID: vcn_harvest_count++; + if (harvest_info->list[i].number_instance == 0) + adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0; + else + adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; break; case DMU_HWID: adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK; @@ -409,16 +548,32 @@ void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) break; } } + /* some IP discovery tables on Navy Flounder don't have this set correctly */ + if ((adev->ip_versions[UVD_HWIP][1] == IP_VERSION(3, 0, 1)) && + (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 2))) + adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; if (vcn_harvest_count == adev->vcn.num_vcn_inst) { adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK; adev->harvest_ip_mask |= AMD_HARVEST_IP_JPEG_MASK; } + if ((adev->pdev->device == 0x731E && + (adev->pdev->revision == 0xC6 || adev->pdev->revision == 0xC7)) || + (adev->pdev->device == 0x7340 && adev->pdev->revision == 0xC9) || + (adev->pdev->device == 0x7360 && adev->pdev->revision == 0xC7)) { + adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK; + adev->harvest_ip_mask |= AMD_HARVEST_IP_JPEG_MASK; + } } +union gc_info { + struct gc_info_v1_0 v1; + struct gc_info_v2_0 v2; +}; + int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev) { struct binary_header *bhdr; - struct gc_info_v1_0 *gc_info; + union gc_info *gc_info; if (!adev->mman.discovery_bin) { DRM_ERROR("ip discovery uninitialized\n"); @@ -426,27 +581,842 @@ int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev) } bhdr = (struct binary_header *)adev->mman.discovery_bin; - gc_info = (struct gc_info_v1_0 *)(adev->mman.discovery_bin + + gc_info = (union gc_info *)(adev->mman.discovery_bin + le16_to_cpu(bhdr->table_list[GC].offset)); + switch (gc_info->v1.header.version_major) { + case 1: + adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->v1.gc_num_se); + adev->gfx.config.max_cu_per_sh = 2 * (le32_to_cpu(gc_info->v1.gc_num_wgp0_per_sa) + + le32_to_cpu(gc_info->v1.gc_num_wgp1_per_sa)); + adev->gfx.config.max_sh_per_se = le32_to_cpu(gc_info->v1.gc_num_sa_per_se); + adev->gfx.config.max_backends_per_se = le32_to_cpu(gc_info->v1.gc_num_rb_per_se); + adev->gfx.config.max_texture_channel_caches = le32_to_cpu(gc_info->v1.gc_num_gl2c); + adev->gfx.config.max_gprs = le32_to_cpu(gc_info->v1.gc_num_gprs); + adev->gfx.config.max_gs_threads = le32_to_cpu(gc_info->v1.gc_num_max_gs_thds); + adev->gfx.config.gs_vgt_table_depth = le32_to_cpu(gc_info->v1.gc_gs_table_depth); + adev->gfx.config.gs_prim_buffer_depth = le32_to_cpu(gc_info->v1.gc_gsprim_buff_depth); + adev->gfx.config.double_offchip_lds_buf = le32_to_cpu(gc_info->v1.gc_double_offchip_lds_buffer); + adev->gfx.cu_info.wave_front_size = le32_to_cpu(gc_info->v1.gc_wave_size); + adev->gfx.cu_info.max_waves_per_simd = le32_to_cpu(gc_info->v1.gc_max_waves_per_simd); + adev->gfx.cu_info.max_scratch_slots_per_cu = le32_to_cpu(gc_info->v1.gc_max_scratch_slots_per_cu); + adev->gfx.cu_info.lds_size = le32_to_cpu(gc_info->v1.gc_lds_size); + adev->gfx.config.num_sc_per_sh = le32_to_cpu(gc_info->v1.gc_num_sc_per_se) / + le32_to_cpu(gc_info->v1.gc_num_sa_per_se); + adev->gfx.config.num_packer_per_sc = le32_to_cpu(gc_info->v1.gc_num_packer_per_sc); + break; + case 2: + adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->v2.gc_num_se); + adev->gfx.config.max_cu_per_sh = le32_to_cpu(gc_info->v2.gc_num_cu_per_sh); + adev->gfx.config.max_sh_per_se = le32_to_cpu(gc_info->v2.gc_num_sh_per_se); + adev->gfx.config.max_backends_per_se = le32_to_cpu(gc_info->v2.gc_num_rb_per_se); + adev->gfx.config.max_texture_channel_caches = le32_to_cpu(gc_info->v2.gc_num_tccs); + adev->gfx.config.max_gprs = le32_to_cpu(gc_info->v2.gc_num_gprs); + adev->gfx.config.max_gs_threads = le32_to_cpu(gc_info->v2.gc_num_max_gs_thds); + adev->gfx.config.gs_vgt_table_depth = le32_to_cpu(gc_info->v2.gc_gs_table_depth); + adev->gfx.config.gs_prim_buffer_depth = le32_to_cpu(gc_info->v2.gc_gsprim_buff_depth); + adev->gfx.config.double_offchip_lds_buf = le32_to_cpu(gc_info->v2.gc_double_offchip_lds_buffer); + adev->gfx.cu_info.wave_front_size = le32_to_cpu(gc_info->v2.gc_wave_size); + adev->gfx.cu_info.max_waves_per_simd = le32_to_cpu(gc_info->v2.gc_max_waves_per_simd); + adev->gfx.cu_info.max_scratch_slots_per_cu = le32_to_cpu(gc_info->v2.gc_max_scratch_slots_per_cu); + adev->gfx.cu_info.lds_size = le32_to_cpu(gc_info->v2.gc_lds_size); + adev->gfx.config.num_sc_per_sh = le32_to_cpu(gc_info->v2.gc_num_sc_per_se) / + le32_to_cpu(gc_info->v2.gc_num_sh_per_se); + adev->gfx.config.num_packer_per_sc = le32_to_cpu(gc_info->v2.gc_num_packer_per_sc); + break; + default: + dev_err(adev->dev, + "Unhandled GC info table %d.%d\n", + gc_info->v1.header.version_major, + gc_info->v1.header.version_minor); + return -EINVAL; + } + return 0; +} + +static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev) +{ + /* what IP to use for this? */ + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 3, 0): + case IP_VERSION(9, 4, 0): + case IP_VERSION(9, 4, 1): + case IP_VERSION(9, 4, 2): + amdgpu_device_ip_block_add(adev, &vega10_common_ip_block); + break; + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 3): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + amdgpu_device_ip_block_add(adev, &nv_common_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add common ip block(GC_HWIP:0x%x)\n", + adev->ip_versions[GC_HWIP][0]); + return -EINVAL; + } + return 0; +} + +static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev) +{ + /* use GC or MMHUB IP version */ + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 3, 0): + case IP_VERSION(9, 4, 0): + case IP_VERSION(9, 4, 1): + case IP_VERSION(9, 4, 2): + amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); + break; + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 3): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add gmc ip block(GC_HWIP:0x%x)\n", + adev->ip_versions[GC_HWIP][0]); + return -EINVAL; + } + return 0; +} + +static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev) +{ + switch (adev->ip_versions[OSSSYS_HWIP][0]) { + case IP_VERSION(4, 0, 0): + case IP_VERSION(4, 0, 1): + case IP_VERSION(4, 1, 0): + case IP_VERSION(4, 1, 1): + case IP_VERSION(4, 3, 0): + amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block); + break; + case IP_VERSION(4, 2, 0): + case IP_VERSION(4, 2, 1): + case IP_VERSION(4, 4, 0): + amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block); + break; + case IP_VERSION(5, 0, 0): + case IP_VERSION(5, 0, 1): + case IP_VERSION(5, 0, 2): + case IP_VERSION(5, 0, 3): + case IP_VERSION(5, 2, 0): + case IP_VERSION(5, 2, 1): + amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add ih ip block(OSSSYS_HWIP:0x%x)\n", + adev->ip_versions[OSSSYS_HWIP][0]); + return -EINVAL; + } + return 0; +} + +static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev) +{ + switch (adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(9, 0, 0): + amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block); + break; + case IP_VERSION(10, 0, 0): + case IP_VERSION(10, 0, 1): + amdgpu_device_ip_block_add(adev, &psp_v10_0_ip_block); + break; + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 4): + case IP_VERSION(11, 0, 5): + case IP_VERSION(11, 0, 9): + case IP_VERSION(11, 0, 7): + case IP_VERSION(11, 0, 11): + case IP_VERSION(11, 0, 12): + case IP_VERSION(11, 0, 13): + case IP_VERSION(11, 5, 0): + amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); + break; + case IP_VERSION(11, 0, 8): + amdgpu_device_ip_block_add(adev, &psp_v11_0_8_ip_block); + break; + case IP_VERSION(11, 0, 3): + case IP_VERSION(12, 0, 1): + amdgpu_device_ip_block_add(adev, &psp_v12_0_ip_block); + break; + case IP_VERSION(13, 0, 1): + case IP_VERSION(13, 0, 2): + case IP_VERSION(13, 0, 3): + amdgpu_device_ip_block_add(adev, &psp_v13_0_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add psp ip block(MP0_HWIP:0x%x)\n", + adev->ip_versions[MP0_HWIP][0]); + return -EINVAL; + } + return 0; +} + +static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev) +{ + switch (adev->ip_versions[MP1_HWIP][0]) { + case IP_VERSION(9, 0, 0): + case IP_VERSION(10, 0, 0): + case IP_VERSION(10, 0, 1): + case IP_VERSION(11, 0, 2): + if (adev->asic_type == CHIP_ARCTURUS) + amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); + else + amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); + break; + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 5): + case IP_VERSION(11, 0, 9): + case IP_VERSION(11, 0, 7): + case IP_VERSION(11, 0, 8): + case IP_VERSION(11, 0, 11): + case IP_VERSION(11, 0, 12): + case IP_VERSION(11, 0, 13): + case IP_VERSION(11, 5, 0): + amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); + break; + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): + amdgpu_device_ip_block_add(adev, &smu_v12_0_ip_block); + break; + case IP_VERSION(13, 0, 1): + case IP_VERSION(13, 0, 2): + case IP_VERSION(13, 0, 3): + amdgpu_device_ip_block_add(adev, &smu_v13_0_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add smu ip block(MP1_HWIP:0x%x)\n", + adev->ip_versions[MP1_HWIP][0]); + return -EINVAL; + } + return 0; +} + +static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) +{ + if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) { + amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); +#if defined(CONFIG_DRM_AMD_DC) + } else if (adev->ip_versions[DCE_HWIP][0]) { + switch (adev->ip_versions[DCE_HWIP][0]) { + case IP_VERSION(1, 0, 0): + case IP_VERSION(1, 0, 1): + case IP_VERSION(2, 0, 2): + case IP_VERSION(2, 0, 0): + case IP_VERSION(2, 0, 3): + case IP_VERSION(2, 1, 0): + case IP_VERSION(3, 0, 0): + case IP_VERSION(3, 0, 2): + case IP_VERSION(3, 0, 3): + case IP_VERSION(3, 0, 1): + case IP_VERSION(3, 1, 2): + case IP_VERSION(3, 1, 3): + amdgpu_device_ip_block_add(adev, &dm_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add dm ip block(DCE_HWIP:0x%x)\n", + adev->ip_versions[DCE_HWIP][0]); + return -EINVAL; + } + } else if (adev->ip_versions[DCI_HWIP][0]) { + switch (adev->ip_versions[DCI_HWIP][0]) { + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): + case IP_VERSION(12, 1, 0): + amdgpu_device_ip_block_add(adev, &dm_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add dm ip block(DCI_HWIP:0x%x)\n", + adev->ip_versions[DCI_HWIP][0]); + return -EINVAL; + } +#endif + } + return 0; +} - adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->gc_num_se); - adev->gfx.config.max_cu_per_sh = 2 * (le32_to_cpu(gc_info->gc_num_wgp0_per_sa) + - le32_to_cpu(gc_info->gc_num_wgp1_per_sa)); - adev->gfx.config.max_sh_per_se = le32_to_cpu(gc_info->gc_num_sa_per_se); - adev->gfx.config.max_backends_per_se = le32_to_cpu(gc_info->gc_num_rb_per_se); - adev->gfx.config.max_texture_channel_caches = le32_to_cpu(gc_info->gc_num_gl2c); - adev->gfx.config.max_gprs = le32_to_cpu(gc_info->gc_num_gprs); - adev->gfx.config.max_gs_threads = le32_to_cpu(gc_info->gc_num_max_gs_thds); - adev->gfx.config.gs_vgt_table_depth = le32_to_cpu(gc_info->gc_gs_table_depth); - adev->gfx.config.gs_prim_buffer_depth = le32_to_cpu(gc_info->gc_gsprim_buff_depth); - adev->gfx.config.double_offchip_lds_buf = le32_to_cpu(gc_info->gc_double_offchip_lds_buffer); - adev->gfx.cu_info.wave_front_size = le32_to_cpu(gc_info->gc_wave_size); - adev->gfx.cu_info.max_waves_per_simd = le32_to_cpu(gc_info->gc_max_waves_per_simd); - adev->gfx.cu_info.max_scratch_slots_per_cu = le32_to_cpu(gc_info->gc_max_scratch_slots_per_cu); - adev->gfx.cu_info.lds_size = le32_to_cpu(gc_info->gc_lds_size); - adev->gfx.config.num_sc_per_sh = le32_to_cpu(gc_info->gc_num_sc_per_se) / - le32_to_cpu(gc_info->gc_num_sa_per_se); - adev->gfx.config.num_packer_per_sc = le32_to_cpu(gc_info->gc_num_packer_per_sc); +static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev) +{ + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 3, 0): + case IP_VERSION(9, 4, 0): + case IP_VERSION(9, 4, 1): + case IP_VERSION(9, 4, 2): + amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); + break; + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): + amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add gfx ip block(GC_HWIP:0x%x)\n", + adev->ip_versions[GC_HWIP][0]); + return -EINVAL; + } + return 0; +} +static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev) +{ + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(4, 0, 0): + case IP_VERSION(4, 0, 1): + case IP_VERSION(4, 1, 0): + case IP_VERSION(4, 1, 1): + case IP_VERSION(4, 1, 2): + case IP_VERSION(4, 2, 0): + case IP_VERSION(4, 2, 2): + case IP_VERSION(4, 4, 0): + amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); + break; + case IP_VERSION(5, 0, 0): + case IP_VERSION(5, 0, 1): + case IP_VERSION(5, 0, 2): + case IP_VERSION(5, 0, 5): + amdgpu_device_ip_block_add(adev, &sdma_v5_0_ip_block); + break; + case IP_VERSION(5, 2, 0): + case IP_VERSION(5, 2, 2): + case IP_VERSION(5, 2, 4): + case IP_VERSION(5, 2, 5): + case IP_VERSION(5, 2, 3): + case IP_VERSION(5, 2, 1): + amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add sdma ip block(SDMA0_HWIP:0x%x)\n", + adev->ip_versions[SDMA0_HWIP][0]); + return -EINVAL; + } return 0; } + +static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) +{ + if (adev->ip_versions[VCE_HWIP][0]) { + switch (adev->ip_versions[UVD_HWIP][0]) { + case IP_VERSION(7, 0, 0): + case IP_VERSION(7, 2, 0): + /* UVD is not supported on vega20 SR-IOV */ + if (!(adev->asic_type == CHIP_VEGA20 && amdgpu_sriov_vf(adev))) + amdgpu_device_ip_block_add(adev, &uvd_v7_0_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add uvd v7 ip block(UVD_HWIP:0x%x)\n", + adev->ip_versions[UVD_HWIP][0]); + return -EINVAL; + } + switch (adev->ip_versions[VCE_HWIP][0]) { + case IP_VERSION(4, 0, 0): + case IP_VERSION(4, 1, 0): + /* VCE is not supported on vega20 SR-IOV */ + if (!(adev->asic_type == CHIP_VEGA20 && amdgpu_sriov_vf(adev))) + amdgpu_device_ip_block_add(adev, &vce_v4_0_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add VCE v4 ip block(VCE_HWIP:0x%x)\n", + adev->ip_versions[VCE_HWIP][0]); + return -EINVAL; + } + } else { + switch (adev->ip_versions[UVD_HWIP][0]) { + case IP_VERSION(1, 0, 0): + case IP_VERSION(1, 0, 1): + amdgpu_device_ip_block_add(adev, &vcn_v1_0_ip_block); + break; + case IP_VERSION(2, 0, 0): + case IP_VERSION(2, 0, 2): + case IP_VERSION(2, 2, 0): + amdgpu_device_ip_block_add(adev, &vcn_v2_0_ip_block); + if (!amdgpu_sriov_vf(adev)) + amdgpu_device_ip_block_add(adev, &jpeg_v2_0_ip_block); + break; + case IP_VERSION(2, 0, 3): + break; + case IP_VERSION(2, 5, 0): + amdgpu_device_ip_block_add(adev, &vcn_v2_5_ip_block); + amdgpu_device_ip_block_add(adev, &jpeg_v2_5_ip_block); + break; + case IP_VERSION(2, 6, 0): + amdgpu_device_ip_block_add(adev, &vcn_v2_6_ip_block); + amdgpu_device_ip_block_add(adev, &jpeg_v2_6_ip_block); + break; + case IP_VERSION(3, 0, 0): + case IP_VERSION(3, 0, 16): + case IP_VERSION(3, 1, 1): + case IP_VERSION(3, 0, 2): + case IP_VERSION(3, 0, 192): + amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); + if (!amdgpu_sriov_vf(adev)) + amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block); + break; + case IP_VERSION(3, 0, 33): + amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); + break; + default: + dev_err(adev->dev, + "Failed to add vcn/jpeg ip block(UVD_HWIP:0x%x)\n", + adev->ip_versions[UVD_HWIP][0]); + return -EINVAL; + } + } + return 0; +} + +static int amdgpu_discovery_set_mes_ip_blocks(struct amdgpu_device *adev) +{ + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 3): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + amdgpu_device_ip_block_add(adev, &mes_v10_1_ip_block); + break; + default: + break; + } + return 0; +} + +int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) +{ + int r; + + switch (adev->asic_type) { + case CHIP_VEGA10: + vega10_reg_base_init(adev); + adev->sdma.num_instances = 2; + adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 0, 0); + adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 0, 0); + adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 0); + adev->ip_versions[HDP_HWIP][0] = IP_VERSION(4, 0, 0); + adev->ip_versions[SDMA0_HWIP][0] = IP_VERSION(4, 0, 0); + adev->ip_versions[SDMA1_HWIP][0] = IP_VERSION(4, 0, 0); + adev->ip_versions[DF_HWIP][0] = IP_VERSION(2, 1, 0); + adev->ip_versions[NBIO_HWIP][0] = IP_VERSION(6, 1, 0); + adev->ip_versions[UMC_HWIP][0] = IP_VERSION(6, 0, 0); + adev->ip_versions[MP0_HWIP][0] = IP_VERSION(9, 0, 0); + adev->ip_versions[MP1_HWIP][0] = IP_VERSION(9, 0, 0); + adev->ip_versions[THM_HWIP][0] = IP_VERSION(9, 0, 0); + adev->ip_versions[SMUIO_HWIP][0] = IP_VERSION(9, 0, 0); + adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 0, 1); + adev->ip_versions[UVD_HWIP][0] = IP_VERSION(7, 0, 0); + adev->ip_versions[VCE_HWIP][0] = IP_VERSION(4, 0, 0); + adev->ip_versions[DCI_HWIP][0] = IP_VERSION(12, 0, 0); + break; + case CHIP_VEGA12: + vega10_reg_base_init(adev); + adev->sdma.num_instances = 2; + adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 3, 0); + adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 3, 0); + adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 1); + adev->ip_versions[HDP_HWIP][0] = IP_VERSION(4, 0, 1); + adev->ip_versions[SDMA0_HWIP][0] = IP_VERSION(4, 0, 1); + adev->ip_versions[SDMA1_HWIP][0] = IP_VERSION(4, 0, 1); + adev->ip_versions[DF_HWIP][0] = IP_VERSION(2, 5, 0); + adev->ip_versions[NBIO_HWIP][0] = IP_VERSION(6, 2, 0); + adev->ip_versions[UMC_HWIP][0] = IP_VERSION(6, 1, 0); + adev->ip_versions[MP0_HWIP][0] = IP_VERSION(9, 0, 0); + adev->ip_versions[MP1_HWIP][0] = IP_VERSION(9, 0, 0); + adev->ip_versions[THM_HWIP][0] = IP_VERSION(9, 0, 0); + adev->ip_versions[SMUIO_HWIP][0] = IP_VERSION(9, 0, 1); + adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 2, 1); + adev->ip_versions[UVD_HWIP][0] = IP_VERSION(7, 0, 0); + adev->ip_versions[VCE_HWIP][0] = IP_VERSION(4, 0, 0); + adev->ip_versions[DCI_HWIP][0] = IP_VERSION(12, 0, 1); + break; + case CHIP_RAVEN: + vega10_reg_base_init(adev); + adev->sdma.num_instances = 1; + adev->vcn.num_vcn_inst = 1; + if (adev->apu_flags & AMD_APU_IS_RAVEN2) { + adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 2, 0); + adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 2, 0); + adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 1, 1); + adev->ip_versions[HDP_HWIP][0] = IP_VERSION(4, 1, 1); + adev->ip_versions[SDMA0_HWIP][0] = IP_VERSION(4, 1, 1); + adev->ip_versions[DF_HWIP][0] = IP_VERSION(2, 1, 1); + adev->ip_versions[NBIO_HWIP][0] = IP_VERSION(7, 0, 1); + adev->ip_versions[UMC_HWIP][0] = IP_VERSION(7, 5, 0); + adev->ip_versions[MP0_HWIP][0] = IP_VERSION(10, 0, 1); + adev->ip_versions[MP1_HWIP][0] = IP_VERSION(10, 0, 1); + adev->ip_versions[THM_HWIP][0] = IP_VERSION(10, 1, 0); + adev->ip_versions[SMUIO_HWIP][0] = IP_VERSION(10, 0, 1); + adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 2, 2); + adev->ip_versions[UVD_HWIP][0] = IP_VERSION(1, 0, 1); + adev->ip_versions[DCE_HWIP][0] = IP_VERSION(1, 0, 1); + } else { + adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 1, 0); + adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 1, 0); + adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 1, 0); + adev->ip_versions[HDP_HWIP][0] = IP_VERSION(4, 1, 0); + adev->ip_versions[SDMA0_HWIP][0] = IP_VERSION(4, 1, 0); + adev->ip_versions[DF_HWIP][0] = IP_VERSION(2, 1, 0); + adev->ip_versions[NBIO_HWIP][0] = IP_VERSION(7, 0, 0); + adev->ip_versions[UMC_HWIP][0] = IP_VERSION(7, 0, 0); + adev->ip_versions[MP0_HWIP][0] = IP_VERSION(10, 0, 0); + adev->ip_versions[MP1_HWIP][0] = IP_VERSION(10, 0, 0); + adev->ip_versions[THM_HWIP][0] = IP_VERSION(10, 0, 0); + adev->ip_versions[SMUIO_HWIP][0] = IP_VERSION(10, 0, 0); + adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 1, 0); + adev->ip_versions[UVD_HWIP][0] = IP_VERSION(1, 0, 0); + adev->ip_versions[DCE_HWIP][0] = IP_VERSION(1, 0, 0); + } + break; + case CHIP_VEGA20: + vega20_reg_base_init(adev); + adev->sdma.num_instances = 2; + adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 0); + adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 0); + adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 0); + adev->ip_versions[HDP_HWIP][0] = IP_VERSION(4, 2, 0); + adev->ip_versions[SDMA0_HWIP][0] = IP_VERSION(4, 2, 0); + adev->ip_versions[SDMA1_HWIP][0] = IP_VERSION(4, 2, 0); + adev->ip_versions[DF_HWIP][0] = IP_VERSION(3, 6, 0); + adev->ip_versions[NBIO_HWIP][0] = IP_VERSION(7, 4, 0); + adev->ip_versions[UMC_HWIP][0] = IP_VERSION(6, 1, 1); + adev->ip_versions[MP0_HWIP][0] = IP_VERSION(11, 0, 2); + adev->ip_versions[MP1_HWIP][0] = IP_VERSION(11, 0, 2); + adev->ip_versions[THM_HWIP][0] = IP_VERSION(11, 0, 2); + adev->ip_versions[SMUIO_HWIP][0] = IP_VERSION(11, 0, 2); + adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 4, 0); + adev->ip_versions[UVD_HWIP][0] = IP_VERSION(7, 2, 0); + adev->ip_versions[UVD_HWIP][1] = IP_VERSION(7, 2, 0); + adev->ip_versions[VCE_HWIP][0] = IP_VERSION(4, 1, 0); + adev->ip_versions[DCI_HWIP][0] = IP_VERSION(12, 1, 0); + break; + case CHIP_ARCTURUS: + arct_reg_base_init(adev); + adev->sdma.num_instances = 8; + adev->vcn.num_vcn_inst = 2; + adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 1); + adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 1); + adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 1); + adev->ip_versions[HDP_HWIP][0] = IP_VERSION(4, 2, 1); + adev->ip_versions[SDMA0_HWIP][0] = IP_VERSION(4, 2, 2); + adev->ip_versions[SDMA1_HWIP][0] = IP_VERSION(4, 2, 2); + adev->ip_versions[SDMA1_HWIP][1] = IP_VERSION(4, 2, 2); + adev->ip_versions[SDMA1_HWIP][2] = IP_VERSION(4, 2, 2); + adev->ip_versions[SDMA1_HWIP][3] = IP_VERSION(4, 2, 2); + adev->ip_versions[SDMA1_HWIP][4] = IP_VERSION(4, 2, 2); + adev->ip_versions[SDMA1_HWIP][5] = IP_VERSION(4, 2, 2); + adev->ip_versions[SDMA1_HWIP][6] = IP_VERSION(4, 2, 2); + adev->ip_versions[DF_HWIP][0] = IP_VERSION(3, 6, 1); + adev->ip_versions[NBIO_HWIP][0] = IP_VERSION(7, 4, 1); + adev->ip_versions[UMC_HWIP][0] = IP_VERSION(6, 1, 2); + adev->ip_versions[MP0_HWIP][0] = IP_VERSION(11, 0, 4); + adev->ip_versions[MP1_HWIP][0] = IP_VERSION(11, 0, 2); + adev->ip_versions[THM_HWIP][0] = IP_VERSION(11, 0, 3); + adev->ip_versions[SMUIO_HWIP][0] = IP_VERSION(11, 0, 3); + adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 4, 1); + adev->ip_versions[UVD_HWIP][0] = IP_VERSION(2, 5, 0); + adev->ip_versions[UVD_HWIP][1] = IP_VERSION(2, 5, 0); + break; + case CHIP_ALDEBARAN: + aldebaran_reg_base_init(adev); + adev->sdma.num_instances = 5; + adev->vcn.num_vcn_inst = 2; + adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 2); + adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 2); + adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 4, 0); + adev->ip_versions[HDP_HWIP][0] = IP_VERSION(4, 4, 0); + adev->ip_versions[SDMA0_HWIP][0] = IP_VERSION(4, 4, 0); + adev->ip_versions[SDMA0_HWIP][1] = IP_VERSION(4, 4, 0); + adev->ip_versions[SDMA0_HWIP][2] = IP_VERSION(4, 4, 0); + adev->ip_versions[SDMA0_HWIP][3] = IP_VERSION(4, 4, 0); + adev->ip_versions[SDMA0_HWIP][4] = IP_VERSION(4, 4, 0); + adev->ip_versions[DF_HWIP][0] = IP_VERSION(3, 6, 2); + adev->ip_versions[NBIO_HWIP][0] = IP_VERSION(7, 4, 4); + adev->ip_versions[UMC_HWIP][0] = IP_VERSION(6, 7, 0); + adev->ip_versions[MP0_HWIP][0] = IP_VERSION(13, 0, 2); + adev->ip_versions[MP1_HWIP][0] = IP_VERSION(13, 0, 2); + adev->ip_versions[THM_HWIP][0] = IP_VERSION(13, 0, 2); + adev->ip_versions[SMUIO_HWIP][0] = IP_VERSION(13, 0, 2); + adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 4, 2); + adev->ip_versions[UVD_HWIP][0] = IP_VERSION(2, 6, 0); + adev->ip_versions[UVD_HWIP][1] = IP_VERSION(2, 6, 0); + adev->ip_versions[XGMI_HWIP][0] = IP_VERSION(6, 1, 0); + break; + default: + r = amdgpu_discovery_reg_base_init(adev); + if (r) + return -EINVAL; + + amdgpu_discovery_harvest_ip(adev); + + if (!adev->mman.discovery_bin) { + DRM_ERROR("ip discovery uninitialized\n"); + return -EINVAL; + } + break; + } + + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 4, 0): + case IP_VERSION(9, 4, 1): + case IP_VERSION(9, 4, 2): + adev->family = AMDGPU_FAMILY_AI; + break; + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 3, 0): + adev->family = AMDGPU_FAMILY_RV; + break; + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + adev->family = AMDGPU_FAMILY_NV; + break; + case IP_VERSION(10, 3, 1): + adev->family = AMDGPU_FAMILY_VGH; + break; + case IP_VERSION(10, 3, 3): + adev->family = AMDGPU_FAMILY_YC; + break; + default: + return -EINVAL; + } + + if (adev->ip_versions[XGMI_HWIP][0] == IP_VERSION(4, 8, 0)) + adev->gmc.xgmi.supported = true; + + /* set NBIO version */ + switch (adev->ip_versions[NBIO_HWIP][0]) { + case IP_VERSION(6, 1, 0): + case IP_VERSION(6, 2, 0): + adev->nbio.funcs = &nbio_v6_1_funcs; + adev->nbio.hdp_flush_reg = &nbio_v6_1_hdp_flush_reg; + break; + case IP_VERSION(7, 0, 0): + case IP_VERSION(7, 0, 1): + case IP_VERSION(2, 5, 0): + adev->nbio.funcs = &nbio_v7_0_funcs; + adev->nbio.hdp_flush_reg = &nbio_v7_0_hdp_flush_reg; + break; + case IP_VERSION(7, 4, 0): + case IP_VERSION(7, 4, 1): + adev->nbio.funcs = &nbio_v7_4_funcs; + adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg; + break; + case IP_VERSION(7, 4, 4): + adev->nbio.funcs = &nbio_v7_4_funcs; + adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg_ald; + break; + case IP_VERSION(7, 2, 0): + case IP_VERSION(7, 2, 1): + case IP_VERSION(7, 5, 0): + adev->nbio.funcs = &nbio_v7_2_funcs; + adev->nbio.hdp_flush_reg = &nbio_v7_2_hdp_flush_reg; + break; + case IP_VERSION(2, 1, 1): + case IP_VERSION(2, 3, 0): + case IP_VERSION(2, 3, 1): + case IP_VERSION(2, 3, 2): + adev->nbio.funcs = &nbio_v2_3_funcs; + adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg; + break; + case IP_VERSION(3, 3, 0): + case IP_VERSION(3, 3, 1): + case IP_VERSION(3, 3, 2): + case IP_VERSION(3, 3, 3): + adev->nbio.funcs = &nbio_v2_3_funcs; + adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg_sc; + break; + default: + break; + } + + switch (adev->ip_versions[HDP_HWIP][0]) { + case IP_VERSION(4, 0, 0): + case IP_VERSION(4, 0, 1): + case IP_VERSION(4, 1, 0): + case IP_VERSION(4, 1, 1): + case IP_VERSION(4, 1, 2): + case IP_VERSION(4, 2, 0): + case IP_VERSION(4, 2, 1): + case IP_VERSION(4, 4, 0): + adev->hdp.funcs = &hdp_v4_0_funcs; + break; + case IP_VERSION(5, 0, 0): + case IP_VERSION(5, 0, 1): + case IP_VERSION(5, 0, 2): + case IP_VERSION(5, 0, 3): + case IP_VERSION(5, 0, 4): + case IP_VERSION(5, 2, 0): + adev->hdp.funcs = &hdp_v5_0_funcs; + break; + default: + break; + } + + switch (adev->ip_versions[DF_HWIP][0]) { + case IP_VERSION(3, 6, 0): + case IP_VERSION(3, 6, 1): + case IP_VERSION(3, 6, 2): + adev->df.funcs = &df_v3_6_funcs; + break; + case IP_VERSION(2, 1, 0): + case IP_VERSION(2, 1, 1): + case IP_VERSION(2, 5, 0): + case IP_VERSION(3, 5, 1): + case IP_VERSION(3, 5, 2): + adev->df.funcs = &df_v1_7_funcs; + break; + default: + break; + } + + switch (adev->ip_versions[SMUIO_HWIP][0]) { + case IP_VERSION(9, 0, 0): + case IP_VERSION(9, 0, 1): + case IP_VERSION(10, 0, 0): + case IP_VERSION(10, 0, 1): + case IP_VERSION(10, 0, 2): + adev->smuio.funcs = &smuio_v9_0_funcs; + break; + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 3): + case IP_VERSION(11, 0, 4): + case IP_VERSION(11, 0, 7): + case IP_VERSION(11, 0, 8): + adev->smuio.funcs = &smuio_v11_0_funcs; + break; + case IP_VERSION(11, 0, 6): + case IP_VERSION(11, 0, 10): + case IP_VERSION(11, 0, 11): + case IP_VERSION(11, 5, 0): + case IP_VERSION(13, 0, 1): + adev->smuio.funcs = &smuio_v11_0_6_funcs; + break; + case IP_VERSION(13, 0, 2): + adev->smuio.funcs = &smuio_v13_0_funcs; + break; + default: + break; + } + + r = amdgpu_discovery_set_common_ip_blocks(adev); + if (r) + return r; + + r = amdgpu_discovery_set_gmc_ip_blocks(adev); + if (r) + return r; + + /* For SR-IOV, PSP needs to be initialized before IH */ + if (amdgpu_sriov_vf(adev)) { + r = amdgpu_discovery_set_psp_ip_blocks(adev); + if (r) + return r; + r = amdgpu_discovery_set_ih_ip_blocks(adev); + if (r) + return r; + } else { + r = amdgpu_discovery_set_ih_ip_blocks(adev); + if (r) + return r; + + if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) { + r = amdgpu_discovery_set_psp_ip_blocks(adev); + if (r) + return r; + } + } + + if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) { + r = amdgpu_discovery_set_smu_ip_blocks(adev); + if (r) + return r; + } + + r = amdgpu_discovery_set_display_ip_blocks(adev); + if (r) + return r; + + r = amdgpu_discovery_set_gc_ip_blocks(adev); + if (r) + return r; + + r = amdgpu_discovery_set_sdma_ip_blocks(adev); + if (r) + return r; + + if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT && + !amdgpu_sriov_vf(adev)) { + r = amdgpu_discovery_set_smu_ip_blocks(adev); + if (r) + return r; + } + + r = amdgpu_discovery_set_mm_ip_blocks(adev); + if (r) + return r; + + if (adev->enable_mes) { + r = amdgpu_discovery_set_mes_ip_blocks(adev); + if (r) + return r; + } + + return 0; +} + diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h index 48e6b88cfdfe..14537cec19db 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h @@ -33,8 +33,7 @@ void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev); int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int number_instance, int *major, int *minor, int *revision); -int amdgpu_discovery_get_vcn_version(struct amdgpu_device *adev, int vcn_instance, - int *major, int *minor, int *revision); int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev); +int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev); #endif /* __AMDGPU_DISCOVERY__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index dc50c05f23fc..82011e75ed85 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -83,9 +83,6 @@ static void amdgpu_display_flip_work_func(struct work_struct *__work) unsigned i; int vpos, hpos; - if (amdgpu_display_flip_handle_fence(work, &work->excl)) - return; - for (i = 0; i < work->shared_count; ++i) if (amdgpu_display_flip_handle_fence(work, &work->shared[i])) return; @@ -203,7 +200,7 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc, goto unpin; } - r = dma_resv_get_fences(new_abo->tbo.base.resv, &work->excl, + r = dma_resv_get_fences(new_abo->tbo.base.resv, NULL, &work->shared_count, &work->shared); if (unlikely(r != 0)) { DRM_ERROR("failed to get fences for buffer\n"); @@ -253,7 +250,6 @@ unreserve: cleanup: amdgpu_bo_unref(&work->old_abo); - dma_fence_put(work->excl); for (i = 0; i < work->shared_count; ++i) dma_fence_put(work->shared[i]); kfree(work->shared); @@ -1364,7 +1360,7 @@ bool amdgpu_display_crtc_scaling_mode_fixup(struct drm_crtc *crtc, if ((!(mode->flags & DRM_MODE_FLAG_INTERLACE)) && ((amdgpu_encoder->underscan_type == UNDERSCAN_ON) || ((amdgpu_encoder->underscan_type == UNDERSCAN_AUTO) && - drm_detect_hdmi_monitor(amdgpu_connector_edid(connector)) && + connector->display_info.is_hdmi && amdgpu_display_is_hdtv_mode(mode)))) { if (amdgpu_encoder->underscan_hborder != 0) amdgpu_crtc->h_border = amdgpu_encoder->underscan_hborder; @@ -1603,13 +1599,10 @@ int amdgpu_display_suspend_helper(struct amdgpu_device *adev) continue; } robj = gem_to_amdgpu_bo(fb->obj[0]); - /* don't unpin kernel fb objects */ - if (!amdgpu_fbdev_robj_is_fb(adev, robj)) { - r = amdgpu_bo_reserve(robj, true); - if (r == 0) { - amdgpu_bo_unpin(robj); - amdgpu_bo_unreserve(robj); - } + r = amdgpu_bo_reserve(robj, true); + if (r == 0) { + amdgpu_bo_unpin(robj); + amdgpu_bo_unreserve(robj); } } return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c index ae6ab93c868b..579adfafe4d0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c @@ -61,9 +61,6 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf, if (pci_p2pdma_distance_many(adev->pdev, &attach->dev, 1, true) < 0) attach->peer2peer = false; - if (attach->dev->driver == adev->dev->driver) - return 0; - r = pm_runtime_get_sync(adev_to_drm(adev)->dev); if (r < 0) goto out; @@ -384,7 +381,7 @@ amdgpu_dma_buf_move_notify(struct dma_buf_attachment *attach) struct amdgpu_vm_bo_base *bo_base; int r; - if (bo->tbo.resource->mem_type == TTM_PL_SYSTEM) + if (!bo->tbo.resource || bo->tbo.resource->mem_type == TTM_PL_SYSTEM) return; r = ttm_bo_validate(&bo->tbo, &placement, &ctx); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index f18240f87387..b63ed1ddf713 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -31,13 +31,14 @@ #include "amdgpu_drv.h" #include <drm/drm_pciids.h> -#include <linux/console.h> #include <linux/module.h> #include <linux/pm_runtime.h> #include <linux/vga_switcheroo.h> #include <drm/drm_probe_helper.h> #include <linux/mmu_notifier.h> #include <linux/suspend.h> +#include <linux/cc_platform.h> +#include <linux/fb.h> #include "amdgpu.h" #include "amdgpu_irq.h" @@ -96,9 +97,11 @@ * - 3.40.0 - Add AMDGPU_IDS_FLAGS_TMZ * - 3.41.0 - Add video codec query * - 3.42.0 - Add 16bpc fixed point display support + * - 3.43.0 - Add device hot plug/unplug support + * - 3.44.0 - DCN3 supports DCC independent block settings: !64B && 128B, 64B && 128B */ #define KMS_DRIVER_MAJOR 3 -#define KMS_DRIVER_MINOR 42 +#define KMS_DRIVER_MINOR 44 #define KMS_DRIVER_PATCHLEVEL 0 int amdgpu_vram_limit; @@ -311,9 +314,12 @@ module_param_named(dpm, amdgpu_dpm, int, 0444); /** * DOC: fw_load_type (int) - * Set different firmware loading type for debugging (0 = direct, 1 = SMU, 2 = PSP). The default is -1 (auto). + * Set different firmware loading type for debugging, if supported. + * Set to 0 to force direct loading if supported by the ASIC. Set + * to -1 to select the default loading mode for the ASIC, as defined + * by the driver. The default is -1 (auto). */ -MODULE_PARM_DESC(fw_load_type, "firmware loading type (0 = direct, 1 = SMU, 2 = PSP, -1 = auto)"); +MODULE_PARM_DESC(fw_load_type, "firmware loading type (0 = force direct if supported, -1 = auto)"); module_param_named(fw_load_type, amdgpu_fw_load_type, int, 0444); /** @@ -325,10 +331,11 @@ module_param_named(aspm, amdgpu_aspm, int, 0444); /** * DOC: runpm (int) - * Override for runtime power management control for dGPUs in PX/HG laptops. The amdgpu driver can dynamically power down - * the dGPU on PX/HG laptops when it is idle. The default is -1 (auto enable). Setting the value to 0 disables this functionality. + * Override for runtime power management control for dGPUs. The amdgpu driver can dynamically power down + * the dGPUs when they are idle if supported. The default is -1 (auto enable). + * Setting the value to 0 disables this functionality. */ -MODULE_PARM_DESC(runpm, "PX runtime pm (2 = force enable with BAMACO, 1 = force enable with BACO, 0 = disable, -1 = PX only default)"); +MODULE_PARM_DESC(runpm, "PX runtime pm (2 = force enable with BAMACO, 1 = force enable with BACO, 0 = disable, -1 = auto)"); module_param_named(runpm, amdgpu_runtime_pm, int, 0444); /** @@ -627,7 +634,7 @@ module_param_named(mcbp, amdgpu_mcbp, int, 0444); /** * DOC: discovery (int) * Allow driver to discover hardware IP information from IP Discovery table at the top of VRAM. - * (-1 = auto (default), 0 = disabled, 1 = enabled) + * (-1 = auto (default), 0 = disabled, 1 = enabled, 2 = use ip_discovery table from file) */ MODULE_PARM_DESC(discovery, "Allow driver to discover hardware IPs from IP Discovery table at the top of VRAM"); @@ -875,7 +882,7 @@ module_param_named(reset_method, amdgpu_reset_method, int, 0444); * 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 = auto(default value), 0 = disable bad page retirement)"); +MODULE_PARM_DESC(bad_page_threshold, "Bad page threshold(-1 = auto(default value), 0 = disable bad page retirement, -2 = ignore bad page 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+)"); @@ -890,6 +897,636 @@ MODULE_PARM_DESC(smu_pptable_id, "specify pptable id to be used (-1 = auto(default) value, 0 = use pptable from vbios, > 0 = soft pptable id)"); module_param_named(smu_pptable_id, amdgpu_smu_pptable_id, int, 0444); +/* These devices are not supported by amdgpu. + * They are supported by the mach64, r128, radeon drivers + */ +static const u16 amdgpu_unsupported_pciidlist[] = { + /* mach64 */ + 0x4354, + 0x4358, + 0x4554, + 0x4742, + 0x4744, + 0x4749, + 0x474C, + 0x474D, + 0x474E, + 0x474F, + 0x4750, + 0x4751, + 0x4752, + 0x4753, + 0x4754, + 0x4755, + 0x4756, + 0x4757, + 0x4758, + 0x4759, + 0x475A, + 0x4C42, + 0x4C44, + 0x4C47, + 0x4C49, + 0x4C4D, + 0x4C4E, + 0x4C50, + 0x4C51, + 0x4C52, + 0x4C53, + 0x5654, + 0x5655, + 0x5656, + /* r128 */ + 0x4c45, + 0x4c46, + 0x4d46, + 0x4d4c, + 0x5041, + 0x5042, + 0x5043, + 0x5044, + 0x5045, + 0x5046, + 0x5047, + 0x5048, + 0x5049, + 0x504A, + 0x504B, + 0x504C, + 0x504D, + 0x504E, + 0x504F, + 0x5050, + 0x5051, + 0x5052, + 0x5053, + 0x5054, + 0x5055, + 0x5056, + 0x5057, + 0x5058, + 0x5245, + 0x5246, + 0x5247, + 0x524b, + 0x524c, + 0x534d, + 0x5446, + 0x544C, + 0x5452, + /* radeon */ + 0x3150, + 0x3151, + 0x3152, + 0x3154, + 0x3155, + 0x3E50, + 0x3E54, + 0x4136, + 0x4137, + 0x4144, + 0x4145, + 0x4146, + 0x4147, + 0x4148, + 0x4149, + 0x414A, + 0x414B, + 0x4150, + 0x4151, + 0x4152, + 0x4153, + 0x4154, + 0x4155, + 0x4156, + 0x4237, + 0x4242, + 0x4336, + 0x4337, + 0x4437, + 0x4966, + 0x4967, + 0x4A48, + 0x4A49, + 0x4A4A, + 0x4A4B, + 0x4A4C, + 0x4A4D, + 0x4A4E, + 0x4A4F, + 0x4A50, + 0x4A54, + 0x4B48, + 0x4B49, + 0x4B4A, + 0x4B4B, + 0x4B4C, + 0x4C57, + 0x4C58, + 0x4C59, + 0x4C5A, + 0x4C64, + 0x4C66, + 0x4C67, + 0x4E44, + 0x4E45, + 0x4E46, + 0x4E47, + 0x4E48, + 0x4E49, + 0x4E4A, + 0x4E4B, + 0x4E50, + 0x4E51, + 0x4E52, + 0x4E53, + 0x4E54, + 0x4E56, + 0x5144, + 0x5145, + 0x5146, + 0x5147, + 0x5148, + 0x514C, + 0x514D, + 0x5157, + 0x5158, + 0x5159, + 0x515A, + 0x515E, + 0x5460, + 0x5462, + 0x5464, + 0x5548, + 0x5549, + 0x554A, + 0x554B, + 0x554C, + 0x554D, + 0x554E, + 0x554F, + 0x5550, + 0x5551, + 0x5552, + 0x5554, + 0x564A, + 0x564B, + 0x564F, + 0x5652, + 0x5653, + 0x5657, + 0x5834, + 0x5835, + 0x5954, + 0x5955, + 0x5974, + 0x5975, + 0x5960, + 0x5961, + 0x5962, + 0x5964, + 0x5965, + 0x5969, + 0x5a41, + 0x5a42, + 0x5a61, + 0x5a62, + 0x5b60, + 0x5b62, + 0x5b63, + 0x5b64, + 0x5b65, + 0x5c61, + 0x5c63, + 0x5d48, + 0x5d49, + 0x5d4a, + 0x5d4c, + 0x5d4d, + 0x5d4e, + 0x5d4f, + 0x5d50, + 0x5d52, + 0x5d57, + 0x5e48, + 0x5e4a, + 0x5e4b, + 0x5e4c, + 0x5e4d, + 0x5e4f, + 0x6700, + 0x6701, + 0x6702, + 0x6703, + 0x6704, + 0x6705, + 0x6706, + 0x6707, + 0x6708, + 0x6709, + 0x6718, + 0x6719, + 0x671c, + 0x671d, + 0x671f, + 0x6720, + 0x6721, + 0x6722, + 0x6723, + 0x6724, + 0x6725, + 0x6726, + 0x6727, + 0x6728, + 0x6729, + 0x6738, + 0x6739, + 0x673e, + 0x6740, + 0x6741, + 0x6742, + 0x6743, + 0x6744, + 0x6745, + 0x6746, + 0x6747, + 0x6748, + 0x6749, + 0x674A, + 0x6750, + 0x6751, + 0x6758, + 0x6759, + 0x675B, + 0x675D, + 0x675F, + 0x6760, + 0x6761, + 0x6762, + 0x6763, + 0x6764, + 0x6765, + 0x6766, + 0x6767, + 0x6768, + 0x6770, + 0x6771, + 0x6772, + 0x6778, + 0x6779, + 0x677B, + 0x6840, + 0x6841, + 0x6842, + 0x6843, + 0x6849, + 0x684C, + 0x6850, + 0x6858, + 0x6859, + 0x6880, + 0x6888, + 0x6889, + 0x688A, + 0x688C, + 0x688D, + 0x6898, + 0x6899, + 0x689b, + 0x689c, + 0x689d, + 0x689e, + 0x68a0, + 0x68a1, + 0x68a8, + 0x68a9, + 0x68b0, + 0x68b8, + 0x68b9, + 0x68ba, + 0x68be, + 0x68bf, + 0x68c0, + 0x68c1, + 0x68c7, + 0x68c8, + 0x68c9, + 0x68d8, + 0x68d9, + 0x68da, + 0x68de, + 0x68e0, + 0x68e1, + 0x68e4, + 0x68e5, + 0x68e8, + 0x68e9, + 0x68f1, + 0x68f2, + 0x68f8, + 0x68f9, + 0x68fa, + 0x68fe, + 0x7100, + 0x7101, + 0x7102, + 0x7103, + 0x7104, + 0x7105, + 0x7106, + 0x7108, + 0x7109, + 0x710A, + 0x710B, + 0x710C, + 0x710E, + 0x710F, + 0x7140, + 0x7141, + 0x7142, + 0x7143, + 0x7144, + 0x7145, + 0x7146, + 0x7147, + 0x7149, + 0x714A, + 0x714B, + 0x714C, + 0x714D, + 0x714E, + 0x714F, + 0x7151, + 0x7152, + 0x7153, + 0x715E, + 0x715F, + 0x7180, + 0x7181, + 0x7183, + 0x7186, + 0x7187, + 0x7188, + 0x718A, + 0x718B, + 0x718C, + 0x718D, + 0x718F, + 0x7193, + 0x7196, + 0x719B, + 0x719F, + 0x71C0, + 0x71C1, + 0x71C2, + 0x71C3, + 0x71C4, + 0x71C5, + 0x71C6, + 0x71C7, + 0x71CD, + 0x71CE, + 0x71D2, + 0x71D4, + 0x71D5, + 0x71D6, + 0x71DA, + 0x71DE, + 0x7200, + 0x7210, + 0x7211, + 0x7240, + 0x7243, + 0x7244, + 0x7245, + 0x7246, + 0x7247, + 0x7248, + 0x7249, + 0x724A, + 0x724B, + 0x724C, + 0x724D, + 0x724E, + 0x724F, + 0x7280, + 0x7281, + 0x7283, + 0x7284, + 0x7287, + 0x7288, + 0x7289, + 0x728B, + 0x728C, + 0x7290, + 0x7291, + 0x7293, + 0x7297, + 0x7834, + 0x7835, + 0x791e, + 0x791f, + 0x793f, + 0x7941, + 0x7942, + 0x796c, + 0x796d, + 0x796e, + 0x796f, + 0x9400, + 0x9401, + 0x9402, + 0x9403, + 0x9405, + 0x940A, + 0x940B, + 0x940F, + 0x94A0, + 0x94A1, + 0x94A3, + 0x94B1, + 0x94B3, + 0x94B4, + 0x94B5, + 0x94B9, + 0x9440, + 0x9441, + 0x9442, + 0x9443, + 0x9444, + 0x9446, + 0x944A, + 0x944B, + 0x944C, + 0x944E, + 0x9450, + 0x9452, + 0x9456, + 0x945A, + 0x945B, + 0x945E, + 0x9460, + 0x9462, + 0x946A, + 0x946B, + 0x947A, + 0x947B, + 0x9480, + 0x9487, + 0x9488, + 0x9489, + 0x948A, + 0x948F, + 0x9490, + 0x9491, + 0x9495, + 0x9498, + 0x949C, + 0x949E, + 0x949F, + 0x94C0, + 0x94C1, + 0x94C3, + 0x94C4, + 0x94C5, + 0x94C6, + 0x94C7, + 0x94C8, + 0x94C9, + 0x94CB, + 0x94CC, + 0x94CD, + 0x9500, + 0x9501, + 0x9504, + 0x9505, + 0x9506, + 0x9507, + 0x9508, + 0x9509, + 0x950F, + 0x9511, + 0x9515, + 0x9517, + 0x9519, + 0x9540, + 0x9541, + 0x9542, + 0x954E, + 0x954F, + 0x9552, + 0x9553, + 0x9555, + 0x9557, + 0x955f, + 0x9580, + 0x9581, + 0x9583, + 0x9586, + 0x9587, + 0x9588, + 0x9589, + 0x958A, + 0x958B, + 0x958C, + 0x958D, + 0x958E, + 0x958F, + 0x9590, + 0x9591, + 0x9593, + 0x9595, + 0x9596, + 0x9597, + 0x9598, + 0x9599, + 0x959B, + 0x95C0, + 0x95C2, + 0x95C4, + 0x95C5, + 0x95C6, + 0x95C7, + 0x95C9, + 0x95CC, + 0x95CD, + 0x95CE, + 0x95CF, + 0x9610, + 0x9611, + 0x9612, + 0x9613, + 0x9614, + 0x9615, + 0x9616, + 0x9640, + 0x9641, + 0x9642, + 0x9643, + 0x9644, + 0x9645, + 0x9647, + 0x9648, + 0x9649, + 0x964a, + 0x964b, + 0x964c, + 0x964e, + 0x964f, + 0x9710, + 0x9711, + 0x9712, + 0x9713, + 0x9714, + 0x9715, + 0x9802, + 0x9803, + 0x9804, + 0x9805, + 0x9806, + 0x9807, + 0x9808, + 0x9809, + 0x980A, + 0x9900, + 0x9901, + 0x9903, + 0x9904, + 0x9905, + 0x9906, + 0x9907, + 0x9908, + 0x9909, + 0x990A, + 0x990B, + 0x990C, + 0x990D, + 0x990E, + 0x990F, + 0x9910, + 0x9913, + 0x9917, + 0x9918, + 0x9919, + 0x9990, + 0x9991, + 0x9992, + 0x9993, + 0x9994, + 0x9995, + 0x9996, + 0x9997, + 0x9998, + 0x9999, + 0x999A, + 0x999B, + 0x999C, + 0x999D, + 0x99A0, + 0x99A2, + 0x99A4, +}; + static const struct pci_device_id pciidlist[] = { #ifdef CONFIG_DRM_AMDGPU_SI {0x1002, 0x6780, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI}, @@ -1239,6 +1876,16 @@ static const struct pci_device_id pciidlist[] = { {0x1002, 0x7423, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BEIGE_GOBY}, {0x1002, 0x743F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BEIGE_GOBY}, + { PCI_DEVICE(0x1002, PCI_ANY_ID), + .class = PCI_CLASS_DISPLAY_VGA << 8, + .class_mask = 0xffffff, + .driver_data = CHIP_IP_DISCOVERY }, + + { PCI_DEVICE(0x1002, PCI_ANY_ID), + .class = PCI_CLASS_DISPLAY_OTHER << 8, + .class_mask = 0xffffff, + .driver_data = CHIP_IP_DISCOVERY }, + {0, 0, 0} }; @@ -1246,14 +1893,47 @@ MODULE_DEVICE_TABLE(pci, pciidlist); static const struct drm_driver amdgpu_kms_driver; +static bool amdgpu_is_fw_framebuffer(resource_size_t base, + resource_size_t size) +{ + bool found = false; +#if IS_REACHABLE(CONFIG_FB) + struct apertures_struct *a; + + a = alloc_apertures(1); + if (!a) + return false; + + a->ranges[0].base = base; + a->ranges[0].size = size; + + found = is_firmware_framebuffer(a); + kfree(a); +#endif + return found; +} + static int amdgpu_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) { struct drm_device *ddev; struct amdgpu_device *adev; unsigned long flags = ent->driver_data; - int ret, retry = 0; + int ret, retry = 0, i; bool supports_atomic = false; + bool is_fw_fb; + resource_size_t base, size; + + /* skip devices which are owned by radeon */ + for (i = 0; i < ARRAY_SIZE(amdgpu_unsupported_pciidlist); i++) { + if (amdgpu_unsupported_pciidlist[i] == pdev->device) + return -ENODEV; + } + + if (flags == 0) { + DRM_INFO("Unsupported asic. Remove me when IP discovery init is in place.\n"); + return -ENODEV; + } if (amdgpu_virtual_display || amdgpu_device_asic_has_dc_support(flags & AMD_ASIC_MASK)) @@ -1269,7 +1949,8 @@ static int amdgpu_pci_probe(struct pci_dev *pdev, * however, SME requires an indirect IOMMU mapping because the encryption * bit is beyond the DMA mask of the chip. */ - if (mem_encrypt_active() && ((flags & AMD_ASIC_MASK) == CHIP_RAVEN)) { + if (cc_platform_has(CC_ATTR_MEM_ENCRYPT) && + ((flags & AMD_ASIC_MASK) == CHIP_RAVEN)) { dev_info(&pdev->dev, "SME is not compatible with RAVEN\n"); return -ENOTSUPP; @@ -1310,6 +1991,10 @@ static int amdgpu_pci_probe(struct pci_dev *pdev, } #endif + base = pci_resource_start(pdev, 0); + size = pci_resource_len(pdev, 0); + is_fw_fb = amdgpu_is_fw_framebuffer(base, size); + /* Get rid of things like offb */ ret = drm_aperture_remove_conflicting_pci_framebuffers(pdev, &amdgpu_kms_driver); if (ret) @@ -1322,6 +2007,7 @@ static int amdgpu_pci_probe(struct pci_dev *pdev, adev->dev = &pdev->dev; adev->pdev = pdev; ddev = adev_to_drm(adev); + adev->is_fw_fb = is_fw_fb; if (!supports_atomic) ddev->driver_features &= ~DRIVER_ATOMIC; @@ -1347,6 +2033,19 @@ retry_init: goto err_pci; } + /* + * 1. don't init fbdev on hw without DCE + * 2. don't init fbdev if there are no connectors + */ + if (adev->mode_info.mode_config_initialized && + !list_empty(&adev_to_drm(adev)->mode_config.connector_list)) { + /* select 8 bpp console on low vram cards */ + if (adev->gmc.real_vram_size <= (32*1024*1024)) + drm_fbdev_generic_setup(adev_to_drm(adev), 8); + else + drm_fbdev_generic_setup(adev_to_drm(adev), 32); + } + ret = amdgpu_debugfs_init(adev); if (ret) DRM_ERROR("Creating debugfs files failed (%d).\n", ret); @@ -1498,7 +2197,10 @@ static int amdgpu_pmops_suspend(struct device *dev) adev->in_s3 = true; r = amdgpu_device_suspend(drm_dev, true); adev->in_s3 = false; - + if (r) + return r; + if (!adev->in_s0ix) + r = amdgpu_asic_reset(adev); return r; } @@ -1508,6 +2210,10 @@ static int amdgpu_pmops_resume(struct device *dev) struct amdgpu_device *adev = drm_to_adev(drm_dev); int r; + /* Avoids registers access if device is physically gone */ + if (!pci_device_is_present(adev->pdev)) + adev->no_hw_access = true; + r = amdgpu_device_resume(drm_dev, true); if (amdgpu_acpi_is_s0ix_active(adev)) adev->in_s0ix = false; @@ -1575,12 +2281,27 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev) if (amdgpu_device_supports_px(drm_dev)) drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; + /* + * By setting mp1_state as PP_MP1_STATE_UNLOAD, MP1 will do some + * proper cleanups and put itself into a state ready for PNP. That + * can address some random resuming failure observed on BOCO capable + * platforms. + * TODO: this may be also needed for PX capable platform. + */ + if (amdgpu_device_supports_boco(drm_dev)) + adev->mp1_state = PP_MP1_STATE_UNLOAD; + ret = amdgpu_device_suspend(drm_dev, false); if (ret) { adev->in_runpm = false; + if (amdgpu_device_supports_boco(drm_dev)) + adev->mp1_state = PP_MP1_STATE_NONE; return ret; } + if (amdgpu_device_supports_boco(drm_dev)) + adev->mp1_state = PP_MP1_STATE_NONE; + if (amdgpu_device_supports_px(drm_dev)) { /* Only need to handle PCI state in the driver for ATPX * PCI core handles it for _PR3. @@ -1857,10 +2578,8 @@ static int __init amdgpu_init(void) { int r; - if (vgacon_text_force()) { - DRM_ERROR("VGACON disables amdgpu kernel modesetting.\n"); + if (drm_firmware_drivers_only()) return -EINVAL; - } r = amdgpu_sync_init(); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h index e3a4f7048042..8178323e4bef 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h @@ -45,4 +45,7 @@ long amdgpu_drm_ioctl(struct file *filp, unsigned int cmd, unsigned long arg); +long amdgpu_kms_compat_ioctl(struct file *filp, + unsigned int cmd, unsigned long arg); + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c index af4ef84e27a7..c96e458ed088 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c @@ -222,7 +222,7 @@ bool amdgpu_dig_monitor_is_duallink(struct drm_encoder *encoder, case DRM_MODE_CONNECTOR_HDMIB: if (amdgpu_connector->use_digital) { /* HDMI 1.3 supports up to 340 Mhz over single link */ - if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) { + if (connector->display_info.is_hdmi) { if (pixel_clock > 340000) return true; else @@ -244,7 +244,7 @@ bool amdgpu_dig_monitor_is_duallink(struct drm_encoder *encoder, return false; else { /* HDMI 1.3 supports up to 340 Mhz over single link */ - if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) { + if (connector->display_info.is_hdmi) { if (pixel_clock > 340000) return true; else diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c deleted file mode 100644 index cd0acbea75da..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c +++ /dev/null @@ -1,388 +0,0 @@ -/* - * Copyright © 2007 David Airlie - * - * 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 (including the next - * paragraph) 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 AUTHORS OR COPYRIGHT HOLDERS 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. - * - * Authors: - * David Airlie - */ - -#include <linux/module.h> -#include <linux/pm_runtime.h> -#include <linux/slab.h> -#include <linux/vga_switcheroo.h> - -#include <drm/amdgpu_drm.h> -#include <drm/drm_crtc.h> -#include <drm/drm_crtc_helper.h> -#include <drm/drm_fb_helper.h> -#include <drm/drm_fourcc.h> - -#include "amdgpu.h" -#include "cikd.h" -#include "amdgpu_gem.h" - -#include "amdgpu_display.h" - -/* object hierarchy - - this contains a helper + a amdgpu fb - the helper contains a pointer to amdgpu framebuffer baseclass. -*/ - -static int -amdgpufb_open(struct fb_info *info, int user) -{ - struct drm_fb_helper *fb_helper = info->par; - int ret = pm_runtime_get_sync(fb_helper->dev->dev); - if (ret < 0 && ret != -EACCES) { - pm_runtime_mark_last_busy(fb_helper->dev->dev); - pm_runtime_put_autosuspend(fb_helper->dev->dev); - return ret; - } - return 0; -} - -static int -amdgpufb_release(struct fb_info *info, int user) -{ - struct drm_fb_helper *fb_helper = info->par; - - pm_runtime_mark_last_busy(fb_helper->dev->dev); - pm_runtime_put_autosuspend(fb_helper->dev->dev); - return 0; -} - -static const struct fb_ops amdgpufb_ops = { - .owner = THIS_MODULE, - DRM_FB_HELPER_DEFAULT_OPS, - .fb_open = amdgpufb_open, - .fb_release = amdgpufb_release, - .fb_fillrect = drm_fb_helper_cfb_fillrect, - .fb_copyarea = drm_fb_helper_cfb_copyarea, - .fb_imageblit = drm_fb_helper_cfb_imageblit, -}; - - -int amdgpu_align_pitch(struct amdgpu_device *adev, int width, int cpp, bool tiled) -{ - int aligned = width; - int pitch_mask = 0; - - switch (cpp) { - case 1: - pitch_mask = 255; - break; - case 2: - pitch_mask = 127; - break; - case 3: - case 4: - pitch_mask = 63; - break; - } - - aligned += pitch_mask; - aligned &= ~pitch_mask; - return aligned * cpp; -} - -static void amdgpufb_destroy_pinned_object(struct drm_gem_object *gobj) -{ - struct amdgpu_bo *abo = gem_to_amdgpu_bo(gobj); - int ret; - - ret = amdgpu_bo_reserve(abo, true); - if (likely(ret == 0)) { - amdgpu_bo_kunmap(abo); - amdgpu_bo_unpin(abo); - amdgpu_bo_unreserve(abo); - } - drm_gem_object_put(gobj); -} - -static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev, - struct drm_mode_fb_cmd2 *mode_cmd, - struct drm_gem_object **gobj_p) -{ - const struct drm_format_info *info; - struct amdgpu_device *adev = rfbdev->adev; - struct drm_gem_object *gobj = NULL; - struct amdgpu_bo *abo = NULL; - bool fb_tiled = false; /* useful for testing */ - u32 tiling_flags = 0, domain; - int ret; - int aligned_size, size; - int height = mode_cmd->height; - u32 cpp; - u64 flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED | - AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS | - AMDGPU_GEM_CREATE_VRAM_CLEARED; - - info = drm_get_format_info(adev_to_drm(adev), mode_cmd); - cpp = info->cpp[0]; - - /* need to align pitch with crtc limits */ - mode_cmd->pitches[0] = amdgpu_align_pitch(adev, mode_cmd->width, cpp, - fb_tiled); - domain = amdgpu_display_supported_domains(adev, flags); - height = ALIGN(mode_cmd->height, 8); - size = mode_cmd->pitches[0] * height; - aligned_size = ALIGN(size, PAGE_SIZE); - ret = amdgpu_gem_object_create(adev, aligned_size, 0, domain, flags, - ttm_bo_type_device, NULL, &gobj); - if (ret) { - pr_err("failed to allocate framebuffer (%d)\n", aligned_size); - return -ENOMEM; - } - abo = gem_to_amdgpu_bo(gobj); - - if (fb_tiled) - tiling_flags = AMDGPU_TILING_SET(ARRAY_MODE, GRPH_ARRAY_2D_TILED_THIN1); - - ret = amdgpu_bo_reserve(abo, false); - if (unlikely(ret != 0)) - goto out_unref; - - if (tiling_flags) { - ret = amdgpu_bo_set_tiling_flags(abo, - tiling_flags); - if (ret) - dev_err(adev->dev, "FB failed to set tiling flags\n"); - } - - ret = amdgpu_bo_pin(abo, domain); - if (ret) { - amdgpu_bo_unreserve(abo); - goto out_unref; - } - - ret = amdgpu_ttm_alloc_gart(&abo->tbo); - if (ret) { - amdgpu_bo_unreserve(abo); - dev_err(adev->dev, "%p bind failed\n", abo); - goto out_unref; - } - - ret = amdgpu_bo_kmap(abo, NULL); - amdgpu_bo_unreserve(abo); - if (ret) { - goto out_unref; - } - - *gobj_p = gobj; - return 0; -out_unref: - amdgpufb_destroy_pinned_object(gobj); - *gobj_p = NULL; - return ret; -} - -static int amdgpufb_create(struct drm_fb_helper *helper, - struct drm_fb_helper_surface_size *sizes) -{ - struct amdgpu_fbdev *rfbdev = (struct amdgpu_fbdev *)helper; - struct amdgpu_device *adev = rfbdev->adev; - struct fb_info *info; - struct drm_framebuffer *fb = NULL; - struct drm_mode_fb_cmd2 mode_cmd; - struct drm_gem_object *gobj = NULL; - struct amdgpu_bo *abo = NULL; - int ret; - - memset(&mode_cmd, 0, sizeof(mode_cmd)); - mode_cmd.width = sizes->surface_width; - mode_cmd.height = sizes->surface_height; - - if (sizes->surface_bpp == 24) - sizes->surface_bpp = 32; - - mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp, - sizes->surface_depth); - - ret = amdgpufb_create_pinned_object(rfbdev, &mode_cmd, &gobj); - if (ret) { - DRM_ERROR("failed to create fbcon object %d\n", ret); - return ret; - } - - abo = gem_to_amdgpu_bo(gobj); - - /* okay we have an object now allocate the framebuffer */ - info = drm_fb_helper_alloc_fbi(helper); - if (IS_ERR(info)) { - ret = PTR_ERR(info); - goto out; - } - - ret = amdgpu_display_gem_fb_init(adev_to_drm(adev), &rfbdev->rfb, - &mode_cmd, gobj); - if (ret) { - DRM_ERROR("failed to initialize framebuffer %d\n", ret); - goto out; - } - - fb = &rfbdev->rfb.base; - - /* setup helper */ - rfbdev->helper.fb = fb; - - info->fbops = &amdgpufb_ops; - - info->fix.smem_start = amdgpu_gmc_vram_cpu_pa(adev, abo); - info->fix.smem_len = amdgpu_bo_size(abo); - info->screen_base = amdgpu_bo_kptr(abo); - info->screen_size = amdgpu_bo_size(abo); - - drm_fb_helper_fill_info(info, &rfbdev->helper, sizes); - - /* setup aperture base/size for vesafb takeover */ - info->apertures->ranges[0].base = adev_to_drm(adev)->mode_config.fb_base; - info->apertures->ranges[0].size = adev->gmc.aper_size; - - /* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */ - - if (info->screen_base == NULL) { - ret = -ENOSPC; - goto out; - } - - DRM_INFO("fb mappable at 0x%lX\n", info->fix.smem_start); - DRM_INFO("vram apper at 0x%lX\n", (unsigned long)adev->gmc.aper_base); - DRM_INFO("size %lu\n", (unsigned long)amdgpu_bo_size(abo)); - DRM_INFO("fb depth is %d\n", fb->format->depth); - DRM_INFO(" pitch is %d\n", fb->pitches[0]); - - vga_switcheroo_client_fb_set(adev->pdev, info); - return 0; - -out: - if (fb && ret) { - drm_gem_object_put(gobj); - drm_framebuffer_unregister_private(fb); - drm_framebuffer_cleanup(fb); - kfree(fb); - } - return ret; -} - -static int amdgpu_fbdev_destroy(struct drm_device *dev, struct amdgpu_fbdev *rfbdev) -{ - struct amdgpu_framebuffer *rfb = &rfbdev->rfb; - int i; - - drm_fb_helper_unregister_fbi(&rfbdev->helper); - - if (rfb->base.obj[0]) { - for (i = 0; i < rfb->base.format->num_planes; i++) - drm_gem_object_put(rfb->base.obj[0]); - amdgpufb_destroy_pinned_object(rfb->base.obj[0]); - rfb->base.obj[0] = NULL; - drm_framebuffer_unregister_private(&rfb->base); - drm_framebuffer_cleanup(&rfb->base); - } - drm_fb_helper_fini(&rfbdev->helper); - - return 0; -} - -static const struct drm_fb_helper_funcs amdgpu_fb_helper_funcs = { - .fb_probe = amdgpufb_create, -}; - -int amdgpu_fbdev_init(struct amdgpu_device *adev) -{ - struct amdgpu_fbdev *rfbdev; - int bpp_sel = 32; - int ret; - - /* don't init fbdev on hw without DCE */ - if (!adev->mode_info.mode_config_initialized) - return 0; - - /* don't init fbdev if there are no connectors */ - if (list_empty(&adev_to_drm(adev)->mode_config.connector_list)) - return 0; - - /* select 8 bpp console on low vram cards */ - if (adev->gmc.real_vram_size <= (32*1024*1024)) - bpp_sel = 8; - - rfbdev = kzalloc(sizeof(struct amdgpu_fbdev), GFP_KERNEL); - if (!rfbdev) - return -ENOMEM; - - rfbdev->adev = adev; - adev->mode_info.rfbdev = rfbdev; - - drm_fb_helper_prepare(adev_to_drm(adev), &rfbdev->helper, - &amdgpu_fb_helper_funcs); - - ret = drm_fb_helper_init(adev_to_drm(adev), &rfbdev->helper); - if (ret) { - kfree(rfbdev); - return ret; - } - - /* disable all the possible outputs/crtcs before entering KMS mode */ - if (!amdgpu_device_has_dc_support(adev) && !amdgpu_virtual_display) - drm_helper_disable_unused_functions(adev_to_drm(adev)); - - drm_fb_helper_initial_config(&rfbdev->helper, bpp_sel); - return 0; -} - -void amdgpu_fbdev_fini(struct amdgpu_device *adev) -{ - if (!adev->mode_info.rfbdev) - return; - - amdgpu_fbdev_destroy(adev_to_drm(adev), adev->mode_info.rfbdev); - kfree(adev->mode_info.rfbdev); - adev->mode_info.rfbdev = NULL; -} - -void amdgpu_fbdev_set_suspend(struct amdgpu_device *adev, int state) -{ - if (adev->mode_info.rfbdev) - drm_fb_helper_set_suspend_unlocked(&adev->mode_info.rfbdev->helper, - state); -} - -int amdgpu_fbdev_total_size(struct amdgpu_device *adev) -{ - struct amdgpu_bo *robj; - int size = 0; - - if (!adev->mode_info.rfbdev) - return 0; - - robj = gem_to_amdgpu_bo(adev->mode_info.rfbdev->rfb.base.obj[0]); - size += amdgpu_bo_size(robj); - return size; -} - -bool amdgpu_fbdev_robj_is_fb(struct amdgpu_device *adev, struct amdgpu_bo *robj) -{ - if (!adev->mode_info.rfbdev) - return false; - if (robj == gem_to_amdgpu_bo(adev->mode_info.rfbdev->rfb.base.obj[0])) - return true; - return false; -} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index 8d682befe0d6..45977a72b5dd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -77,11 +77,13 @@ void amdgpu_fence_slab_fini(void) * Cast helper */ static const struct dma_fence_ops amdgpu_fence_ops; +static const struct dma_fence_ops amdgpu_job_fence_ops; static inline struct amdgpu_fence *to_amdgpu_fence(struct dma_fence *f) { struct amdgpu_fence *__f = container_of(f, struct amdgpu_fence, base); - if (__f->base.ops == &amdgpu_fence_ops) + if (__f->base.ops == &amdgpu_fence_ops || + __f->base.ops == &amdgpu_job_fence_ops) return __f; return NULL; @@ -158,19 +160,18 @@ int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **f, struct amd } seq = ++ring->fence_drv.sync_seq; - if (job != NULL && job->job_run_counter) { + if (job && job->job_run_counter) { /* reinit seq for resubmitted jobs */ fence->seqno = seq; } else { - dma_fence_init(fence, &amdgpu_fence_ops, - &ring->fence_drv.lock, - adev->fence_context + ring->idx, - seq); - } - - if (job != NULL) { - /* mark this fence has a parent job */ - set_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &fence->flags); + if (job) + dma_fence_init(fence, &amdgpu_job_fence_ops, + &ring->fence_drv.lock, + adev->fence_context + ring->idx, seq); + else + dma_fence_init(fence, &amdgpu_fence_ops, + &ring->fence_drv.lock, + adev->fence_context + ring->idx, seq); } amdgpu_ring_emit_fence(ring, ring->fence_drv.gpu_addr, @@ -266,7 +267,6 @@ bool amdgpu_fence_process(struct amdgpu_ring *ring) struct amdgpu_fence_driver *drv = &ring->fence_drv; struct amdgpu_device *adev = ring->adev; uint32_t seq, last_seq; - int r; do { last_seq = atomic_read(&ring->fence_drv.last_seq); @@ -298,12 +298,7 @@ bool amdgpu_fence_process(struct amdgpu_ring *ring) if (!fence) continue; - r = dma_fence_signal(fence); - if (!r) - DMA_FENCE_TRACE(fence, "signaled from irq context\n"); - else - BUG(); - + dma_fence_signal(fence); dma_fence_put(fence); pm_runtime_mark_last_busy(adev_to_drm(adev)->dev); pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); @@ -552,11 +547,8 @@ void amdgpu_fence_driver_hw_fini(struct amdgpu_device *adev) if (!ring || !ring->fence_drv.initialized) continue; - if (!ring->no_scheduler) - drm_sched_stop(&ring->sched, NULL); - /* You can't wait for HW to signal if it's gone */ - if (!drm_dev_is_unplugged(&adev->ddev)) + if (!drm_dev_is_unplugged(adev_to_drm(adev))) r = amdgpu_fence_wait_empty(ring); else r = -ENODEV; @@ -614,11 +606,6 @@ void amdgpu_fence_driver_hw_init(struct amdgpu_device *adev) if (!ring || !ring->fence_drv.initialized) continue; - if (!ring->no_scheduler) { - drm_sched_resubmit_jobs(&ring->sched); - drm_sched_start(&ring->sched, true); - } - /* enable the interrupt */ if (ring->fence_drv.irq_src) amdgpu_irq_get(adev, ring->fence_drv.irq_src, @@ -627,6 +614,25 @@ void amdgpu_fence_driver_hw_init(struct amdgpu_device *adev) } /** + * amdgpu_fence_driver_clear_job_fences - clear job embedded fences of ring + * + * @ring: fence of the ring to be cleared + * + */ +void amdgpu_fence_driver_clear_job_fences(struct amdgpu_ring *ring) +{ + int i; + struct dma_fence *old, **ptr; + + for (i = 0; i <= ring->fence_drv.num_fences_mask; i++) { + ptr = &ring->fence_drv.fences[i]; + old = rcu_dereference_protected(*ptr, 1); + if (old && old->ops == &amdgpu_job_fence_ops) + RCU_INIT_POINTER(*ptr, NULL); + } +} + +/** * amdgpu_fence_driver_force_completion - force signal latest fence of ring * * @ring: fence of the ring to signal @@ -649,16 +655,14 @@ static const char *amdgpu_fence_get_driver_name(struct dma_fence *fence) static const char *amdgpu_fence_get_timeline_name(struct dma_fence *f) { - struct amdgpu_ring *ring; + return (const char *)to_amdgpu_fence(f)->ring->name; +} - if (test_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &f->flags)) { - struct amdgpu_job *job = container_of(f, struct amdgpu_job, hw_fence); +static const char *amdgpu_job_fence_get_timeline_name(struct dma_fence *f) +{ + struct amdgpu_job *job = container_of(f, struct amdgpu_job, hw_fence); - ring = to_amdgpu_ring(job->base.sched); - } else { - ring = to_amdgpu_fence(f)->ring; - } - return (const char *)ring->name; + return (const char *)to_amdgpu_ring(job->base.sched)->name; } /** @@ -671,20 +675,25 @@ static const char *amdgpu_fence_get_timeline_name(struct dma_fence *f) */ static bool amdgpu_fence_enable_signaling(struct dma_fence *f) { - struct amdgpu_ring *ring; - - if (test_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &f->flags)) { - struct amdgpu_job *job = container_of(f, struct amdgpu_job, hw_fence); + if (!timer_pending(&to_amdgpu_fence(f)->ring->fence_drv.fallback_timer)) + amdgpu_fence_schedule_fallback(to_amdgpu_fence(f)->ring); - ring = to_amdgpu_ring(job->base.sched); - } else { - ring = to_amdgpu_fence(f)->ring; - } + return true; +} - if (!timer_pending(&ring->fence_drv.fallback_timer)) - amdgpu_fence_schedule_fallback(ring); +/** + * amdgpu_job_fence_enable_signaling - enable signalling on job fence + * @f: fence + * + * This is the simliar function with amdgpu_fence_enable_signaling above, it + * only handles the job embedded fence. + */ +static bool amdgpu_job_fence_enable_signaling(struct dma_fence *f) +{ + struct amdgpu_job *job = container_of(f, struct amdgpu_job, hw_fence); - DMA_FENCE_TRACE(f, "armed on ring %i!\n", ring->idx); + if (!timer_pending(&to_amdgpu_ring(job->base.sched)->fence_drv.fallback_timer)) + amdgpu_fence_schedule_fallback(to_amdgpu_ring(job->base.sched)); return true; } @@ -700,19 +709,23 @@ static void amdgpu_fence_free(struct rcu_head *rcu) { struct dma_fence *f = container_of(rcu, struct dma_fence, rcu); - if (test_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &f->flags)) { - /* free job if fence has a parent job */ - struct amdgpu_job *job; - - job = container_of(f, struct amdgpu_job, hw_fence); - kfree(job); - } else { /* free fence_slab if it's separated fence*/ - struct amdgpu_fence *fence; + kmem_cache_free(amdgpu_fence_slab, to_amdgpu_fence(f)); +} - fence = to_amdgpu_fence(f); - kmem_cache_free(amdgpu_fence_slab, fence); - } +/** + * amdgpu_job_fence_free - free up the job with embedded fence + * + * @rcu: RCU callback head + * + * Free up the job with embedded fence after the RCU grace period. + */ +static void amdgpu_job_fence_free(struct rcu_head *rcu) +{ + struct dma_fence *f = container_of(rcu, struct dma_fence, rcu); + + /* free job if fence has a parent job */ + kfree(container_of(f, struct amdgpu_job, hw_fence)); } /** @@ -728,6 +741,19 @@ static void amdgpu_fence_release(struct dma_fence *f) call_rcu(&f->rcu, amdgpu_fence_free); } +/** + * amdgpu_job_fence_release - callback that job embedded fence can be freed + * + * @f: fence + * + * This is the simliar function with amdgpu_fence_release above, it + * only handles the job embedded fence. + */ +static void amdgpu_job_fence_release(struct dma_fence *f) +{ + call_rcu(&f->rcu, amdgpu_job_fence_free); +} + static const struct dma_fence_ops amdgpu_fence_ops = { .get_driver_name = amdgpu_fence_get_driver_name, .get_timeline_name = amdgpu_fence_get_timeline_name, @@ -735,6 +761,12 @@ static const struct dma_fence_ops amdgpu_fence_ops = { .release = amdgpu_fence_release, }; +static const struct dma_fence_ops amdgpu_job_fence_ops = { + .get_driver_name = amdgpu_fence_get_driver_name, + .get_timeline_name = amdgpu_job_fence_get_timeline_name, + .enable_signaling = amdgpu_job_fence_enable_signaling, + .release = amdgpu_job_fence_release, +}; /* * Fence debugfs diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c index 7709caeb233d..2a786e788627 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c @@ -56,6 +56,9 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev) return true; else return false; + case CHIP_ALDEBARAN: + /* All Aldebaran SKUs have the FRU */ + return true; default: return false; } @@ -88,13 +91,17 @@ static int amdgpu_fru_read_eeprom(struct amdgpu_device *adev, uint32_t addrptr, int amdgpu_fru_get_product_info(struct amdgpu_device *adev) { - unsigned char buff[34]; + unsigned char buff[AMDGPU_PRODUCT_NAME_LEN+2]; u32 addrptr; int size, len; + int offset = 2; if (!is_fru_eeprom_supported(adev)) return 0; + if (adev->asic_type == CHIP_ALDEBARAN) + offset = 0; + /* If algo exists, it means that the i2c_adapter's initialized */ if (!adev->pm.smu_i2c.algo) { DRM_WARN("Cannot access FRU, EEPROM accessor not initialized"); @@ -131,15 +138,13 @@ int amdgpu_fru_get_product_info(struct amdgpu_device *adev) } len = size; - /* Product name should only be 32 characters. Any more, - * and something could be wrong. Cap it at 32 to be safe - */ - if (len >= sizeof(adev->product_name)) { - DRM_WARN("FRU Product Number is larger than 32 characters. This is likely a mistake"); - len = sizeof(adev->product_name) - 1; + if (len >= AMDGPU_PRODUCT_NAME_LEN) { + DRM_WARN("FRU Product Name is larger than %d characters. This is likely a mistake", + AMDGPU_PRODUCT_NAME_LEN); + len = AMDGPU_PRODUCT_NAME_LEN - 1; } /* Start at 2 due to buff using fields 0 and 1 for the address */ - memcpy(adev->product_name, &buff[2], len); + memcpy(adev->product_name, &buff[offset], len); adev->product_name[len] = '\0'; addrptr += size + 1; @@ -157,7 +162,7 @@ int amdgpu_fru_get_product_info(struct amdgpu_device *adev) DRM_WARN("FRU Product Number is larger than 16 characters. This is likely a mistake"); len = sizeof(adev->product_number) - 1; } - memcpy(adev->product_number, &buff[2], len); + memcpy(adev->product_number, &buff[offset], len); adev->product_number[len] = '\0'; addrptr += size + 1; @@ -184,7 +189,7 @@ int amdgpu_fru_get_product_info(struct amdgpu_device *adev) DRM_WARN("FRU Serial Number is larger than 16 characters. This is likely a mistake"); len = sizeof(adev->serial) - 1; } - memcpy(adev->serial, &buff[2], len); + memcpy(adev->serial, &buff[offset], len); adev->serial[len] = '\0'; return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c index 76efd5f8950f..d3e4203f6217 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c @@ -34,6 +34,7 @@ #include <asm/set_memory.h> #endif #include "amdgpu.h" +#include <drm/drm_drv.h> /* * GART @@ -230,12 +231,16 @@ int amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset, u64 page_base; /* Starting from VEGA10, system bit must be 0 to mean invalid. */ uint64_t flags = 0; + int idx; if (!adev->gart.ready) { WARN(1, "trying to unbind memory from uninitialized GART !\n"); return -EINVAL; } + if (!drm_dev_enter(adev_to_drm(adev), &idx)) + return 0; + t = offset / AMDGPU_GPU_PAGE_SIZE; p = t / AMDGPU_GPU_PAGES_IN_CPU_PAGE; for (i = 0; i < pages; i++, p++) { @@ -254,6 +259,7 @@ int amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset, for (i = 0; i < adev->num_vmhubs; i++) amdgpu_gmc_flush_gpu_tlb(adev, 0, i, 0); + drm_dev_exit(idx); return 0; } @@ -276,12 +282,16 @@ int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset, { uint64_t page_base; unsigned i, j, t; + int idx; if (!adev->gart.ready) { WARN(1, "trying to bind memory to uninitialized GART !\n"); return -EINVAL; } + if (!drm_dev_enter(adev_to_drm(adev), &idx)) + return 0; + t = offset / AMDGPU_GPU_PAGE_SIZE; for (i = 0; i < pages; i++) { @@ -291,6 +301,7 @@ int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset, page_base += AMDGPU_GPU_PAGE_SIZE; } } + drm_dev_exit(idx); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index d6aa032890ee..9a6507af1670 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -61,7 +61,7 @@ static vm_fault_t amdgpu_gem_fault(struct vm_fault *vmf) } ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot, - TTM_BO_VM_NUM_PREFAULT, 1); + TTM_BO_VM_NUM_PREFAULT); drm_dev_exit(idx); } else { @@ -264,6 +264,9 @@ static int amdgpu_gem_object_mmap(struct drm_gem_object *obj, struct vm_area_str !(vma->vm_flags & (VM_READ | VM_WRITE | VM_EXEC))) vma->vm_flags &= ~VM_MAYWRITE; + if (bo->kfd_bo) + vma->vm_flags |= VM_DONTCOPY; + return drm_gem_ttm_mmap(obj, vma); } @@ -877,6 +880,32 @@ out: return r; } +static int amdgpu_gem_align_pitch(struct amdgpu_device *adev, + int width, + int cpp, + bool tiled) +{ + int aligned = width; + int pitch_mask = 0; + + switch (cpp) { + case 1: + pitch_mask = 255; + break; + case 2: + pitch_mask = 127; + break; + case 3: + case 4: + pitch_mask = 63; + break; + } + + aligned += pitch_mask; + aligned &= ~pitch_mask; + return aligned * cpp; +} + int amdgpu_mode_dumb_create(struct drm_file *file_priv, struct drm_device *dev, struct drm_mode_create_dumb *args) @@ -885,7 +914,8 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv, struct drm_gem_object *gobj; uint32_t handle; u64 flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED | - AMDGPU_GEM_CREATE_CPU_GTT_USWC; + AMDGPU_GEM_CREATE_CPU_GTT_USWC | + AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS; u32 domain; int r; @@ -897,8 +927,8 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv, if (adev->mman.buffer_funcs_enabled) flags |= AMDGPU_GEM_CREATE_VRAM_CLEARED; - args->pitch = amdgpu_align_pitch(adev, args->width, - DIV_ROUND_UP(args->bpp, 8), 0); + args->pitch = amdgpu_gem_align_pitch(adev, args->width, + DIV_ROUND_UP(args->bpp, 8), 0); args->size = (u64)args->pitch * args->height; args->size = ALIGN(args->size, PAGE_SIZE); domain = amdgpu_bo_get_preferred_domain(adev, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h index d43fe2ed8116..f851196c83a5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h @@ -42,10 +42,9 @@ #define AMDGPU_MAX_GFX_QUEUES KGD_MAX_QUEUES #define AMDGPU_MAX_COMPUTE_QUEUES KGD_MAX_QUEUES -enum gfx_pipe_priority { - AMDGPU_GFX_PIPE_PRIO_NORMAL = 1, - AMDGPU_GFX_PIPE_PRIO_HIGH, - AMDGPU_GFX_PIPE_PRIO_MAX +enum amdgpu_gfx_pipe_priority { + AMDGPU_GFX_PIPE_PRIO_NORMAL = AMDGPU_RING_PRIO_1, + AMDGPU_GFX_PIPE_PRIO_HIGH = AMDGPU_RING_PRIO_2 }; /* Argument for PPSMC_MSG_GpuChangeState */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c index 9ff600a38559..2430d6223c2d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c @@ -153,10 +153,6 @@ int amdgpu_gmc_set_pte_pde(struct amdgpu_device *adev, void *cpu_pt_addr, { void __iomem *ptr = (void *)cpu_pt_addr; uint64_t value; - int idx; - - if (!drm_dev_enter(&adev->ddev, &idx)) - return 0; /* * The following is for PTE only. GART does not have PDEs. @@ -165,8 +161,6 @@ int amdgpu_gmc_set_pte_pde(struct amdgpu_device *adev, void *cpu_pt_addr, value |= flags; writeq(value, ptr + (gpu_page_idx * 8)); - drm_dev_exit(idx); - return 0; } @@ -356,6 +350,7 @@ static inline uint64_t amdgpu_gmc_fault_key(uint64_t addr, uint16_t pasid) * amdgpu_gmc_filter_faults - filter VM faults * * @adev: amdgpu device structure + * @ih: interrupt ring that the fault received from * @addr: address of the VM fault * @pasid: PASID of the process causing the fault * @timestamp: timestamp of the fault @@ -364,7 +359,8 @@ static inline uint64_t amdgpu_gmc_fault_key(uint64_t addr, uint16_t pasid) * True if the fault was filtered and should not be processed further. * False if the fault is a new one and needs to be handled. */ -bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev, uint64_t addr, +bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih, uint64_t addr, uint16_t pasid, uint64_t timestamp) { struct amdgpu_gmc *gmc = &adev->gmc; @@ -372,6 +368,10 @@ bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev, uint64_t addr, struct amdgpu_gmc_fault *fault; uint32_t hash; + /* Stale retry fault if timestamp goes backward */ + if (amdgpu_ih_ts_after(timestamp, ih->processed_timestamp)) + return true; + /* If we don't have space left in the ring buffer return immediately */ stamp = max(timestamp, AMDGPU_GMC_FAULT_TIMEOUT + 1) - AMDGPU_GMC_FAULT_TIMEOUT; @@ -749,6 +749,10 @@ void amdgpu_gmc_init_pdb0(struct amdgpu_device *adev) adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size; u64 vram_end = vram_addr + vram_size; u64 gart_ptb_gpu_pa = amdgpu_gmc_vram_pa(adev, adev->gart.bo); + int idx; + + if (!drm_dev_enter(adev_to_drm(adev), &idx)) + return; flags |= AMDGPU_PTE_VALID | AMDGPU_PTE_READABLE; flags |= AMDGPU_PTE_WRITEABLE; @@ -770,6 +774,7 @@ void amdgpu_gmc_init_pdb0(struct amdgpu_device *adev) flags |= AMDGPU_PDE_BFS(0) | AMDGPU_PTE_SNOOPED; /* Requires gart_ptb_gpu_pa to be 4K aligned */ amdgpu_gmc_set_pte_pde(adev, adev->gmc.ptr_pdb0, i, gart_ptb_gpu_pa, flags); + drm_dev_exit(idx); } /** diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h index e55201134a01..8458cebc6d5b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h @@ -316,7 +316,8 @@ void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc); void amdgpu_gmc_agp_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc); -bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev, uint64_t addr, +bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih, uint64_t addr, uint16_t pasid, uint64_t timestamp); void amdgpu_gmc_filter_faults_remove(struct amdgpu_device *adev, uint64_t addr, uint16_t pasid); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index c076a6b9a5a2..bc1297dcdf97 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -300,20 +300,15 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, */ int amdgpu_ib_pool_init(struct amdgpu_device *adev) { - unsigned size; int r, i; if (adev->ib_pool_ready) return 0; for (i = 0; i < AMDGPU_IB_POOL_MAX; i++) { - if (i == AMDGPU_IB_POOL_DIRECT) - size = PAGE_SIZE * 6; - else - size = AMDGPU_IB_POOL_SIZE; - r = amdgpu_sa_bo_manager_init(adev, &adev->ib_pools[i], - size, AMDGPU_GPU_PAGE_SIZE, + AMDGPU_IB_POOL_SIZE, + AMDGPU_GPU_PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT); if (r) goto error; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c index f3d62e196901..3df146579ad9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c @@ -164,52 +164,32 @@ void amdgpu_ih_ring_write(struct amdgpu_ih_ring *ih, const uint32_t *iv, } } -/* Waiter helper that checks current rptr matches or passes checkpoint wptr */ -static bool amdgpu_ih_has_checkpoint_processed(struct amdgpu_device *adev, - struct amdgpu_ih_ring *ih, - uint32_t checkpoint_wptr, - uint32_t *prev_rptr) -{ - uint32_t cur_rptr = ih->rptr | (*prev_rptr & ~ih->ptr_mask); - - /* rptr has wrapped. */ - if (cur_rptr < *prev_rptr) - cur_rptr += ih->ptr_mask + 1; - *prev_rptr = cur_rptr; - - /* check ring is empty to workaround missing wptr overflow flag */ - return cur_rptr >= checkpoint_wptr || - (cur_rptr & ih->ptr_mask) == amdgpu_ih_get_wptr(adev, ih); -} - /** - * amdgpu_ih_wait_on_checkpoint_process - wait to process IVs up to checkpoint + * amdgpu_ih_wait_on_checkpoint_process_ts - wait to process IVs up to checkpoint * * @adev: amdgpu_device pointer * @ih: ih ring to process * * Used to ensure ring has processed IVs up to the checkpoint write pointer. */ -int amdgpu_ih_wait_on_checkpoint_process(struct amdgpu_device *adev, +int amdgpu_ih_wait_on_checkpoint_process_ts(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih) { - uint32_t checkpoint_wptr, rptr; + uint32_t checkpoint_wptr; + uint64_t checkpoint_ts; + long timeout = HZ; if (!ih->enabled || adev->shutdown) return -ENODEV; checkpoint_wptr = amdgpu_ih_get_wptr(adev, ih); - /* Order wptr with rptr. */ + /* Order wptr with ring data. */ rmb(); - rptr = READ_ONCE(ih->rptr); - - /* wptr has wrapped. */ - if (rptr > checkpoint_wptr) - checkpoint_wptr += ih->ptr_mask + 1; + checkpoint_ts = amdgpu_ih_decode_iv_ts(adev, ih, checkpoint_wptr, -1); - return wait_event_interruptible(ih->wait_process, - amdgpu_ih_has_checkpoint_processed(adev, ih, - checkpoint_wptr, &rptr)); + return wait_event_interruptible_timeout(ih->wait_process, + amdgpu_ih_ts_after(checkpoint_ts, ih->processed_timestamp) || + ih->rptr == amdgpu_ih_get_wptr(adev, ih), timeout); } /** @@ -223,7 +203,7 @@ int amdgpu_ih_wait_on_checkpoint_process(struct amdgpu_device *adev, */ int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih) { - unsigned int count = AMDGPU_IH_MAX_NUM_IVS; + unsigned int count; u32 wptr; if (!ih->enabled || adev->shutdown) @@ -232,6 +212,7 @@ int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih) wptr = amdgpu_ih_get_wptr(adev, ih); restart_ih: + count = AMDGPU_IH_MAX_NUM_IVS; DRM_DEBUG("%s: rptr %d, wptr %d\n", __func__, ih->rptr, wptr); /* Order reading of wptr vs. reading of IH ring data */ @@ -298,3 +279,18 @@ void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev, /* wptr/rptr are in bytes! */ ih->rptr += 32; } + +uint64_t amdgpu_ih_decode_iv_ts_helper(struct amdgpu_ih_ring *ih, u32 rptr, + signed int offset) +{ + uint32_t iv_size = 32; + uint32_t ring_index; + uint32_t dw1, dw2; + + rptr += iv_size * offset; + ring_index = (rptr & ih->ptr_mask) >> 2; + + dw1 = le32_to_cpu(ih->ring[ring_index + 1]); + dw2 = le32_to_cpu(ih->ring[ring_index + 2]); + return dw1 | ((u64)(dw2 & 0xffff) << 32); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h index 0649b59830a5..dd1c2eded6b9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h @@ -68,20 +68,30 @@ struct amdgpu_ih_ring { /* For waiting on IH processing at checkpoint. */ wait_queue_head_t wait_process; + uint64_t processed_timestamp; }; +/* return true if time stamp t2 is after t1 with 48bit wrap around */ +#define amdgpu_ih_ts_after(t1, t2) \ + (((int64_t)((t2) << 16) - (int64_t)((t1) << 16)) > 0LL) + /* provided by the ih block */ struct amdgpu_ih_funcs { /* ring read/write ptr handling, called from interrupt context */ u32 (*get_wptr)(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih); void (*decode_iv)(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, struct amdgpu_iv_entry *entry); + uint64_t (*decode_iv_ts)(struct amdgpu_ih_ring *ih, u32 rptr, + signed int offset); void (*set_rptr)(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih); }; #define amdgpu_ih_get_wptr(adev, ih) (adev)->irq.ih_funcs->get_wptr((adev), (ih)) #define amdgpu_ih_decode_iv(adev, iv) \ (adev)->irq.ih_funcs->decode_iv((adev), (ih), (iv)) +#define amdgpu_ih_decode_iv_ts(adev, ih, rptr, offset) \ + (WARN_ON_ONCE(!(adev)->irq.ih_funcs->decode_iv_ts) ? 0 : \ + (adev)->irq.ih_funcs->decode_iv_ts((ih), (rptr), (offset))) #define amdgpu_ih_set_rptr(adev, ih) (adev)->irq.ih_funcs->set_rptr((adev), (ih)) int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, @@ -89,10 +99,12 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih); void amdgpu_ih_ring_write(struct amdgpu_ih_ring *ih, const uint32_t *iv, unsigned int num_dw); -int amdgpu_ih_wait_on_checkpoint_process(struct amdgpu_device *adev, - struct amdgpu_ih_ring *ih); +int amdgpu_ih_wait_on_checkpoint_process_ts(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih); int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih); void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, struct amdgpu_iv_entry *entry); +uint64_t amdgpu_ih_decode_iv_ts_helper(struct amdgpu_ih_ring *ih, u32 rptr, + signed int offset); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ioc32.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ioc32.c index 5cf142e849bb..a1cbd7c3deb2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ioc32.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ioc32.c @@ -1,4 +1,4 @@ -/** +/* * \file amdgpu_ioc32.c * * 32-bit ioctl compatibility routines for the AMDGPU DRM. @@ -37,12 +37,9 @@ long amdgpu_kms_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { unsigned int nr = DRM_IOCTL_NR(cmd); - int ret; if (nr < DRM_COMMAND_BASE) return drm_compat_ioctl(filp, cmd, arg); - ret = amdgpu_drm_ioctl(filp, cmd, arg); - - return ret; + return amdgpu_drm_ioctl(filp, cmd, arg); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c index cc2e0c9cfe0a..f5cbc2747ac6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c @@ -333,7 +333,6 @@ int amdgpu_irq_init(struct amdgpu_device *adev) if (!amdgpu_device_has_dc_support(adev)) { if (!adev->enable_virtual_display) /* Disable vblank IRQs aggressively for power-saving */ - /* XXX: can this be enabled for DC? */ adev_to_drm(adev)->vblank_disable_immediate = true; r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc); @@ -391,7 +390,7 @@ void amdgpu_irq_fini_hw(struct amdgpu_device *adev) } /** - * amdgpu_irq_fini - shut down interrupt handling + * amdgpu_irq_fini_sw - shut down interrupt handling * * @adev: amdgpu device pointer * @@ -529,6 +528,9 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev, /* Send it to amdkfd as well if it isn't already handled */ if (!handled) amdgpu_amdkfd_interrupt(adev, entry.iv_entry); + + if (amdgpu_ih_ts_after(ih->processed_timestamp, entry.timestamp)) + ih->processed_timestamp = entry.timestamp; } /** diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index de29518673dd..bfc47bea23db 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -38,7 +38,7 @@ static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job) struct amdgpu_device *adev = ring->adev; int idx; - if (!drm_dev_enter(&adev->ddev, &idx)) { + if (!drm_dev_enter(adev_to_drm(adev), &idx)) { DRM_INFO("%s - device unplugged skipping recovery on scheduler:%s", __func__, s_job->sched->name); @@ -182,9 +182,11 @@ int amdgpu_job_submit(struct amdgpu_job *job, struct drm_sched_entity *entity, if (r) return r; + drm_sched_job_arm(&job->base); + *f = dma_fence_get(&job->base.s_fence->finished); amdgpu_job_free_resources(job); - drm_sched_entity_push_job(&job->base, entity); + drm_sched_entity_push_job(&job->base); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index 7e45640fbee0..09ad17944eb2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -206,6 +206,12 @@ int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags) adev->runpm = true; break; } + /* XXX: disable runtime pm if we are the primary adapter + * to avoid displays being re-enabled after DPMS. + * This needs to be sorted out and fixed properly. + */ + if (adev->is_fw_fb) + adev->runpm = false; if (adev->runpm) dev_info(adev->dev, "Using BACO for runtime pm\n"); } @@ -340,28 +346,35 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info, case AMDGPU_INFO_FW_TA: switch (query_fw->index) { case TA_FW_TYPE_PSP_XGMI: - fw_info->ver = adev->psp.ta_fw_version; - fw_info->feature = adev->psp.xgmi.feature_version; + fw_info->ver = adev->psp.xgmi_context.context.bin_desc.fw_version; + fw_info->feature = adev->psp.xgmi_context.context + .bin_desc.feature_version; break; case TA_FW_TYPE_PSP_RAS: - fw_info->ver = adev->psp.ta_fw_version; - fw_info->feature = adev->psp.ras.feature_version; + fw_info->ver = adev->psp.ras_context.context.bin_desc.fw_version; + fw_info->feature = adev->psp.ras_context.context + .bin_desc.feature_version; break; case TA_FW_TYPE_PSP_HDCP: - fw_info->ver = adev->psp.ta_fw_version; - fw_info->feature = adev->psp.hdcp.feature_version; + fw_info->ver = adev->psp.hdcp_context.context.bin_desc.fw_version; + fw_info->feature = adev->psp.hdcp_context.context + .bin_desc.feature_version; break; case TA_FW_TYPE_PSP_DTM: - fw_info->ver = adev->psp.ta_fw_version; - fw_info->feature = adev->psp.dtm.feature_version; + fw_info->ver = adev->psp.dtm_context.context.bin_desc.fw_version; + fw_info->feature = adev->psp.dtm_context.context + .bin_desc.feature_version; break; case TA_FW_TYPE_PSP_RAP: - fw_info->ver = adev->psp.ta_fw_version; - fw_info->feature = adev->psp.rap.feature_version; + fw_info->ver = adev->psp.rap_context.context.bin_desc.fw_version; + fw_info->feature = adev->psp.rap_context.context + .bin_desc.feature_version; break; case TA_FW_TYPE_PSP_SECUREDISPLAY: - fw_info->ver = adev->psp.ta_fw_version; - fw_info->feature = adev->psp.securedisplay.feature_version; + fw_info->ver = adev->psp.securedisplay_context.context.bin_desc.fw_version; + fw_info->feature = + adev->psp.securedisplay_context.context.bin_desc + .feature_version; break; default: return -EINVAL; @@ -378,8 +391,8 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info, fw_info->feature = adev->psp.sos.feature_version; break; case AMDGPU_INFO_FW_ASD: - fw_info->ver = adev->psp.asd.fw_version; - fw_info->feature = adev->psp.asd.feature_version; + fw_info->ver = adev->psp.asd_context.bin_desc.fw_version; + fw_info->feature = adev->psp.asd_context.bin_desc.feature_version; break; case AMDGPU_INFO_FW_DMCU: fw_info->ver = adev->dm.dmcu_fw_version; @@ -1416,6 +1429,8 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused) struct drm_amdgpu_info_firmware fw_info; struct drm_amdgpu_query_fw query_fw; struct atom_context *ctx = adev->mode_info.atom_context; + uint8_t smu_minor, smu_debug; + uint16_t smu_major; int ret, i; static const char *ta_fw_name[TA_FW_TYPE_MAX_INDEX] = { @@ -1561,8 +1576,11 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused) ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); if (ret) return ret; - seq_printf(m, "SMC feature version: %u, firmware version: 0x%08x\n", - fw_info.feature, fw_info.ver); + smu_major = (fw_info.ver >> 16) & 0xffff; + smu_minor = (fw_info.ver >> 8) & 0xff; + smu_debug = (fw_info.ver >> 0) & 0xff; + seq_printf(m, "SMC feature version: %u, firmware version: 0x%08x (%d.%d.%d)\n", + fw_info.feature, fw_info.ver, smu_major, smu_minor, smu_debug); /* SDMA */ query_fw.fw_type = AMDGPU_INFO_FW_SDMA; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c index a2d3dbbf7d25..ce538f4819f9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c @@ -31,7 +31,7 @@ void amdgpu_mca_query_correctable_error_count(struct amdgpu_device *adev, uint64_t mc_status_addr, unsigned long *error_count) { - uint64_t mc_status = RREG64_PCIE(mc_status_addr * 4); + uint64_t mc_status = RREG64_PCIE(mc_status_addr); if (REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 && REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1) @@ -42,7 +42,7 @@ void amdgpu_mca_query_uncorrectable_error_count(struct amdgpu_device *adev, uint64_t mc_status_addr, unsigned long *error_count) { - uint64_t mc_status = RREG64_PCIE(mc_status_addr * 4); + uint64_t mc_status = RREG64_PCIE(mc_status_addr); if ((REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) && (REG_GET_FIELD(mc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1 || @@ -56,7 +56,7 @@ void amdgpu_mca_query_uncorrectable_error_count(struct amdgpu_device *adev, void amdgpu_mca_reset_error_count(struct amdgpu_device *adev, uint64_t mc_status_addr) { - WREG64_PCIE(mc_status_addr * 4, 0x0ULL); + WREG64_PCIE(mc_status_addr, 0x0ULL); } void amdgpu_mca_query_ras_error_count(struct amdgpu_device *adev, @@ -87,8 +87,8 @@ int amdgpu_mca_ras_late_init(struct amdgpu_device *adev, if (!mca_dev->ras_if) return -ENOMEM; mca_dev->ras_if->block = mca_dev->ras_funcs->ras_block; + mca_dev->ras_if->sub_block_index = mca_dev->ras_funcs->ras_sub_block; mca_dev->ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE; - mca_dev->ras_if->sub_block_index = 0; } ih_info.head = fs_info.head = *mca_dev->ras_if; r = amdgpu_ras_late_init(adev, mca_dev->ras_if, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h index f860f2f0e296..c74bc7177066 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h @@ -29,6 +29,7 @@ struct amdgpu_mca_ras_funcs { void (*query_ras_error_address)(struct amdgpu_device *adev, void *ras_error_status); uint32_t ras_block; + uint32_t ras_sub_block; const char* sysfs_name; }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h index 89fb372ed49c..6043bf6fd414 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h @@ -232,8 +232,6 @@ struct amdgpu_i2c_chan { struct mutex mutex; }; -struct amdgpu_fbdev; - struct amdgpu_afmt { bool enabled; int offset; @@ -309,13 +307,6 @@ struct amdgpu_framebuffer { uint64_t address; }; -struct amdgpu_fbdev { - struct drm_fb_helper helper; - struct amdgpu_framebuffer rfb; - struct list_head fbdev_list; - struct amdgpu_device *adev; -}; - struct amdgpu_mode_info { struct atom_context *atom_context; struct card_info *atom_card_info; @@ -341,8 +332,6 @@ struct amdgpu_mode_info { struct edid *bios_hardcoded_edid; int bios_hardcoded_edid_size; - /* pointer to fbdev info structure */ - struct amdgpu_fbdev *rfbdev; /* firmware flags */ u32 firmware_flags; /* pointer to backlight encoder */ @@ -631,15 +620,6 @@ bool amdgpu_crtc_get_scanout_position(struct drm_crtc *crtc, int *hpos, ktime_t *stime, ktime_t *etime, const struct drm_display_mode *mode); -/* fbdev layer */ -int amdgpu_fbdev_init(struct amdgpu_device *adev); -void amdgpu_fbdev_fini(struct amdgpu_device *adev); -void amdgpu_fbdev_set_suspend(struct amdgpu_device *adev, int state); -int amdgpu_fbdev_total_size(struct amdgpu_device *adev); -bool amdgpu_fbdev_robj_is_fb(struct amdgpu_device *adev, struct amdgpu_bo *robj); - -int amdgpu_align_pitch(struct amdgpu_device *adev, int width, int bpp, bool tiled); - /* amdgpu_display.c */ void amdgpu_display_print_display_setup(struct drm_device *dev); int amdgpu_display_modeset_create_props(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 01a78c786536..3a7b56e57cec 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -695,40 +695,6 @@ int amdgpu_bo_create_vm(struct amdgpu_device *adev, } /** - * amdgpu_bo_validate - validate an &amdgpu_bo buffer object - * @bo: pointer to the buffer object - * - * Sets placement according to domain; and changes placement and caching - * policy of the buffer object according to the placement. - * This is used for validating shadow bos. It calls ttm_bo_validate() to - * make sure the buffer is resident where it needs to be. - * - * Returns: - * 0 for success or a negative error code on failure. - */ -int amdgpu_bo_validate(struct amdgpu_bo *bo) -{ - struct ttm_operation_ctx ctx = { false, false }; - uint32_t domain; - int r; - - if (bo->tbo.pin_count) - return 0; - - domain = bo->preferred_domains; - -retry: - amdgpu_bo_placement_from_domain(bo, domain); - r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); - if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) { - domain = bo->allowed_domains; - goto retry; - } - - return r; -} - -/** * amdgpu_bo_add_to_shadow_list - add a BO to the shadow list * * @vmbo: BO that will be inserted into the shadow list @@ -1038,29 +1004,6 @@ void amdgpu_bo_unpin(struct amdgpu_bo *bo) } } -/** - * amdgpu_bo_evict_vram - evict VRAM buffers - * @adev: amdgpu device object - * - * Evicts all VRAM buffers on the lru list of the memory type. - * Mainly used for evicting vram at suspend time. - * - * Returns: - * 0 for success or a negative error code on failure. - */ -int amdgpu_bo_evict_vram(struct amdgpu_device *adev) -{ - struct ttm_resource_manager *man; - - if (adev->in_s3 && (adev->flags & AMD_IS_APU)) { - /* No need to evict vram on APUs for suspend to ram */ - return 0; - } - - man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM); - return ttm_resource_manager_evict_all(&adev->mman.bdev, man); -} - static const char *amdgpu_vram_names[] = { "UNKNOWN", "GDDR1", @@ -1089,9 +1032,14 @@ int amdgpu_bo_init(struct amdgpu_device *adev) /* On A+A platform, VRAM can be mapped as WB */ if (!adev->gmc.xgmi.connected_to_cpu) { /* reserve PAT memory space to WC for VRAM */ - arch_io_reserve_memtype_wc(adev->gmc.aper_base, + int r = arch_io_reserve_memtype_wc(adev->gmc.aper_base, adev->gmc.aper_size); + if (r) { + DRM_ERROR("Unable to set WC memtype for the aperture base\n"); + return r; + } + /* Add an MTRR for the VRAM */ adev->gmc.vram_mtrr = arch_phys_wc_add(adev->gmc.aper_base, adev->gmc.aper_size); @@ -1331,7 +1279,7 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo) abo = ttm_to_amdgpu_bo(bo); if (abo->kfd_bo) - amdgpu_amdkfd_unreserve_memory_limit(abo); + amdgpu_amdkfd_release_notify(abo); /* We only remove the fence if the resv has individualized. */ WARN_ON_ONCE(bo->type == ttm_bo_type_kernel diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h index 9d6c001c15f8..4c9cbdc66995 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h @@ -304,7 +304,6 @@ int amdgpu_bo_pin(struct amdgpu_bo *bo, u32 domain); int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain, u64 min_offset, u64 max_offset); void amdgpu_bo_unpin(struct amdgpu_bo *bo); -int amdgpu_bo_evict_vram(struct amdgpu_device *adev); int amdgpu_bo_init(struct amdgpu_device *adev); void amdgpu_bo_fini(struct amdgpu_device *adev); int amdgpu_bo_set_tiling_flags(struct amdgpu_bo *bo, u64 tiling_flags); @@ -327,7 +326,6 @@ int amdgpu_bo_sync_wait_resv(struct amdgpu_device *adev, struct dma_resv *resv, int amdgpu_bo_sync_wait(struct amdgpu_bo *bo, void *owner, bool intr); u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo); u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo); -int amdgpu_bo_validate(struct amdgpu_bo *bo); void amdgpu_bo_get_memory(struct amdgpu_bo *bo, uint64_t *vram_mem, uint64_t *gtt_mem, uint64_t *cpu_mem); void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo_vm *vmbo); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c index 4eaec446b49d..0bb2466d539a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c @@ -69,6 +69,7 @@ static void amdgpu_pll_reduce_ratio(unsigned *nom, unsigned *den, /** * amdgpu_pll_get_fb_ref_div - feedback and ref divider calculation * + * @adev: amdgpu_device pointer * @nom: nominator * @den: denominator * @post_div: post divider @@ -106,6 +107,7 @@ static void amdgpu_pll_get_fb_ref_div(struct amdgpu_device *adev, unsigned int n /** * amdgpu_pll_compute - compute PLL paramaters * + * @adev: amdgpu_device pointer * @pll: information about the PLL * @freq: requested frequency * @dot_clock_p: resulting pixel clock diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c index 82e9ecf84352..71ee361d0972 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c @@ -233,6 +233,10 @@ static void amdgpu_perf_start(struct perf_event *event, int flags) if (WARN_ON_ONCE(!(hwc->state & PERF_HES_STOPPED))) return; + if ((!pe->adev->df.funcs) || + (!pe->adev->df.funcs->pmc_start)) + return; + WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE)); hwc->state = 0; @@ -268,6 +272,10 @@ static void amdgpu_perf_read(struct perf_event *event) pmu); u64 count, prev; + if ((!pe->adev->df.funcs) || + (!pe->adev->df.funcs->pmc_get_count)) + return; + do { prev = local64_read(&hwc->prev_count); @@ -297,6 +305,10 @@ static void amdgpu_perf_stop(struct perf_event *event, int flags) if (hwc->state & PERF_HES_UPTODATE) return; + if ((!pe->adev->df.funcs) || + (!pe->adev->df.funcs->pmc_stop)) + return; + switch (hwc->config_base) { case AMDGPU_PMU_EVENT_CONFIG_TYPE_DF: case AMDGPU_PMU_EVENT_CONFIG_TYPE_XGMI: @@ -326,6 +338,10 @@ static int amdgpu_perf_add(struct perf_event *event, int flags) struct amdgpu_pmu_entry, pmu); + if ((!pe->adev->df.funcs) || + (!pe->adev->df.funcs->pmc_start)) + return -EINVAL; + switch (pe->pmu_perf_type) { case AMDGPU_PMU_PERF_TYPE_DF: hwc->config_base = AMDGPU_PMU_EVENT_CONFIG_TYPE_DF; @@ -371,6 +387,9 @@ static void amdgpu_perf_del(struct perf_event *event, int flags) struct amdgpu_pmu_entry *pe = container_of(event->pmu, struct amdgpu_pmu_entry, pmu); + if ((!pe->adev->df.funcs) || + (!pe->adev->df.funcs->pmc_stop)) + return; amdgpu_perf_stop(event, PERF_EF_UPDATE); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c index d02c8637f909..786afe4f58f9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c @@ -59,7 +59,7 @@ static DEVICE_ATTR_RO(mem_info_preempt_used); * @man: TTM memory type manager * @tbo: TTM BO we need this range for * @place: placement flags and restrictions - * @mem: the resulting mem object + * @res: TTM memory object * * Dummy, just count the space used without allocating resources or any limit. */ @@ -85,7 +85,7 @@ static int amdgpu_preempt_mgr_new(struct ttm_resource_manager *man, * amdgpu_preempt_mgr_del - free ranges * * @man: TTM memory type manager - * @mem: TTM memory object + * @res: TTM memory object * * Free the allocated GTT again. */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index 9b41cb8c3de5..dee17a0e1187 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -46,6 +46,10 @@ static int psp_sysfs_init(struct amdgpu_device *adev); static void psp_sysfs_fini(struct amdgpu_device *adev); static int psp_load_smu_fw(struct psp_context *psp); +static int psp_ta_unload(struct psp_context *psp, struct ta_context *context); +static int psp_ta_load(struct psp_context *psp, struct ta_context *context); +static int psp_rap_terminate(struct psp_context *psp); +static int psp_securedisplay_terminate(struct psp_context *psp); /* * Due to DF Cstate management centralized to PMFW, the firmware @@ -61,23 +65,32 @@ static int psp_load_smu_fw(struct psp_context *psp); * * This new sequence is required for * - Arcturus and onwards - * - Navi12 and onwards */ static void psp_check_pmfw_centralized_cstate_management(struct psp_context *psp) { struct amdgpu_device *adev = psp->adev; - psp->pmfw_centralized_cstate_management = false; - - if (amdgpu_sriov_vf(adev)) - return; - - if (adev->flags & AMD_IS_APU) + if (amdgpu_sriov_vf(adev)) { + psp->pmfw_centralized_cstate_management = false; return; + } - if ((adev->asic_type >= CHIP_ARCTURUS) || - (adev->asic_type >= CHIP_NAVI12)) + switch (adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 4): + case IP_VERSION(11, 0, 5): + case IP_VERSION(11, 0, 7): + case IP_VERSION(11, 0, 9): + case IP_VERSION(11, 0, 11): + case IP_VERSION(11, 0, 12): + case IP_VERSION(11, 0, 13): + case IP_VERSION(13, 0, 2): psp->pmfw_centralized_cstate_management = true; + break; + default: + psp->pmfw_centralized_cstate_management = false; + break; + } } static int psp_early_init(void *handle) @@ -85,43 +98,45 @@ static int psp_early_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; struct psp_context *psp = &adev->psp; - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA12: + switch (adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(9, 0, 0): psp_v3_1_set_psp_funcs(psp); psp->autoload_supported = false; break; - case CHIP_RAVEN: + case IP_VERSION(10, 0, 0): + case IP_VERSION(10, 0, 1): psp_v10_0_set_psp_funcs(psp); psp->autoload_supported = false; break; - case CHIP_VEGA20: - case CHIP_ARCTURUS: + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 4): psp_v11_0_set_psp_funcs(psp); psp->autoload_supported = false; break; - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 5): + case IP_VERSION(11, 0, 9): + case IP_VERSION(11, 0, 7): + case IP_VERSION(11, 0, 11): + case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 0, 12): + case IP_VERSION(11, 0, 13): psp_v11_0_set_psp_funcs(psp); psp->autoload_supported = true; break; - case CHIP_RENOIR: + case IP_VERSION(11, 0, 3): + case IP_VERSION(12, 0, 1): psp_v12_0_set_psp_funcs(psp); break; - case CHIP_ALDEBARAN: + case IP_VERSION(13, 0, 2): psp_v13_0_set_psp_funcs(psp); break; - case CHIP_YELLOW_CARP: + case IP_VERSION(13, 0, 1): + case IP_VERSION(13, 0, 3): psp_v13_0_set_psp_funcs(psp); psp->autoload_supported = true; break; - case CHIP_CYAN_SKILLFISH: + case IP_VERSION(11, 0, 8): if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2) { psp_v11_0_8_set_psp_funcs(psp); psp->autoload_supported = false; @@ -264,7 +279,8 @@ static int psp_sw_init(void *handle) DRM_ERROR("Failed to load psp firmware!\n"); return ret; } - } else if (amdgpu_sriov_vf(adev) && adev->asic_type == CHIP_ALDEBARAN) { + } else if (amdgpu_sriov_vf(adev) && + adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 2)) { ret = psp_init_ta_microcode(psp, "aldebaran"); if (ret) { DRM_ERROR("Failed to initialize ta microcode!\n"); @@ -307,7 +323,8 @@ static int psp_sw_init(void *handle) } } - if (adev->asic_type == CHIP_NAVI10 || adev->asic_type == CHIP_SIENNA_CICHLID) { + if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 0) || + adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 7)) { ret= psp_sysfs_init(adev); if (ret) { return ret; @@ -337,8 +354,8 @@ static int psp_sw_fini(void *handle) psp->ta_fw = NULL; } - if (adev->asic_type == CHIP_NAVI10 || - adev->asic_type == CHIP_SIENNA_CICHLID) + if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 0) || + adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 7)) psp_sysfs_fini(adev); kfree(cmd); @@ -424,7 +441,7 @@ psp_cmd_submit_buf(struct psp_context *psp, if (psp->adev->no_hw_access) return 0; - if (!drm_dev_enter(&psp->adev->ddev, &idx)) + if (!drm_dev_enter(adev_to_drm(psp->adev), &idx)) return 0; memset(psp->cmd_buf_mem, 0, PSP_CMD_BUFFER_SIZE); @@ -501,7 +518,7 @@ static struct psp_gfx_cmd_resp *acquire_psp_cmd_buf(struct psp_context *psp) return cmd; } -void release_psp_cmd_buf(struct psp_context *psp) +static void release_psp_cmd_buf(struct psp_context *psp) { mutex_unlock(&psp->mutex); } @@ -597,10 +614,10 @@ static int psp_tmr_init(struct psp_context *psp) static bool psp_skip_tmr(struct psp_context *psp) { - switch (psp->adev->asic_type) { - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_ALDEBARAN: + switch (psp->adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(11, 0, 9): + case IP_VERSION(11, 0, 7): + case IP_VERSION(13, 0, 2): return true; default: return false; @@ -778,46 +795,29 @@ static int psp_rl_load(struct amdgpu_device *adev) return ret; } -static void psp_prep_asd_load_cmd_buf(struct psp_gfx_cmd_resp *cmd, - uint64_t asd_mc, uint32_t size) +static int psp_asd_load(struct psp_context *psp) { - cmd->cmd_id = GFX_CMD_ID_LOAD_ASD; - cmd->cmd.cmd_load_ta.app_phy_addr_lo = lower_32_bits(asd_mc); - cmd->cmd.cmd_load_ta.app_phy_addr_hi = upper_32_bits(asd_mc); - cmd->cmd.cmd_load_ta.app_len = size; - - cmd->cmd.cmd_load_ta.cmd_buf_phy_addr_lo = 0; - cmd->cmd.cmd_load_ta.cmd_buf_phy_addr_hi = 0; - cmd->cmd.cmd_load_ta.cmd_buf_len = 0; + return psp_ta_load(psp, &psp->asd_context); } -static int psp_asd_load(struct psp_context *psp) +static int psp_asd_initialize(struct psp_context *psp) { int ret; - struct psp_gfx_cmd_resp *cmd; /* If PSP version doesn't match ASD version, asd loading will be failed. * add workaround to bypass it for sriov now. * TODO: add version check to make it common */ - if (amdgpu_sriov_vf(psp->adev) || !psp->asd.size_bytes) + if (amdgpu_sriov_vf(psp->adev) || !psp->asd_context.bin_desc.size_bytes) return 0; - cmd = acquire_psp_cmd_buf(psp); - - psp_copy_fw(psp, psp->asd.start_addr, psp->asd.size_bytes); - - psp_prep_asd_load_cmd_buf(cmd, psp->fw_pri_mc_addr, - psp->asd.size_bytes); + psp->asd_context.mem_context.shared_mc_addr = 0; + psp->asd_context.mem_context.shared_mem_size = PSP_ASD_SHARED_MEM_SIZE; + psp->asd_context.ta_load_type = GFX_CMD_ID_LOAD_ASD; - ret = psp_cmd_submit_buf(psp, NULL, cmd, - psp->fence_buf_mc_addr); - if (!ret) { - psp->asd_context.asd_initialized = true; - psp->asd_context.session_id = cmd->resp.session_id; - } - - release_psp_cmd_buf(psp); + ret = psp_asd_load(psp); + if (!ret) + psp->asd_context.initialized = true; return ret; } @@ -829,27 +829,39 @@ static void psp_prep_ta_unload_cmd_buf(struct psp_gfx_cmd_resp *cmd, cmd->cmd.cmd_unload_ta.session_id = session_id; } +static int psp_ta_unload(struct psp_context *psp, struct ta_context *context) +{ + int ret; + struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp); + + psp_prep_ta_unload_cmd_buf(cmd, context->session_id); + + ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); + + release_psp_cmd_buf(psp); + + return ret; +} + static int psp_asd_unload(struct psp_context *psp) { + return psp_ta_unload(psp, &psp->asd_context); +} + +static int psp_asd_terminate(struct psp_context *psp) +{ int ret; - struct psp_gfx_cmd_resp *cmd; if (amdgpu_sriov_vf(psp->adev)) return 0; - if (!psp->asd_context.asd_initialized) + if (!psp->asd_context.initialized) return 0; - cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_unload_cmd_buf(cmd, psp->asd_context.session_id); + ret = psp_asd_unload(psp); - ret = psp_cmd_submit_buf(psp, NULL, cmd, - psp->fence_buf_mc_addr); if (!ret) - psp->asd_context.asd_initialized = false; - - release_psp_cmd_buf(psp); + psp->asd_context.initialized = false; return ret; } @@ -885,23 +897,22 @@ int psp_reg_program(struct psp_context *psp, enum psp_reg_prog_id reg, static void psp_prep_ta_load_cmd_buf(struct psp_gfx_cmd_resp *cmd, uint64_t ta_bin_mc, - uint32_t ta_bin_size, - uint64_t ta_shared_mc, - uint32_t ta_shared_size) + struct ta_context *context) { - cmd->cmd_id = GFX_CMD_ID_LOAD_TA; + cmd->cmd_id = context->ta_load_type; cmd->cmd.cmd_load_ta.app_phy_addr_lo = lower_32_bits(ta_bin_mc); cmd->cmd.cmd_load_ta.app_phy_addr_hi = upper_32_bits(ta_bin_mc); - cmd->cmd.cmd_load_ta.app_len = ta_bin_size; + cmd->cmd.cmd_load_ta.app_len = context->bin_desc.size_bytes; - cmd->cmd.cmd_load_ta.cmd_buf_phy_addr_lo = lower_32_bits(ta_shared_mc); - cmd->cmd.cmd_load_ta.cmd_buf_phy_addr_hi = upper_32_bits(ta_shared_mc); - cmd->cmd.cmd_load_ta.cmd_buf_len = ta_shared_size; + cmd->cmd.cmd_load_ta.cmd_buf_phy_addr_lo = + lower_32_bits(context->mem_context.shared_mc_addr); + cmd->cmd.cmd_load_ta.cmd_buf_phy_addr_hi = + upper_32_bits(context->mem_context.shared_mc_addr); + cmd->cmd.cmd_load_ta.cmd_buf_len = context->mem_context.shared_mem_size; } static int psp_ta_init_shared_buf(struct psp_context *psp, - struct ta_mem_context *mem_ctx, - uint32_t shared_mem_size) + struct ta_mem_context *mem_ctx) { int ret; @@ -909,8 +920,8 @@ static int psp_ta_init_shared_buf(struct psp_context *psp, * Allocate 16k memory aligned to 4k from Frame Buffer (local * physical) for ta to host memory */ - ret = amdgpu_bo_create_kernel(psp->adev, shared_mem_size, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, + ret = amdgpu_bo_create_kernel(psp->adev, mem_ctx->shared_mem_size, + PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM, &mem_ctx->shared_bo, &mem_ctx->shared_mc_addr, &mem_ctx->shared_buf); @@ -926,8 +937,7 @@ static void psp_ta_free_shared_buf(struct ta_mem_context *mem_ctx) static int psp_xgmi_init_shared_buf(struct psp_context *psp) { - return psp_ta_init_shared_buf(psp, &psp->xgmi_context.context.mem_context, - PSP_XGMI_SHARED_MEM_SIZE); + return psp_ta_init_shared_buf(psp, &psp->xgmi_context.context.mem_context); } static void psp_prep_ta_invoke_cmd_buf(struct psp_gfx_cmd_resp *cmd, @@ -941,12 +951,12 @@ static void psp_prep_ta_invoke_cmd_buf(struct psp_gfx_cmd_resp *cmd, static int psp_ta_invoke(struct psp_context *psp, uint32_t ta_cmd_id, - uint32_t session_id) + struct ta_context *context) { int ret; struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp); - psp_prep_ta_invoke_cmd_buf(cmd, ta_cmd_id, session_id); + psp_prep_ta_invoke_cmd_buf(cmd, ta_cmd_id, context->session_id); ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); @@ -956,31 +966,23 @@ static int psp_ta_invoke(struct psp_context *psp, return ret; } -static int psp_xgmi_load(struct psp_context *psp) +static int psp_ta_load(struct psp_context *psp, struct ta_context *context) { int ret; struct psp_gfx_cmd_resp *cmd; - /* - * TODO: bypass the loading in sriov for now - */ - cmd = acquire_psp_cmd_buf(psp); - psp_copy_fw(psp, psp->xgmi.start_addr, psp->xgmi.size_bytes); + psp_copy_fw(psp, context->bin_desc.start_addr, + context->bin_desc.size_bytes); - psp_prep_ta_load_cmd_buf(cmd, - psp->fw_pri_mc_addr, - psp->xgmi.size_bytes, - psp->xgmi_context.context.mem_context.shared_mc_addr, - PSP_XGMI_SHARED_MEM_SIZE); + psp_prep_ta_load_cmd_buf(cmd, psp->fw_pri_mc_addr, context); ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); if (!ret) { - psp->xgmi_context.context.initialized = true; - psp->xgmi_context.context.session_id = cmd->resp.session_id; + context->session_id = cmd->resp.session_id; } release_psp_cmd_buf(psp); @@ -988,41 +990,31 @@ static int psp_xgmi_load(struct psp_context *psp) return ret; } -static int psp_xgmi_unload(struct psp_context *psp) +static int psp_xgmi_load(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd; - struct amdgpu_device *adev = psp->adev; - - /* XGMI TA unload currently is not supported on Arcturus/Aldebaran A+A */ - if (adev->asic_type == CHIP_ARCTURUS || - (adev->asic_type == CHIP_ALDEBARAN && adev->gmc.xgmi.connected_to_cpu)) - return 0; - - /* - * TODO: bypass the unloading in sriov for now - */ - - cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_unload_cmd_buf(cmd, psp->xgmi_context.context.session_id); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, - psp->fence_buf_mc_addr); - - release_psp_cmd_buf(psp); + return psp_ta_load(psp, &psp->xgmi_context.context); +} - return ret; +static int psp_xgmi_unload(struct psp_context *psp) +{ + return psp_ta_unload(psp, &psp->xgmi_context.context); } int psp_xgmi_invoke(struct psp_context *psp, uint32_t ta_cmd_id) { - return psp_ta_invoke(psp, ta_cmd_id, psp->xgmi_context.context.session_id); + return psp_ta_invoke(psp, ta_cmd_id, &psp->xgmi_context.context); } int psp_xgmi_terminate(struct psp_context *psp) { int ret; + struct amdgpu_device *adev = psp->adev; + + /* XGMI TA unload currently is not supported on Arcturus/Aldebaran A+A */ + if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 4) || + (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 2) && + adev->gmc.xgmi.connected_to_cpu)) + return 0; if (!psp->xgmi_context.context.initialized) return 0; @@ -1045,13 +1037,16 @@ int psp_xgmi_initialize(struct psp_context *psp, bool set_extended_data, bool lo int ret; if (!psp->ta_fw || - !psp->xgmi.size_bytes || - !psp->xgmi.start_addr) + !psp->xgmi_context.context.bin_desc.size_bytes || + !psp->xgmi_context.context.bin_desc.start_addr) return -ENOENT; if (!load_ta) goto invoke; + psp->xgmi_context.context.mem_context.shared_mem_size = PSP_XGMI_SHARED_MEM_SIZE; + psp->xgmi_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA; + if (!psp->xgmi_context.context.initialized) { ret = psp_xgmi_init_shared_buf(psp); if (ret) @@ -1060,7 +1055,9 @@ int psp_xgmi_initialize(struct psp_context *psp, bool set_extended_data, bool lo /* Load XGMI TA */ ret = psp_xgmi_load(psp); - if (ret) + if (!ret) + psp->xgmi_context.context.initialized = true; + else return ret; invoke: @@ -1117,8 +1114,8 @@ int psp_xgmi_get_node_id(struct psp_context *psp, uint64_t *node_id) static bool psp_xgmi_peer_link_info_supported(struct psp_context *psp) { - return psp->adev->asic_type == CHIP_ALDEBARAN && - psp->xgmi.feature_version >= 0x2000000b; + return psp->adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 2) && + psp->xgmi_context.context.bin_desc.fw_version >= 0x2000000b; } /* @@ -1282,80 +1279,40 @@ int psp_xgmi_set_topology_info(struct psp_context *psp, // ras begin static int psp_ras_init_shared_buf(struct psp_context *psp) { - return psp_ta_init_shared_buf(psp, &psp->ras_context.context.mem_context, - PSP_RAS_SHARED_MEM_SIZE); + return psp_ta_init_shared_buf(psp, &psp->ras_context.context.mem_context); } static int psp_ras_load(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd; - struct ta_ras_shared_memory *ras_cmd; - - /* - * TODO: bypass the loading in sriov for now - */ - if (amdgpu_sriov_vf(psp->adev)) - return 0; - - psp_copy_fw(psp, psp->ras.start_addr, psp->ras.size_bytes); - - ras_cmd = (struct ta_ras_shared_memory *)psp->ras_context.context.mem_context.shared_buf; - - if (psp->adev->gmc.xgmi.connected_to_cpu) - ras_cmd->ras_in_message.init_flags.poison_mode_en = 1; - else - ras_cmd->ras_in_message.init_flags.dgpu_mode = 1; - - cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_load_cmd_buf(cmd, - psp->fw_pri_mc_addr, - psp->ras.size_bytes, - psp->ras_context.context.mem_context.shared_mc_addr, - PSP_RAS_SHARED_MEM_SIZE); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, - psp->fence_buf_mc_addr); - - if (!ret) { - psp->ras_context.context.session_id = cmd->resp.session_id; - - if (!ras_cmd->ras_status) - psp->ras_context.context.initialized = true; - else - dev_warn(psp->adev->dev, "RAS Init Status: 0x%X\n", ras_cmd->ras_status); - } - - release_psp_cmd_buf(psp); - - if (ret || ras_cmd->ras_status) - amdgpu_ras_fini(psp->adev); - - return ret; + return psp_ta_load(psp, &psp->ras_context.context); } static int psp_ras_unload(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd; - - /* - * TODO: bypass the unloading in sriov for now - */ - if (amdgpu_sriov_vf(psp->adev)) - return 0; - - cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_unload_cmd_buf(cmd, psp->ras_context.context.session_id); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, - psp->fence_buf_mc_addr); + return psp_ta_unload(psp, &psp->ras_context.context); +} - release_psp_cmd_buf(psp); +static void psp_ras_ta_check_status(struct psp_context *psp) +{ + struct ta_ras_shared_memory *ras_cmd = + (struct ta_ras_shared_memory *)psp->ras_context.context.mem_context.shared_buf; - return ret; + switch (ras_cmd->ras_status) { + case TA_RAS_STATUS__ERROR_UNSUPPORTED_IP: + dev_warn(psp->adev->dev, + "RAS WARNING: cmd failed due to unsupported ip\n"); + break; + case TA_RAS_STATUS__ERROR_UNSUPPORTED_ERROR_INJ: + dev_warn(psp->adev->dev, + "RAS WARNING: cmd failed due to unsupported error injection\n"); + break; + case TA_RAS_STATUS__SUCCESS: + break; + default: + dev_warn(psp->adev->dev, + "RAS WARNING: ras status = 0x%X\n", ras_cmd->ras_status); + break; + } } int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id) @@ -1371,7 +1328,7 @@ int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id) if (amdgpu_sriov_vf(psp->adev)) return 0; - ret = psp_ta_invoke(psp, ta_cmd_id, psp->ras_context.context.session_id); + ret = psp_ta_invoke(psp, ta_cmd_id, &psp->ras_context.context); if (amdgpu_ras_intr_triggered()) return ret; @@ -1391,31 +1348,8 @@ int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id) else if (ras_cmd->ras_out_message.flags.reg_access_failure_flag) dev_warn(psp->adev->dev, "RAS internal register access blocked\n"); - } - return ret; -} - -static int psp_ras_status_to_errno(struct amdgpu_device *adev, - enum ta_ras_status ras_status) -{ - int ret = -EINVAL; - - switch (ras_status) { - case TA_RAS_STATUS__SUCCESS: - ret = 0; - break; - case TA_RAS_STATUS__RESET_NEEDED: - ret = -EAGAIN; - break; - case TA_RAS_STATUS__ERROR_RAS_NOT_AVAILABLE: - dev_warn(adev->dev, "RAS WARN: ras function unavailable\n"); - break; - case TA_RAS_STATUS__ERROR_ASD_READ_WRITE: - dev_warn(adev->dev, "RAS WARN: asd read or write failed\n"); - break; - default: - dev_err(adev->dev, "RAS ERROR: ras function failed ret 0x%X\n", ret); + psp_ras_ta_check_status(psp); } return ret; @@ -1444,7 +1378,7 @@ int psp_ras_enable_features(struct psp_context *psp, if (ret) return -EINVAL; - return psp_ras_status_to_errno(psp->adev, ras_cmd->ras_status); + return 0; } static int psp_ras_terminate(struct psp_context *psp) @@ -1477,6 +1411,7 @@ static int psp_ras_initialize(struct psp_context *psp) int ret; uint32_t boot_cfg = 0xFF; struct amdgpu_device *adev = psp->adev; + struct ta_ras_shared_memory *ras_cmd; /* * TODO: bypass the initialize in sriov for now @@ -1484,8 +1419,8 @@ static int psp_ras_initialize(struct psp_context *psp) if (amdgpu_sriov_vf(adev)) return 0; - if (!adev->psp.ras.size_bytes || - !adev->psp.ras.start_addr) { + if (!adev->psp.ras_context.context.bin_desc.size_bytes || + !adev->psp.ras_context.context.bin_desc.start_addr) { dev_info(adev->dev, "RAS: optional ras ta ucode is not available\n"); return 0; } @@ -1531,17 +1466,34 @@ static int psp_ras_initialize(struct psp_context *psp) } } + psp->ras_context.context.mem_context.shared_mem_size = PSP_RAS_SHARED_MEM_SIZE; + psp->ras_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA; + if (!psp->ras_context.context.initialized) { ret = psp_ras_init_shared_buf(psp); if (ret) return ret; } + ras_cmd = (struct ta_ras_shared_memory *)psp->ras_context.context.mem_context.shared_buf; + memset(ras_cmd, 0, sizeof(struct ta_ras_shared_memory)); + + if (amdgpu_ras_is_poison_mode_supported(adev)) + ras_cmd->ras_in_message.init_flags.poison_mode_en = 1; + if (!adev->gmc.xgmi.connected_to_cpu) + ras_cmd->ras_in_message.init_flags.dgpu_mode = 1; + ret = psp_ras_load(psp); - if (ret) - return ret; - return 0; + if (!ret && !ras_cmd->ras_status) + psp->ras_context.context.initialized = true; + else { + if (ras_cmd->ras_status) + dev_warn(psp->adev->dev, "RAS Init Status: 0x%X\n", ras_cmd->ras_status); + amdgpu_ras_fini(psp->adev); + } + + return ret; } int psp_ras_trigger_error(struct psp_context *psp, @@ -1568,51 +1520,24 @@ int psp_ras_trigger_error(struct psp_context *psp, if (amdgpu_ras_intr_triggered()) return 0; - return psp_ras_status_to_errno(psp->adev, ras_cmd->ras_status); + if (ras_cmd->ras_status) + return -EINVAL; + + return 0; } // ras end // HDCP start static int psp_hdcp_init_shared_buf(struct psp_context *psp) { - return psp_ta_init_shared_buf(psp, &psp->hdcp_context.context.mem_context, - PSP_HDCP_SHARED_MEM_SIZE); + return psp_ta_init_shared_buf(psp, &psp->hdcp_context.context.mem_context); } static int psp_hdcp_load(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd; - - /* - * TODO: bypass the loading in sriov for now - */ - if (amdgpu_sriov_vf(psp->adev)) - return 0; - - psp_copy_fw(psp, psp->hdcp.start_addr, - psp->hdcp.size_bytes); - - cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_load_cmd_buf(cmd, - psp->fw_pri_mc_addr, - psp->hdcp.size_bytes, - psp->hdcp_context.context.mem_context.shared_mc_addr, - PSP_HDCP_SHARED_MEM_SIZE); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); - - if (!ret) { - psp->hdcp_context.context.initialized = true; - psp->hdcp_context.context.session_id = cmd->resp.session_id; - mutex_init(&psp->hdcp_context.mutex); - } - - release_psp_cmd_buf(psp); - - return ret; + return psp_ta_load(psp, &psp->hdcp_context.context); } + static int psp_hdcp_initialize(struct psp_context *psp) { int ret; @@ -1623,12 +1548,15 @@ static int psp_hdcp_initialize(struct psp_context *psp) if (amdgpu_sriov_vf(psp->adev)) return 0; - if (!psp->hdcp.size_bytes || - !psp->hdcp.start_addr) { + if (!psp->hdcp_context.context.bin_desc.size_bytes || + !psp->hdcp_context.context.bin_desc.start_addr) { dev_info(psp->adev->dev, "HDCP: optional hdcp ta ucode is not available\n"); return 0; } + psp->hdcp_context.context.mem_context.shared_mem_size = PSP_HDCP_SHARED_MEM_SIZE; + psp->hdcp_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA; + if (!psp->hdcp_context.context.initialized) { ret = psp_hdcp_init_shared_buf(psp); if (ret) @@ -1636,32 +1564,17 @@ static int psp_hdcp_initialize(struct psp_context *psp) } ret = psp_hdcp_load(psp); - if (ret) - return ret; + if (!ret) { + psp->hdcp_context.context.initialized = true; + mutex_init(&psp->hdcp_context.mutex); + } - return 0; + return ret; } static int psp_hdcp_unload(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd; - - /* - * TODO: bypass the unloading in sriov for now - */ - if (amdgpu_sriov_vf(psp->adev)) - return 0; - - cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_unload_cmd_buf(cmd, psp->hdcp_context.context.session_id); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); - - release_psp_cmd_buf(psp); - - return ret; + return psp_ta_unload(psp, &psp->hdcp_context.context); } int psp_hdcp_invoke(struct psp_context *psp, uint32_t ta_cmd_id) @@ -1672,7 +1585,7 @@ int psp_hdcp_invoke(struct psp_context *psp, uint32_t ta_cmd_id) if (amdgpu_sriov_vf(psp->adev)) return 0; - return psp_ta_invoke(psp, ta_cmd_id, psp->hdcp_context.context.session_id); + return psp_ta_invoke(psp, ta_cmd_id, &psp->hdcp_context.context); } static int psp_hdcp_terminate(struct psp_context *psp) @@ -1709,42 +1622,12 @@ out: // DTM start static int psp_dtm_init_shared_buf(struct psp_context *psp) { - return psp_ta_init_shared_buf(psp, &psp->dtm_context.context.mem_context, - PSP_DTM_SHARED_MEM_SIZE); + return psp_ta_init_shared_buf(psp, &psp->dtm_context.context.mem_context); } static int psp_dtm_load(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd; - - /* - * TODO: bypass the loading in sriov for now - */ - if (amdgpu_sriov_vf(psp->adev)) - return 0; - - psp_copy_fw(psp, psp->dtm.start_addr, psp->dtm.size_bytes); - - cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_load_cmd_buf(cmd, - psp->fw_pri_mc_addr, - psp->dtm.size_bytes, - psp->dtm_context.context.mem_context.shared_mc_addr, - PSP_DTM_SHARED_MEM_SIZE); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); - - if (!ret) { - psp->dtm_context.context.initialized = true; - psp->dtm_context.context.session_id = cmd->resp.session_id; - mutex_init(&psp->dtm_context.mutex); - } - - release_psp_cmd_buf(psp); - - return ret; + return psp_ta_load(psp, &psp->dtm_context.context); } static int psp_dtm_initialize(struct psp_context *psp) @@ -1757,12 +1640,15 @@ static int psp_dtm_initialize(struct psp_context *psp) if (amdgpu_sriov_vf(psp->adev)) return 0; - if (!psp->dtm.size_bytes || - !psp->dtm.start_addr) { + if (!psp->dtm_context.context.bin_desc.size_bytes || + !psp->dtm_context.context.bin_desc.start_addr) { dev_info(psp->adev->dev, "DTM: optional dtm ta ucode is not available\n"); return 0; } + psp->dtm_context.context.mem_context.shared_mem_size = PSP_DTM_SHARED_MEM_SIZE; + psp->dtm_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA; + if (!psp->dtm_context.context.initialized) { ret = psp_dtm_init_shared_buf(psp); if (ret) @@ -1770,32 +1656,17 @@ static int psp_dtm_initialize(struct psp_context *psp) } ret = psp_dtm_load(psp); - if (ret) - return ret; + if (!ret) { + psp->dtm_context.context.initialized = true; + mutex_init(&psp->dtm_context.mutex); + } - return 0; + return ret; } static int psp_dtm_unload(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd; - - /* - * TODO: bypass the unloading in sriov for now - */ - if (amdgpu_sriov_vf(psp->adev)) - return 0; - - cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_unload_cmd_buf(cmd, psp->dtm_context.context.session_id); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); - - release_psp_cmd_buf(psp); - - return ret; + return psp_ta_unload(psp, &psp->dtm_context.context); } int psp_dtm_invoke(struct psp_context *psp, uint32_t ta_cmd_id) @@ -1806,7 +1677,7 @@ int psp_dtm_invoke(struct psp_context *psp, uint32_t ta_cmd_id) if (amdgpu_sriov_vf(psp->adev)) return 0; - return psp_ta_invoke(psp, ta_cmd_id, psp->dtm_context.context.session_id); + return psp_ta_invoke(psp, ta_cmd_id, &psp->dtm_context.context); } static int psp_dtm_terminate(struct psp_context *psp) @@ -1843,50 +1714,17 @@ out: // RAP start static int psp_rap_init_shared_buf(struct psp_context *psp) { - return psp_ta_init_shared_buf(psp, &psp->rap_context.context.mem_context, - PSP_RAP_SHARED_MEM_SIZE); + return psp_ta_init_shared_buf(psp, &psp->rap_context.context.mem_context); } static int psp_rap_load(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd; - - psp_copy_fw(psp, psp->rap.start_addr, psp->rap.size_bytes); - - cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_load_cmd_buf(cmd, - psp->fw_pri_mc_addr, - psp->rap.size_bytes, - psp->rap_context.context.mem_context.shared_mc_addr, - PSP_RAP_SHARED_MEM_SIZE); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); - - if (!ret) { - psp->rap_context.context.initialized = true; - psp->rap_context.context.session_id = cmd->resp.session_id; - mutex_init(&psp->rap_context.mutex); - } - - release_psp_cmd_buf(psp); - - return ret; + return psp_ta_load(psp, &psp->rap_context.context); } static int psp_rap_unload(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_unload_cmd_buf(cmd, psp->rap_context.context.session_id); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); - - release_psp_cmd_buf(psp); - - return ret; + return psp_ta_unload(psp, &psp->rap_context.context); } static int psp_rap_initialize(struct psp_context *psp) @@ -1900,12 +1738,15 @@ static int psp_rap_initialize(struct psp_context *psp) if (amdgpu_sriov_vf(psp->adev)) return 0; - if (!psp->rap.size_bytes || - !psp->rap.start_addr) { + if (!psp->rap_context.context.bin_desc.size_bytes || + !psp->rap_context.context.bin_desc.start_addr) { dev_info(psp->adev->dev, "RAP: optional rap ta ucode is not available\n"); return 0; } + psp->rap_context.context.mem_context.shared_mem_size = PSP_RAP_SHARED_MEM_SIZE; + psp->rap_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA; + if (!psp->rap_context.context.initialized) { ret = psp_rap_init_shared_buf(psp); if (ret) @@ -1913,16 +1754,15 @@ static int psp_rap_initialize(struct psp_context *psp) } ret = psp_rap_load(psp); - if (ret) + if (!ret) { + psp->rap_context.context.initialized = true; + mutex_init(&psp->rap_context.mutex); + } else return ret; ret = psp_rap_invoke(psp, TA_CMD_RAP__INITIALIZE, &status); if (ret || status != TA_RAP_STATUS__SUCCESS) { - psp_rap_unload(psp); - - psp_ta_free_shared_buf(&psp->rap_context.context.mem_context); - - psp->rap_context.context.initialized = false; + psp_rap_terminate(psp); dev_warn(psp->adev->dev, "RAP TA initialize fail (%d) status %d.\n", ret, status); @@ -1971,7 +1811,7 @@ int psp_rap_invoke(struct psp_context *psp, uint32_t ta_cmd_id, enum ta_rap_stat rap_cmd->cmd_id = ta_cmd_id; rap_cmd->validation_method_id = METHOD_A; - ret = psp_ta_invoke(psp, rap_cmd->cmd_id, psp->rap_context.context.session_id); + ret = psp_ta_invoke(psp, rap_cmd->cmd_id, &psp->rap_context.context); if (ret) goto out_unlock; @@ -1989,49 +1829,17 @@ out_unlock: static int psp_securedisplay_init_shared_buf(struct psp_context *psp) { return psp_ta_init_shared_buf( - psp, &psp->securedisplay_context.context.mem_context, - PSP_SECUREDISPLAY_SHARED_MEM_SIZE); + psp, &psp->securedisplay_context.context.mem_context); } static int psp_securedisplay_load(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp); - - memset(psp->fw_pri_buf, 0, PSP_1_MEG); - memcpy(psp->fw_pri_buf, psp->securedisplay.start_addr, psp->securedisplay.size_bytes); - - psp_prep_ta_load_cmd_buf(cmd, - psp->fw_pri_mc_addr, - psp->securedisplay.size_bytes, - psp->securedisplay_context.context.mem_context.shared_mc_addr, - PSP_SECUREDISPLAY_SHARED_MEM_SIZE); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); - - if (!ret) { - psp->securedisplay_context.context.initialized = true; - psp->securedisplay_context.context.session_id = cmd->resp.session_id; - mutex_init(&psp->securedisplay_context.mutex); - } - - release_psp_cmd_buf(psp); - - return ret; + return psp_ta_load(psp, &psp->securedisplay_context.context); } static int psp_securedisplay_unload(struct psp_context *psp) { - int ret; - struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp); - - psp_prep_ta_unload_cmd_buf(cmd, psp->securedisplay_context.context.session_id); - - ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); - - release_psp_cmd_buf(psp); - - return ret; + return psp_ta_unload(psp, &psp->securedisplay_context.context); } static int psp_securedisplay_initialize(struct psp_context *psp) @@ -2045,12 +1853,16 @@ static int psp_securedisplay_initialize(struct psp_context *psp) if (amdgpu_sriov_vf(psp->adev)) return 0; - if (!psp->securedisplay.size_bytes || - !psp->securedisplay.start_addr) { + if (!psp->securedisplay_context.context.bin_desc.size_bytes || + !psp->securedisplay_context.context.bin_desc.start_addr) { dev_info(psp->adev->dev, "SECUREDISPLAY: securedisplay ta ucode is not available\n"); return 0; } + psp->securedisplay_context.context.mem_context.shared_mem_size = + PSP_SECUREDISPLAY_SHARED_MEM_SIZE; + psp->securedisplay_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA; + if (!psp->securedisplay_context.context.initialized) { ret = psp_securedisplay_init_shared_buf(psp); if (ret) @@ -2058,7 +1870,10 @@ static int psp_securedisplay_initialize(struct psp_context *psp) } ret = psp_securedisplay_load(psp); - if (ret) + if (!ret) { + psp->securedisplay_context.context.initialized = true; + mutex_init(&psp->securedisplay_context.mutex); + } else return ret; psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd, @@ -2066,12 +1881,7 @@ static int psp_securedisplay_initialize(struct psp_context *psp) ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__QUERY_TA); if (ret) { - psp_securedisplay_unload(psp); - - psp_ta_free_shared_buf(&psp->securedisplay_context.context.mem_context); - - psp->securedisplay_context.context.initialized = false; - + psp_securedisplay_terminate(psp); dev_err(psp->adev->dev, "SECUREDISPLAY TA initialize fail.\n"); return -EINVAL; } @@ -2123,7 +1933,7 @@ int psp_securedisplay_invoke(struct psp_context *psp, uint32_t ta_cmd_id) mutex_lock(&psp->securedisplay_context.mutex); - ret = psp_ta_invoke(psp, ta_cmd_id, psp->securedisplay_context.context.session_id); + ret = psp_ta_invoke(psp, ta_cmd_id, &psp->securedisplay_context.context); mutex_unlock(&psp->securedisplay_context.mutex); @@ -2207,12 +2017,16 @@ static int psp_hw_start(struct psp_context *psp) return ret; } + if (amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev)) + goto skip_pin_bo; + ret = psp_tmr_init(psp); if (ret) { DRM_ERROR("PSP tmr init failed!\n"); return ret; } +skip_pin_bo: /* * For ASICs with DF Cstate management centralized * to PMFW, TMR setup should be performed after PMFW @@ -2443,8 +2257,8 @@ static int psp_load_smu_fw(struct psp_context *psp) if ((amdgpu_in_reset(adev) && ras && adev->ras_enabled && - (adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_VEGA20))) { + (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 4) || + adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 2)))) { ret = amdgpu_dpm_set_mp1_state(adev, PP_MP1_STATE_UNLOAD); if (ret) { DRM_WARN("Failed to set MP1 state prepare for reload\n"); @@ -2541,8 +2355,9 @@ static int psp_load_non_psp_fw(struct psp_context *psp) continue; if (psp->autoload_supported && - (adev->asic_type >= CHIP_SIENNA_CICHLID && - adev->asic_type <= CHIP_DIMGREY_CAVEFISH) && + (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 7) || + adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 11) || + adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 12)) && (ucode->ucode_id == AMDGPU_UCODE_ID_SDMA1 || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA2 || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA3)) @@ -2629,7 +2444,7 @@ skip_memalloc: if (ret) goto failed; - ret = psp_asd_load(psp); + ret = psp_asd_initialize(psp); if (ret) { DRM_ERROR("PSP load asd failed!\n"); return ret; @@ -2641,6 +2456,18 @@ skip_memalloc: return ret; } + if (amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev)) { + if (adev->gmc.xgmi.num_physical_nodes > 1) { + ret = psp_xgmi_initialize(psp, false, true); + /* Warning the XGMI seesion initialize failure + * Instead of stop driver initialization + */ + if (ret) + dev_err(psp->adev->dev, + "XGMI: Failed to initialize XGMI session\n"); + } + } + if (psp->ta_fw) { ret = psp_ras_initialize(psp); if (ret) @@ -2721,7 +2548,7 @@ static int psp_hw_fini(void *handle) psp_hdcp_terminate(psp); } - psp_asd_unload(psp); + psp_asd_terminate(psp); psp_tmr_terminate(psp); psp_ring_destroy(psp, PSP_RING_TYPE__KM); @@ -2779,9 +2606,9 @@ static int psp_suspend(void *handle) } } - ret = psp_asd_unload(psp); + ret = psp_asd_terminate(psp); if (ret) { - DRM_ERROR("Failed to unload asd\n"); + DRM_ERROR("Failed to terminate asd\n"); return ret; } @@ -2826,12 +2653,18 @@ static int psp_resume(void *handle) if (ret) goto failed; - ret = psp_asd_load(psp); + ret = psp_asd_initialize(psp); if (ret) { DRM_ERROR("PSP load asd failed!\n"); goto failed; } + ret = psp_rl_load(adev); + if (ret) { + dev_err(adev->dev, "PSP load RL failed!\n"); + goto failed; + } + if (adev->gmc.xgmi.num_physical_nodes > 1) { ret = psp_xgmi_initialize(psp, false, true); /* Warning the XGMI seesion initialize failure @@ -2994,10 +2827,10 @@ int psp_init_asd_microcode(struct psp_context *psp, goto out; asd_hdr = (const struct psp_firmware_header_v1_0 *)adev->psp.asd_fw->data; - adev->psp.asd.fw_version = le32_to_cpu(asd_hdr->header.ucode_version); - adev->psp.asd.feature_version = le32_to_cpu(asd_hdr->sos.fw_version); - adev->psp.asd.size_bytes = le32_to_cpu(asd_hdr->header.ucode_size_bytes); - adev->psp.asd.start_addr = (uint8_t *)asd_hdr + + adev->psp.asd_context.bin_desc.fw_version = le32_to_cpu(asd_hdr->header.ucode_version); + adev->psp.asd_context.bin_desc.feature_version = le32_to_cpu(asd_hdr->sos.fw_version); + adev->psp.asd_context.bin_desc.size_bytes = le32_to_cpu(asd_hdr->header.ucode_size_bytes); + adev->psp.asd_context.bin_desc.start_addr = (uint8_t *)asd_hdr + le32_to_cpu(asd_hdr->header.ucode_array_offset_bytes); return 0; out: @@ -3129,7 +2962,8 @@ static int psp_init_sos_base_fw(struct amdgpu_device *adev) ucode_array_start_addr = (uint8_t *)sos_hdr + le32_to_cpu(sos_hdr->header.ucode_array_offset_bytes); - if (adev->gmc.xgmi.connected_to_cpu || (adev->asic_type != CHIP_ALDEBARAN)) { + if (adev->gmc.xgmi.connected_to_cpu || + (adev->ip_versions[MP0_HWIP][0] != IP_VERSION(13, 0, 2))) { adev->psp.sos.fw_version = le32_to_cpu(sos_hdr->header.ucode_version); adev->psp.sos.feature_version = le32_to_cpu(sos_hdr->sos.fw_version); @@ -3284,40 +3118,43 @@ static int parse_ta_bin_descriptor(struct psp_context *psp, switch (desc->fw_type) { case TA_FW_TYPE_PSP_ASD: - psp->asd.fw_version = le32_to_cpu(desc->fw_version); - psp->asd.feature_version = le32_to_cpu(desc->fw_version); - psp->asd.size_bytes = le32_to_cpu(desc->size_bytes); - psp->asd.start_addr = ucode_start_addr; + psp->asd_context.bin_desc.fw_version = le32_to_cpu(desc->fw_version); + psp->asd_context.bin_desc.feature_version = le32_to_cpu(desc->fw_version); + psp->asd_context.bin_desc.size_bytes = le32_to_cpu(desc->size_bytes); + psp->asd_context.bin_desc.start_addr = ucode_start_addr; break; case TA_FW_TYPE_PSP_XGMI: - psp->xgmi.feature_version = le32_to_cpu(desc->fw_version); - psp->xgmi.size_bytes = le32_to_cpu(desc->size_bytes); - psp->xgmi.start_addr = ucode_start_addr; + psp->xgmi_context.context.bin_desc.fw_version = le32_to_cpu(desc->fw_version); + psp->xgmi_context.context.bin_desc.size_bytes = le32_to_cpu(desc->size_bytes); + psp->xgmi_context.context.bin_desc.start_addr = ucode_start_addr; break; case TA_FW_TYPE_PSP_RAS: - psp->ras.feature_version = le32_to_cpu(desc->fw_version); - psp->ras.size_bytes = le32_to_cpu(desc->size_bytes); - psp->ras.start_addr = ucode_start_addr; + psp->ras_context.context.bin_desc.fw_version = le32_to_cpu(desc->fw_version); + psp->ras_context.context.bin_desc.size_bytes = le32_to_cpu(desc->size_bytes); + psp->ras_context.context.bin_desc.start_addr = ucode_start_addr; break; case TA_FW_TYPE_PSP_HDCP: - psp->hdcp.feature_version = le32_to_cpu(desc->fw_version); - psp->hdcp.size_bytes = le32_to_cpu(desc->size_bytes); - psp->hdcp.start_addr = ucode_start_addr; + psp->hdcp_context.context.bin_desc.fw_version = le32_to_cpu(desc->fw_version); + psp->hdcp_context.context.bin_desc.size_bytes = le32_to_cpu(desc->size_bytes); + psp->hdcp_context.context.bin_desc.start_addr = ucode_start_addr; break; case TA_FW_TYPE_PSP_DTM: - psp->dtm.feature_version = le32_to_cpu(desc->fw_version); - psp->dtm.size_bytes = le32_to_cpu(desc->size_bytes); - psp->dtm.start_addr = ucode_start_addr; + psp->dtm_context.context.bin_desc.fw_version = le32_to_cpu(desc->fw_version); + psp->dtm_context.context.bin_desc.size_bytes = le32_to_cpu(desc->size_bytes); + psp->dtm_context.context.bin_desc.start_addr = ucode_start_addr; break; case TA_FW_TYPE_PSP_RAP: - psp->rap.feature_version = le32_to_cpu(desc->fw_version); - psp->rap.size_bytes = le32_to_cpu(desc->size_bytes); - psp->rap.start_addr = ucode_start_addr; + psp->rap_context.context.bin_desc.fw_version = le32_to_cpu(desc->fw_version); + psp->rap_context.context.bin_desc.size_bytes = le32_to_cpu(desc->size_bytes); + psp->rap_context.context.bin_desc.start_addr = ucode_start_addr; break; case TA_FW_TYPE_PSP_SECUREDISPLAY: - psp->securedisplay.feature_version = le32_to_cpu(desc->fw_version); - psp->securedisplay.size_bytes = le32_to_cpu(desc->size_bytes); - psp->securedisplay.start_addr = ucode_start_addr; + psp->securedisplay_context.context.bin_desc.fw_version = + le32_to_cpu(desc->fw_version); + psp->securedisplay_context.context.bin_desc.size_bytes = + le32_to_cpu(desc->size_bytes); + psp->securedisplay_context.context.bin_desc.start_addr = + ucode_start_addr; break; default: dev_warn(psp->adev->dev, "Unsupported TA type: %d\n", desc->fw_type); @@ -3478,7 +3315,7 @@ void psp_copy_fw(struct psp_context *psp, uint8_t *start_addr, uint32_t bin_size { int idx; - if (!drm_dev_enter(&psp->adev->ddev, &idx)) + if (!drm_dev_enter(adev_to_drm(psp->adev), &idx)) return; memset(psp->fw_pri_buf, 0, PSP_1_MEG); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h index 8ef2d28af92a..f29afabbff1f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h @@ -34,17 +34,20 @@ #define PSP_FENCE_BUFFER_SIZE 0x1000 #define PSP_CMD_BUFFER_SIZE 0x1000 -#define PSP_XGMI_SHARED_MEM_SIZE 0x4000 -#define PSP_RAS_SHARED_MEM_SIZE 0x4000 #define PSP_1_MEG 0x100000 #define PSP_TMR_SIZE(adev) ((adev)->asic_type == CHIP_ALDEBARAN ? 0x800000 : 0x400000) -#define PSP_HDCP_SHARED_MEM_SIZE 0x4000 -#define PSP_DTM_SHARED_MEM_SIZE 0x4000 -#define PSP_RAP_SHARED_MEM_SIZE 0x4000 -#define PSP_SECUREDISPLAY_SHARED_MEM_SIZE 0x4000 -#define PSP_SHARED_MEM_SIZE 0x4000 #define PSP_FW_NAME_LEN 0x24 +enum psp_shared_mem_size { + PSP_ASD_SHARED_MEM_SIZE = 0x0, + PSP_XGMI_SHARED_MEM_SIZE = 0x4000, + PSP_RAS_SHARED_MEM_SIZE = 0x4000, + PSP_HDCP_SHARED_MEM_SIZE = 0x4000, + PSP_DTM_SHARED_MEM_SIZE = 0x4000, + PSP_RAP_SHARED_MEM_SIZE = 0x4000, + PSP_SECUREDISPLAY_SHARED_MEM_SIZE = 0x4000, +}; + struct psp_context; struct psp_xgmi_node_info; struct psp_xgmi_topology_info; @@ -131,21 +134,26 @@ struct psp_xgmi_topology_info { struct psp_xgmi_node_info nodes[AMDGPU_XGMI_MAX_CONNECTED_NODES]; }; -struct psp_asd_context { - bool asd_initialized; - uint32_t session_id; +struct psp_bin_desc { + uint32_t fw_version; + uint32_t feature_version; + uint32_t size_bytes; + uint8_t *start_addr; }; struct ta_mem_context { struct amdgpu_bo *shared_bo; uint64_t shared_mc_addr; void *shared_buf; + enum psp_shared_mem_size shared_mem_size; }; struct ta_context { bool initialized; uint32_t session_id; struct ta_mem_context mem_context; + struct psp_bin_desc bin_desc; + enum psp_gfx_cmd_id ta_load_type; }; struct ta_cp_context { @@ -263,13 +271,6 @@ struct psp_runtime_boot_cfg_entry { uint32_t reserved; }; -struct psp_bin_desc { - uint32_t fw_version; - uint32_t feature_version; - uint32_t size_bytes; - uint8_t *start_addr; -}; - struct psp_context { struct amdgpu_device *adev; @@ -301,7 +302,6 @@ struct psp_context /* asd firmware */ const struct firmware *asd_fw; - struct psp_bin_desc asd; /* toc firmware */ const struct firmware *toc_fw; @@ -326,14 +326,8 @@ struct psp_context /* xgmi ta firmware and buffer */ const struct firmware *ta_fw; uint32_t ta_fw_version; - struct psp_bin_desc xgmi; - struct psp_bin_desc ras; - struct psp_bin_desc hdcp; - struct psp_bin_desc dtm; - struct psp_bin_desc rap; - struct psp_bin_desc securedisplay; - - struct psp_asd_context asd_context; + + struct ta_context asd_context; struct psp_xgmi_context xgmi_context; struct psp_ras_context ras_context; struct ta_cp_context hdcp_context; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 96a8fd0ca1df..91e6e87562ac 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -35,7 +35,11 @@ #include "amdgpu_xgmi.h" #include "ivsrcid/nbio/irqsrcs_nbif_7_4.h" #include "atom.h" +#ifdef CONFIG_X86_MCE_AMD +#include <asm/mce.h> +static bool notifier_registered; +#endif static const char *RAS_FS_NAME = "ras"; const char *ras_error_string[] = { @@ -61,8 +65,30 @@ const char *ras_block_string[] = { "mp0", "mp1", "fuse", + "mca", }; +const char *ras_mca_block_string[] = { + "mca_mp0", + "mca_mp1", + "mca_mpio", + "mca_iohc", +}; + +const char *get_ras_block_str(struct ras_common_if *ras_block) +{ + if (!ras_block) + return "NULL"; + + if (ras_block->block >= AMDGPU_RAS_BLOCK_COUNT) + return "OUT OF RANGE"; + + if (ras_block->block == AMDGPU_RAS_BLOCK__MCA) + return ras_mca_block_string[ras_block->sub_block_index]; + + return ras_block_string[ras_block->block]; +} + #define ras_err_str(i) (ras_error_string[ffs(i)]) #define RAS_DEFAULT_FLAGS (AMDGPU_RAS_FLAG_INIT_BY_VBIOS) @@ -85,6 +111,14 @@ static bool amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con, uint64_t addr); static bool amdgpu_ras_check_bad_page(struct amdgpu_device *adev, uint64_t addr); +#ifdef CONFIG_X86_MCE_AMD +static void amdgpu_register_bad_pages_mca_notifier(struct amdgpu_device *adev); +struct mce_notifier_adev_list { + struct amdgpu_device *devs[MAX_GPU_INSTANCE]; + int num_gpu; +}; +static struct mce_notifier_adev_list mce_adev_list; +#endif void amdgpu_ras_set_error_query_ready(struct amdgpu_device *adev, bool ready) { @@ -187,7 +221,7 @@ static int amdgpu_ras_find_block_id_by_name(const char *name, int *block_id) for (i = 0; i < ARRAY_SIZE(ras_block_string); i++) { *block_id = i; - if (strcmp(name, ras_block_str(i)) == 0) + if (strcmp(name, ras_block_string[i]) == 0) return 0; } return -EINVAL; @@ -509,7 +543,6 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev, if (amdgpu_ras_query_error_status(obj->adev, &info)) return -EINVAL; - if (obj->adev->asic_type == CHIP_ALDEBARAN) { if (amdgpu_ras_reset_error_status(obj->adev, info.head.block)) DRM_WARN("Failed to reset error counter and error status"); @@ -529,7 +562,7 @@ static inline void put_obj(struct ras_manager *obj) if (obj && (--obj->use == 0)) list_del(&obj->node); if (obj && (obj->use < 0)) - DRM_ERROR("RAS ERROR: Unbalance obj(%s) use\n", ras_block_str(obj->head.block)); + DRM_ERROR("RAS ERROR: Unbalance obj(%s) use\n", get_ras_block_str(&obj->head)); } /* make one obj and return it. */ @@ -545,7 +578,14 @@ static struct ras_manager *amdgpu_ras_create_obj(struct amdgpu_device *adev, if (head->block >= AMDGPU_RAS_BLOCK_COUNT) return NULL; - obj = &con->objs[head->block]; + if (head->block == AMDGPU_RAS_BLOCK__MCA) { + if (head->sub_block_index >= AMDGPU_RAS_MCA_BLOCK__LAST) + return NULL; + + obj = &con->objs[AMDGPU_RAS_BLOCK__LAST + head->sub_block_index]; + } else + obj = &con->objs[head->block]; + /* already exist. return obj? */ if (alive_obj(obj)) return NULL; @@ -573,19 +613,21 @@ struct ras_manager *amdgpu_ras_find_obj(struct amdgpu_device *adev, if (head->block >= AMDGPU_RAS_BLOCK_COUNT) return NULL; - obj = &con->objs[head->block]; + if (head->block == AMDGPU_RAS_BLOCK__MCA) { + if (head->sub_block_index >= AMDGPU_RAS_MCA_BLOCK__LAST) + return NULL; - if (alive_obj(obj)) { - WARN_ON(head->block != obj->head.block); + obj = &con->objs[AMDGPU_RAS_BLOCK__LAST + head->sub_block_index]; + } else + obj = &con->objs[head->block]; + + if (alive_obj(obj)) return obj; - } } else { - for (i = 0; i < AMDGPU_RAS_BLOCK_COUNT; i++) { + for (i = 0; i < AMDGPU_RAS_BLOCK_COUNT + AMDGPU_RAS_MCA_BLOCK_COUNT; i++) { obj = &con->objs[i]; - if (alive_obj(obj)) { - WARN_ON(i != obj->head.block); + if (alive_obj(obj)) return obj; - } } } @@ -626,8 +668,6 @@ static int __amdgpu_ras_feature_enable(struct amdgpu_device *adev, */ if (!amdgpu_ras_is_feature_allowed(adev, head)) return 0; - if (!(!!enable ^ !!amdgpu_ras_is_feature_enabled(adev, head))) - return 0; if (enable) { if (!obj) { @@ -678,19 +718,14 @@ int amdgpu_ras_feature_enable(struct amdgpu_device *adev, /* Do not enable if it is not allowed. */ WARN_ON(enable && !amdgpu_ras_is_feature_allowed(adev, head)); - /* Are we alerady in that state we are going to set? */ - if (!(!!enable ^ !!amdgpu_ras_is_feature_enabled(adev, head))) { - ret = 0; - goto out; - } if (!amdgpu_ras_intr_triggered()) { ret = psp_ras_enable_features(&adev->psp, info, enable); if (ret) { - dev_err(adev->dev, "ras %s %s failed %d\n", + dev_err(adev->dev, "ras %s %s failed poison:%d ret:%d\n", enable ? "enable":"disable", - ras_block_str(head->block), - ret); + get_ras_block_str(head), + amdgpu_ras_is_poison_mode_supported(adev), ret); goto out; } } @@ -731,7 +766,7 @@ int amdgpu_ras_feature_enable_on_boot(struct amdgpu_device *adev, if (!ret) dev_info(adev->dev, "RAS INFO: %s setup object\n", - ras_block_str(head->block)); + get_ras_block_str(head)); } } else { /* setup the object then issue a ras TA disable cmd.*/ @@ -781,17 +816,39 @@ static int amdgpu_ras_enable_all_features(struct amdgpu_device *adev, bool bypass) { struct amdgpu_ras *con = amdgpu_ras_get_context(adev); - int ras_block_count = AMDGPU_RAS_BLOCK_COUNT; int i; - const enum amdgpu_ras_error_type default_ras_type = - AMDGPU_RAS_ERROR__NONE; + const enum amdgpu_ras_error_type default_ras_type = AMDGPU_RAS_ERROR__NONE; - for (i = 0; i < ras_block_count; i++) { + for (i = 0; i < AMDGPU_RAS_BLOCK_COUNT; i++) { struct ras_common_if head = { .block = i, .type = default_ras_type, .sub_block_index = 0, }; + + if (i == AMDGPU_RAS_BLOCK__MCA) + continue; + + if (bypass) { + /* + * bypass psp. vbios enable ras for us. + * so just create the obj + */ + if (__amdgpu_ras_feature_enable(adev, &head, 1)) + break; + } else { + if (amdgpu_ras_feature_enable(adev, &head, 1)) + break; + } + } + + for (i = 0; i < AMDGPU_RAS_MCA_BLOCK_COUNT; i++) { + struct ras_common_if head = { + .block = AMDGPU_RAS_BLOCK__MCA, + .type = default_ras_type, + .sub_block_index = i, + }; + if (bypass) { /* * bypass psp. vbios enable ras for us. @@ -809,6 +866,64 @@ static int amdgpu_ras_enable_all_features(struct amdgpu_device *adev, } /* feature ctl end */ + +static void amdgpu_ras_mca_query_error_status(struct amdgpu_device *adev, + struct ras_common_if *ras_block, + struct ras_err_data *err_data) +{ + switch (ras_block->sub_block_index) { + case AMDGPU_RAS_MCA_BLOCK__MP0: + if (adev->mca.mp0.ras_funcs && + adev->mca.mp0.ras_funcs->query_ras_error_count) + adev->mca.mp0.ras_funcs->query_ras_error_count(adev, &err_data); + break; + case AMDGPU_RAS_MCA_BLOCK__MP1: + if (adev->mca.mp1.ras_funcs && + adev->mca.mp1.ras_funcs->query_ras_error_count) + adev->mca.mp1.ras_funcs->query_ras_error_count(adev, &err_data); + break; + case AMDGPU_RAS_MCA_BLOCK__MPIO: + if (adev->mca.mpio.ras_funcs && + adev->mca.mpio.ras_funcs->query_ras_error_count) + adev->mca.mpio.ras_funcs->query_ras_error_count(adev, &err_data); + break; + default: + break; + } +} + +static void amdgpu_ras_get_ecc_info(struct amdgpu_device *adev, struct ras_err_data *err_data) +{ + struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); + int ret = 0; + + /* + * choosing right query method according to + * whether smu support query error information + */ + ret = smu_get_ecc_info(&adev->smu, (void *)&(ras->umc_ecc)); + if (ret == -EOPNOTSUPP) { + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->query_ras_error_count) + adev->umc.ras_funcs->query_ras_error_count(adev, err_data); + + /* umc query_ras_error_address is also responsible for clearing + * error status + */ + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->query_ras_error_address) + adev->umc.ras_funcs->query_ras_error_address(adev, err_data); + } else if (!ret) { + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->ecc_info_query_ras_error_count) + adev->umc.ras_funcs->ecc_info_query_ras_error_count(adev, err_data); + + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->ecc_info_query_ras_error_address) + adev->umc.ras_funcs->ecc_info_query_ras_error_address(adev, err_data); + } +} + /* query/inject/cure begin */ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, struct ras_query_if *info) @@ -822,15 +937,7 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, switch (info->head.block) { case AMDGPU_RAS_BLOCK__UMC: - if (adev->umc.ras_funcs && - adev->umc.ras_funcs->query_ras_error_count) - adev->umc.ras_funcs->query_ras_error_count(adev, &err_data); - /* umc query_ras_error_address is also responsible for clearing - * error status - */ - if (adev->umc.ras_funcs && - adev->umc.ras_funcs->query_ras_error_address) - adev->umc.ras_funcs->query_ras_error_address(adev, &err_data); + amdgpu_ras_get_ecc_info(adev, &err_data); break; case AMDGPU_RAS_BLOCK__SDMA: if (adev->sdma.funcs->query_ras_error_count) { @@ -872,6 +979,9 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, adev->hdp.ras_funcs->query_ras_error_count) adev->hdp.ras_funcs->query_ras_error_count(adev, &err_data); break; + case AMDGPU_RAS_BLOCK__MCA: + amdgpu_ras_mca_query_error_status(adev, &info->head, &err_data); + break; default: break; } @@ -893,13 +1003,13 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, adev->smuio.funcs->get_socket_id(adev), adev->smuio.funcs->get_die_id(adev), obj->err_data.ce_count, - ras_block_str(info->head.block)); + get_ras_block_str(&info->head)); } else { dev_info(adev->dev, "%ld correctable hardware errors " "detected in %s block, no user " "action is needed.\n", obj->err_data.ce_count, - ras_block_str(info->head.block)); + get_ras_block_str(&info->head)); } } if (err_data.ue_count) { @@ -912,15 +1022,18 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, adev->smuio.funcs->get_socket_id(adev), adev->smuio.funcs->get_die_id(adev), obj->err_data.ue_count, - ras_block_str(info->head.block)); + get_ras_block_str(&info->head)); } else { dev_info(adev->dev, "%ld uncorrectable hardware errors " "detected in %s block\n", obj->err_data.ue_count, - ras_block_str(info->head.block)); + get_ras_block_str(&info->head)); } } + if (!amdgpu_persistent_edc_harvesting_supported(adev)) + amdgpu_ras_reset_error_status(adev, info->head.block); + return 0; } @@ -1027,6 +1140,7 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev, case AMDGPU_RAS_BLOCK__SDMA: case AMDGPU_RAS_BLOCK__MMHUB: case AMDGPU_RAS_BLOCK__PCIE_BIF: + case AMDGPU_RAS_BLOCK__MCA: ret = psp_ras_trigger_error(&adev->psp, &block_info); break; case AMDGPU_RAS_BLOCK__XGMI_WAFL: @@ -1034,22 +1148,22 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev, break; default: dev_info(adev->dev, "%s error injection is not supported yet\n", - ras_block_str(info->head.block)); + get_ras_block_str(&info->head)); ret = -EINVAL; } if (ret) dev_err(adev->dev, "ras inject %s failed %d\n", - ras_block_str(info->head.block), ret); + get_ras_block_str(&info->head), ret); return ret; } /** * amdgpu_ras_query_error_count -- Get error counts of all IPs - * adev: pointer to AMD GPU device - * ce_count: pointer to an integer to be set to the count of correctible errors. - * ue_count: pointer to an integer to be set to the count of uncorrectible + * @adev: pointer to AMD GPU device + * @ce_count: pointer to an integer to be set to the count of correctible errors. + * @ue_count: pointer to an integer to be set to the count of uncorrectible * errors. * * If set, @ce_count or @ue_count, count and return the corresponding @@ -1383,7 +1497,7 @@ void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev) if (amdgpu_ras_is_supported(adev, obj->head.block) && (obj->attr_inuse == 1)) { sprintf(fs_info.debugfs_name, "%s_err_inject", - ras_block_str(obj->head.block)); + get_ras_block_str(&obj->head)); fs_info.head = obj->head; amdgpu_ras_debugfs_create(adev, &fs_info, dir); } @@ -1469,22 +1583,28 @@ static void amdgpu_ras_interrupt_handler(struct ras_manager *obj) data->rptr = (data->aligned_element_size + data->rptr) % data->ring_size; - /* Let IP handle its data, maybe we need get the output - * from the callback to udpate the error type/count, etc - */ if (data->cb) { - ret = data->cb(obj->adev, &err_data, &entry); - /* ue will trigger an interrupt, and in that case - * we need do a reset to recovery the whole system. - * But leave IP do that recovery, here we just dispatch - * the error. - */ - if (ret == AMDGPU_RAS_SUCCESS) { - /* these counts could be left as 0 if - * some blocks do not count error number + if (amdgpu_ras_is_poison_mode_supported(obj->adev) && + obj->head.block == AMDGPU_RAS_BLOCK__UMC) + dev_info(obj->adev->dev, + "Poison is created, no user action is needed.\n"); + else { + /* Let IP handle its data, maybe we need get the output + * from the callback to udpate the error type/count, etc + */ + ret = data->cb(obj->adev, &err_data, &entry); + /* ue will trigger an interrupt, and in that case + * we need do a reset to recovery the whole system. + * But leave IP do that recovery, here we just dispatch + * the error. */ - obj->err_data.ue_count += err_data.ue_count; - obj->err_data.ce_count += err_data.ce_count; + if (ret == AMDGPU_RAS_SUCCESS) { + /* these counts could be left as 0 if + * some blocks do not count error number + */ + obj->err_data.ue_count += err_data.ue_count; + obj->err_data.ce_count += err_data.ce_count; + } } } } @@ -1627,6 +1747,16 @@ static void amdgpu_ras_log_on_err_counter(struct amdgpu_device *adev) if (info.head.block == AMDGPU_RAS_BLOCK__PCIE_BIF) continue; + /* + * this is a workaround for aldebaran, skip send msg to + * smu to get ecc_info table due to smu handle get ecc + * info table failed temporarily. + * should be removed until smu fix handle ecc_info table. + */ + if ((info.head.block == AMDGPU_RAS_BLOCK__UMC) && + (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(13, 0, 2))) + continue; + amdgpu_ras_query_error_status(adev, &info); } } @@ -1839,9 +1969,11 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev) if (!con || !con->eh_data) return 0; + mutex_lock(&con->recovery_lock); control = &con->eeprom_control; data = con->eh_data; save_count = data->count - control->ras_num_recs; + mutex_unlock(&con->recovery_lock); /* only new entries are saved */ if (save_count > 0) { if (amdgpu_ras_eeprom_append(control, @@ -2014,6 +2146,11 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev) adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.ras_num_recs); } +#ifdef CONFIG_X86_MCE_AMD + if ((adev->asic_type == CHIP_ALDEBARAN) && + (adev->gmc.xgmi.connected_to_cpu)) + amdgpu_register_bad_pages_mca_notifier(adev); +#endif return 0; free: @@ -2056,19 +2193,6 @@ static int amdgpu_ras_recovery_fini(struct amdgpu_device *adev) } /* recovery end */ -/* return 0 if ras will reset gpu and repost.*/ -int amdgpu_ras_request_reset_on_boot(struct amdgpu_device *adev, - unsigned int block) -{ - struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); - - if (!ras) - return -EINVAL; - - ras->flags |= AMDGPU_RAS_FLAG_INIT_NEED_RESET; - return 0; -} - static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev) { return adev->asic_type == CHIP_VEGA10 || @@ -2176,12 +2300,14 @@ int amdgpu_ras_init(struct amdgpu_device *adev) { struct amdgpu_ras *con = amdgpu_ras_get_context(adev); int r; + bool df_poison, umc_poison; if (con) return 0; con = kmalloc(sizeof(struct amdgpu_ras) + - sizeof(struct ras_manager) * AMDGPU_RAS_BLOCK_COUNT, + sizeof(struct ras_manager) * AMDGPU_RAS_BLOCK_COUNT + + sizeof(struct ras_manager) * AMDGPU_RAS_MCA_BLOCK_COUNT, GFP_KERNEL|__GFP_ZERO); if (!con) return -ENOMEM; @@ -2245,6 +2371,27 @@ int amdgpu_ras_init(struct amdgpu_device *adev) goto release_con; } + /* Init poison supported flag, the default value is false */ + if (adev->gmc.xgmi.connected_to_cpu) { + /* enabled by default when GPU is connected to CPU */ + con->poison_supported = true; + } + else if (adev->df.funcs && + adev->df.funcs->query_ras_poison_mode && + adev->umc.ras_funcs && + adev->umc.ras_funcs->query_ras_poison_mode) { + df_poison = + adev->df.funcs->query_ras_poison_mode(adev); + umc_poison = + adev->umc.ras_funcs->query_ras_poison_mode(adev); + /* Only poison is set in both DF and UMC, we can support it */ + if (df_poison && umc_poison) + con->poison_supported = true; + else if (df_poison != umc_poison) + dev_warn(adev->dev, "Poison setting is inconsistent in DF/UMC(%d:%d)!\n", + df_poison, umc_poison); + } + if (amdgpu_ras_fs_init(adev)) { r = -EINVAL; goto release_con; @@ -2288,6 +2435,16 @@ static int amdgpu_persistent_edc_harvesting(struct amdgpu_device *adev, return 0; } +bool amdgpu_ras_is_poison_mode_supported(struct amdgpu_device *adev) +{ + struct amdgpu_ras *con = amdgpu_ras_get_context(adev); + + if (!con) + return false; + + return con->poison_supported; +} + /* helper function to handle common stuff in ip late init phase */ int amdgpu_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block, @@ -2306,12 +2463,7 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev, r = amdgpu_ras_feature_enable_on_boot(adev, ras_block, 1); if (r) { - if (r == -EAGAIN) { - /* request gpu reset. will run again */ - amdgpu_ras_request_reset_on_boot(adev, - ras_block->block); - return 0; - } else if (adev->in_suspend || amdgpu_in_reset(adev)) { + if (adev->in_suspend || amdgpu_in_reset(adev)) { /* in resume phase, if fail to enable ras, * clean up all ras fs nodes, and disable ras */ goto cleanup; @@ -2365,7 +2517,6 @@ void amdgpu_ras_late_fini(struct amdgpu_device *adev, amdgpu_ras_sysfs_remove(adev, ras_block); if (ih_info->cb) amdgpu_ras_interrupt_remove_handler(adev, ih_info); - amdgpu_ras_feature_enable(adev, ras_block, 0); } /* do some init work after IP late init as dependence. @@ -2403,19 +2554,6 @@ void amdgpu_ras_resume(struct amdgpu_device *adev) } } } - - if (con->flags & AMDGPU_RAS_FLAG_INIT_NEED_RESET) { - con->flags &= ~AMDGPU_RAS_FLAG_INIT_NEED_RESET; - /* setup ras obj state as disabled. - * for init_by_vbios case. - * if we want to enable ras, just enable it in a normal way. - * If we want do disable it, need setup ras obj as enabled, - * then issue another TA disable cmd. - * See feature_enable_on_boot - */ - amdgpu_ras_disable_all_features(adev, 1); - amdgpu_ras_reset_gpu(adev); - } } void amdgpu_ras_suspend(struct amdgpu_device *adev) @@ -2507,3 +2645,136 @@ void amdgpu_release_ras_context(struct amdgpu_device *adev) kfree(con); } } + +#ifdef CONFIG_X86_MCE_AMD +static struct amdgpu_device *find_adev(uint32_t node_id) +{ + int i; + struct amdgpu_device *adev = NULL; + + for (i = 0; i < mce_adev_list.num_gpu; i++) { + adev = mce_adev_list.devs[i]; + + if (adev && adev->gmc.xgmi.connected_to_cpu && + adev->gmc.xgmi.physical_node_id == node_id) + break; + adev = NULL; + } + + return adev; +} + +#define GET_MCA_IPID_GPUID(m) (((m) >> 44) & 0xF) +#define GET_UMC_INST(m) (((m) >> 21) & 0x7) +#define GET_CHAN_INDEX(m) ((((m) >> 12) & 0x3) | (((m) >> 18) & 0x4)) +#define GPU_ID_OFFSET 8 + +static int amdgpu_bad_page_notifier(struct notifier_block *nb, + unsigned long val, void *data) +{ + struct mce *m = (struct mce *)data; + struct amdgpu_device *adev = NULL; + uint32_t gpu_id = 0; + uint32_t umc_inst = 0; + uint32_t ch_inst, channel_index = 0; + struct ras_err_data err_data = {0, 0, 0, NULL}; + struct eeprom_table_record err_rec; + uint64_t retired_page; + + /* + * If the error was generated in UMC_V2, which belongs to GPU UMCs, + * and error occurred in DramECC (Extended error code = 0) then only + * process the error, else bail out. + */ + if (!m || !((smca_get_bank_type(m->extcpu, m->bank) == SMCA_UMC_V2) && + (XEC(m->status, 0x3f) == 0x0))) + return NOTIFY_DONE; + + /* + * If it is correctable error, return. + */ + if (mce_is_correctable(m)) + return NOTIFY_OK; + + /* + * GPU Id is offset by GPU_ID_OFFSET in MCA_IPID_UMC register. + */ + gpu_id = GET_MCA_IPID_GPUID(m->ipid) - GPU_ID_OFFSET; + + adev = find_adev(gpu_id); + if (!adev) { + DRM_WARN("%s: Unable to find adev for gpu_id: %d\n", __func__, + gpu_id); + return NOTIFY_DONE; + } + + /* + * If it is uncorrectable error, then find out UMC instance and + * channel index. + */ + umc_inst = GET_UMC_INST(m->ipid); + ch_inst = GET_CHAN_INDEX(m->ipid); + + dev_info(adev->dev, "Uncorrectable error detected in UMC inst: %d, chan_idx: %d", + umc_inst, ch_inst); + + memset(&err_rec, 0x0, sizeof(struct eeprom_table_record)); + + /* + * Translate UMC channel address to Physical address + */ + channel_index = + adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + + ch_inst]; + + retired_page = ADDR_OF_8KB_BLOCK(m->addr) | + ADDR_OF_256B_BLOCK(channel_index) | + OFFSET_IN_256B_BLOCK(m->addr); + + err_rec.address = m->addr; + err_rec.retired_page = retired_page >> AMDGPU_GPU_PAGE_SHIFT; + err_rec.ts = (uint64_t)ktime_get_real_seconds(); + err_rec.err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE; + err_rec.cu = 0; + err_rec.mem_channel = channel_index; + err_rec.mcumc_id = umc_inst; + + err_data.err_addr = &err_rec; + err_data.err_addr_cnt = 1; + + if (amdgpu_bad_page_threshold != 0) { + amdgpu_ras_add_bad_pages(adev, err_data.err_addr, + err_data.err_addr_cnt); + amdgpu_ras_save_bad_pages(adev); + } + + return NOTIFY_OK; +} + +static struct notifier_block amdgpu_bad_page_nb = { + .notifier_call = amdgpu_bad_page_notifier, + .priority = MCE_PRIO_UC, +}; + +static void amdgpu_register_bad_pages_mca_notifier(struct amdgpu_device *adev) +{ + /* + * Add the adev to the mce_adev_list. + * During mode2 reset, amdgpu device is temporarily + * removed from the mgpu_info list which can cause + * page retirement to fail. + * Use this list instead of mgpu_info to find the amdgpu + * device on which the UMC error was reported. + */ + mce_adev_list.devs[mce_adev_list.num_gpu++] = adev; + + /* + * Register the x86 notifier only once + * with MCE subsystem. + */ + if (notifier_registered == false) { + mce_register_decode_chain(&amdgpu_bad_page_nb); + notifier_registered = true; + } +} +#endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h index eae604fd90b8..1c708122d492 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h @@ -32,7 +32,6 @@ #include "amdgpu_ras_eeprom.h" #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS (0x1 << 0) -#define AMDGPU_RAS_FLAG_INIT_NEED_RESET (0x1 << 1) enum amdgpu_ras_block { AMDGPU_RAS_BLOCK__UMC = 0, @@ -49,15 +48,22 @@ enum amdgpu_ras_block { AMDGPU_RAS_BLOCK__MP0, AMDGPU_RAS_BLOCK__MP1, AMDGPU_RAS_BLOCK__FUSE, - AMDGPU_RAS_BLOCK__MPIO, + AMDGPU_RAS_BLOCK__MCA, AMDGPU_RAS_BLOCK__LAST }; -extern const char *ras_block_string[]; +enum amdgpu_ras_mca_block { + AMDGPU_RAS_MCA_BLOCK__MP0 = 0, + AMDGPU_RAS_MCA_BLOCK__MP1, + AMDGPU_RAS_MCA_BLOCK__MPIO, + AMDGPU_RAS_MCA_BLOCK__IOHC, + + AMDGPU_RAS_MCA_BLOCK__LAST +}; -#define ras_block_str(i) (ras_block_string[i]) #define AMDGPU_RAS_BLOCK_COUNT AMDGPU_RAS_BLOCK__LAST +#define AMDGPU_RAS_MCA_BLOCK_COUNT AMDGPU_RAS_MCA_BLOCK__LAST #define AMDGPU_RAS_BLOCK_MASK ((1ULL << AMDGPU_RAS_BLOCK_COUNT) - 1) enum amdgpu_ras_gfx_subblock { @@ -313,6 +319,19 @@ struct ras_common_if { char name[32]; }; +#define MAX_UMC_CHANNEL_NUM 32 + +struct ecc_info_per_ch { + uint16_t ce_count_lo_chip; + uint16_t ce_count_hi_chip; + uint64_t mca_umc_status; + uint64_t mca_umc_addr; +}; + +struct umc_ecc_info { + struct ecc_info_per_ch ecc[MAX_UMC_CHANNEL_NUM]; +}; + struct amdgpu_ras { /* ras infrastructure */ /* for ras itself. */ @@ -345,10 +364,16 @@ struct amdgpu_ras { /* disable ras error count harvest in recovery */ bool disable_ras_err_cnt_harvest; + /* is poison mode supported */ + bool poison_supported; + /* RAS count errors delayed work */ struct delayed_work ras_counte_delay_work; atomic_t ras_ue_count; atomic_t ras_ce_count; + + /* record umc error info queried from smu */ + struct umc_ecc_info umc_ecc; }; struct ras_fs_data { @@ -488,8 +513,6 @@ static inline int amdgpu_ras_is_supported(struct amdgpu_device *adev, } int amdgpu_ras_recovery_init(struct amdgpu_device *adev); -int amdgpu_ras_request_reset_on_boot(struct amdgpu_device *adev, - unsigned int block); void amdgpu_ras_resume(struct amdgpu_device *adev); void amdgpu_ras_suspend(struct amdgpu_device *adev); @@ -544,6 +567,8 @@ amdgpu_ras_block_to_ta(enum amdgpu_ras_block block) { return TA_RAS_BLOCK__MP1; case AMDGPU_RAS_BLOCK__FUSE: return TA_RAS_BLOCK__FUSE; + case AMDGPU_RAS_BLOCK__MCA: + return TA_RAS_BLOCK__MCA; default: WARN_ONCE(1, "RAS ERROR: unexpected block id %d\n", block); return TA_RAS_BLOCK__UMC; @@ -638,4 +663,8 @@ void amdgpu_release_ras_context(struct amdgpu_device *adev); int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev); +const char *get_ras_block_str(struct ras_common_if *ras_block); + +bool amdgpu_ras_is_poison_mode_supported(struct amdgpu_device *adev); + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c index 98732518543e..05117eda105b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c @@ -1077,6 +1077,13 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control, if (res) DRM_ERROR("RAS table incorrect checksum or error:%d\n", res); + + /* Warn if we are at 90% of the threshold or above + */ + if (10 * control->ras_num_recs >= 9 * ras->bad_page_cnt_threshold) + dev_warn(adev->dev, "RAS records:%u exceeds 90%% of threshold:%d", + control->ras_num_recs, + ras->bad_page_cnt_threshold); } else if (hdr->header == RAS_TABLE_HDR_BAD && amdgpu_bad_page_threshold != 0) { res = __verify_ras_table_checksum(control); @@ -1098,11 +1105,18 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control, res = amdgpu_ras_eeprom_correct_header_tag(control, RAS_TABLE_HDR_VAL); } else { - *exceed_err_limit = true; - dev_err(adev->dev, - "RAS records:%d exceed threshold:%d, " - "maybe retire this GPU?", + dev_err(adev->dev, "RAS records:%d exceed threshold:%d", control->ras_num_recs, ras->bad_page_cnt_threshold); + if (amdgpu_bad_page_threshold == -2) { + dev_warn(adev->dev, "GPU will be initialized due to bad_page_threshold = -2."); + res = 0; + } else { + *exceed_err_limit = 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_recs, ras->bad_page_cnt_threshold); + } } } else { DRM_INFO("Creating a new EEPROM table"); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index 0554576d3695..ab2351ba9574 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -415,26 +415,20 @@ static const struct file_operations amdgpu_debugfs_ring_fops = { #endif -int amdgpu_debugfs_ring_init(struct amdgpu_device *adev, - struct amdgpu_ring *ring) +void amdgpu_debugfs_ring_init(struct amdgpu_device *adev, + struct amdgpu_ring *ring) { #if defined(CONFIG_DEBUG_FS) struct drm_minor *minor = adev_to_drm(adev)->primary; - struct dentry *ent, *root = minor->debugfs_root; + struct dentry *root = minor->debugfs_root; char name[32]; sprintf(name, "amdgpu_ring_%s", ring->name); + debugfs_create_file_size(name, S_IFREG | S_IRUGO, root, ring, + &amdgpu_debugfs_ring_fops, + ring->ring_size + 12); - ent = debugfs_create_file(name, - S_IFREG | S_IRUGO, root, - ring, &amdgpu_debugfs_ring_fops); - if (IS_ERR(ent)) - return PTR_ERR(ent); - - i_size_write(ent->d_inode, ring->ring_size + 12); - ring->ent = ent; #endif - return 0; } /** diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index e713d31619fe..fae7d185ad0d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -36,8 +36,13 @@ #define AMDGPU_MAX_VCE_RINGS 3 #define AMDGPU_MAX_UVD_ENC_RINGS 2 -#define AMDGPU_RING_PRIO_DEFAULT 1 -#define AMDGPU_RING_PRIO_MAX AMDGPU_GFX_PIPE_PRIO_MAX +enum amdgpu_ring_priority_level { + AMDGPU_RING_PRIO_0, + AMDGPU_RING_PRIO_1, + AMDGPU_RING_PRIO_DEFAULT = 1, + AMDGPU_RING_PRIO_2, + AMDGPU_RING_PRIO_MAX +}; /* some special values for the owner field */ #define AMDGPU_FENCE_OWNER_UNDEFINED ((void *)0ul) @@ -48,9 +53,6 @@ #define AMDGPU_FENCE_FLAG_INT (1 << 1) #define AMDGPU_FENCE_FLAG_TC_WB_ONLY (1 << 2) -/* fence flag bit to indicate the face is embedded in job*/ -#define AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT (DMA_FENCE_FLAG_USER_BITS + 1) - #define to_amdgpu_ring(s) container_of((s), struct amdgpu_ring, sched) #define AMDGPU_IB_POOL_SIZE (1024 * 1024) @@ -109,6 +111,7 @@ struct amdgpu_fence_driver { struct dma_fence **fences; }; +void amdgpu_fence_driver_clear_job_fences(struct amdgpu_ring *ring); void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring); int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring, @@ -248,10 +251,6 @@ struct amdgpu_ring { bool has_compute_vm_bug; bool no_scheduler; int hw_prio; - -#if defined(CONFIG_DEBUG_FS) - struct dentry *ent; -#endif }; #define amdgpu_ring_parse_cs(r, p, ib) ((r)->funcs->parse_cs((p), (ib))) @@ -351,8 +350,6 @@ static inline void amdgpu_ring_write_multiple(struct amdgpu_ring *ring, int amdgpu_ring_test_helper(struct amdgpu_ring *ring); -int amdgpu_debugfs_ring_init(struct amdgpu_device *adev, - struct amdgpu_ring *ring); -void amdgpu_debugfs_ring_fini(struct amdgpu_ring *ring); - +void amdgpu_debugfs_ring_init(struct amdgpu_device *adev, + struct amdgpu_ring *ring); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c index b7d861ed5284..e9b45089a28a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c @@ -32,37 +32,9 @@ #include "amdgpu_sched.h" #include "amdgpu_vm.h" -int amdgpu_to_sched_priority(int amdgpu_priority, - enum drm_sched_priority *prio) -{ - switch (amdgpu_priority) { - case AMDGPU_CTX_PRIORITY_VERY_HIGH: - *prio = DRM_SCHED_PRIORITY_HIGH; - break; - case AMDGPU_CTX_PRIORITY_HIGH: - *prio = DRM_SCHED_PRIORITY_HIGH; - break; - case AMDGPU_CTX_PRIORITY_NORMAL: - *prio = DRM_SCHED_PRIORITY_NORMAL; - break; - case AMDGPU_CTX_PRIORITY_LOW: - case AMDGPU_CTX_PRIORITY_VERY_LOW: - *prio = DRM_SCHED_PRIORITY_MIN; - break; - case AMDGPU_CTX_PRIORITY_UNSET: - *prio = DRM_SCHED_PRIORITY_UNSET; - break; - default: - WARN(1, "Invalid context priority %d\n", amdgpu_priority); - return -EINVAL; - } - - return 0; -} - static int amdgpu_sched_process_priority_override(struct amdgpu_device *adev, int fd, - enum drm_sched_priority priority) + int32_t priority) { struct fd f = fdget(fd); struct amdgpu_fpriv *fpriv; @@ -89,7 +61,7 @@ static int amdgpu_sched_process_priority_override(struct amdgpu_device *adev, static int amdgpu_sched_context_priority_override(struct amdgpu_device *adev, int fd, unsigned ctx_id, - enum drm_sched_priority priority) + int32_t priority) { struct fd f = fdget(fd); struct amdgpu_fpriv *fpriv; @@ -124,7 +96,6 @@ int amdgpu_sched_ioctl(struct drm_device *dev, void *data, { union drm_amdgpu_sched *args = data; struct amdgpu_device *adev = drm_to_adev(dev); - enum drm_sched_priority priority; int r; /* First check the op, then the op's argument. @@ -138,21 +109,22 @@ int amdgpu_sched_ioctl(struct drm_device *dev, void *data, return -EINVAL; } - r = amdgpu_to_sched_priority(args->in.priority, &priority); - if (r) - return r; + if (!amdgpu_ctx_priority_is_valid(args->in.priority)) { + WARN(1, "Invalid context priority %d\n", args->in.priority); + return -EINVAL; + } switch (args->in.op) { case AMDGPU_SCHED_OP_PROCESS_PRIORITY_OVERRIDE: r = amdgpu_sched_process_priority_override(adev, args->in.fd, - priority); + args->in.priority); break; case AMDGPU_SCHED_OP_CONTEXT_PRIORITY_OVERRIDE: r = amdgpu_sched_context_priority_override(adev, args->in.fd, args->in.ctx_id, - priority); + args->in.priority); break; default: /* Impossible. diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c index 862eb3c1c4c5..f7d8487799b2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c @@ -252,41 +252,25 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync, struct dma_resv *resv, enum amdgpu_sync_mode mode, void *owner) { - struct dma_resv_list *flist; + struct dma_resv_iter cursor; struct dma_fence *f; - unsigned i; - int r = 0; + int r; if (resv == NULL) return -EINVAL; - /* always sync to the exclusive fence */ - f = dma_resv_excl_fence(resv); - dma_fence_chain_for_each(f, f) { - struct dma_fence_chain *chain = to_dma_fence_chain(f); - - if (amdgpu_sync_test_fence(adev, mode, owner, chain ? - chain->fence : f)) { - r = amdgpu_sync_fence(sync, f); - dma_fence_put(f); - if (r) - return r; - break; - } - } - - flist = dma_resv_shared_list(resv); - if (!flist) - return 0; - - for (i = 0; i < flist->shared_count; ++i) { - f = rcu_dereference_protected(flist->shared[i], - dma_resv_held(resv)); - - if (amdgpu_sync_test_fence(adev, mode, owner, f)) { - r = amdgpu_sync_fence(sync, f); - if (r) - return r; + dma_resv_for_each_fence(&cursor, resv, true, f) { + dma_fence_chain_for_each(f, f) { + struct dma_fence_chain *chain = to_dma_fence_chain(f); + + if (amdgpu_sync_test_fence(adev, mode, owner, chain ? + chain->fence : f)) { + r = amdgpu_sync_fence(sync, f); + dma_fence_put(f); + if (r) + return r; + break; + } } } return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 94126dc39688..fb0d8bffdce2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -41,6 +41,7 @@ #include <linux/swiotlb.h> #include <linux/dma-buf.h> #include <linux/sizes.h> +#include <linux/module.h> #include <drm/ttm/ttm_bo_api.h> #include <drm/ttm/ttm_bo_driver.h> @@ -59,6 +60,8 @@ #include "amdgpu_res_cursor.h" #include "bif/bif_4_1_d.h" +MODULE_IMPORT_NS(DMA_BUF); + #define AMDGPU_TTM_VRAM_MAX_DW_READ (size_t)128 static int amdgpu_ttm_backend_bind(struct ttm_device *bdev, @@ -113,17 +116,8 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo, abo = ttm_to_amdgpu_bo(bo); if (abo->flags & AMDGPU_AMDKFD_CREATE_SVM_BO) { - struct dma_fence *fence; - struct dma_resv *resv = &bo->base._resv; - - rcu_read_lock(); - fence = rcu_dereference(resv->fence_excl); - if (fence && !fence->ops->signaled) - dma_fence_enable_sw_signaling(fence); - placement->num_placement = 0; placement->num_busy_placement = 0; - rcu_read_unlock(); return; } @@ -696,6 +690,9 @@ int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages) true, NULL); out_unlock: mmap_read_unlock(mm); + if (r) + pr_debug("failed %d to get user pages 0x%lx\n", r, start); + mmput(mm); return r; @@ -894,7 +891,7 @@ static int amdgpu_ttm_backend_bind(struct ttm_device *bdev, DRM_ERROR("failed to pin userptr\n"); return r; } - } else if (ttm->page_flags & TTM_PAGE_FLAG_SG) { + } else if (ttm->page_flags & TTM_TT_FLAG_EXTERNAL) { if (!ttm->sg) { struct dma_buf_attachment *attach; struct sg_table *sgt; @@ -916,11 +913,6 @@ static int amdgpu_ttm_backend_bind(struct ttm_device *bdev, ttm->num_pages, bo_mem, ttm); } - if (bo_mem->mem_type == AMDGPU_PL_GDS || - bo_mem->mem_type == AMDGPU_PL_GWS || - bo_mem->mem_type == AMDGPU_PL_OA) - return -EINVAL; - if (bo_mem->mem_type != TTM_PL_TT || !amdgpu_gtt_mgr_has_gart_addr(bo_mem)) { gtt->offset = AMDGPU_BO_INVALID_OFFSET; @@ -1066,8 +1058,6 @@ static void amdgpu_ttm_backend_destroy(struct ttm_device *bdev, { struct amdgpu_ttm_tt *gtt = (void *)ttm; - amdgpu_ttm_backend_unbind(bdev, ttm); - ttm_tt_destroy_common(bdev, ttm); if (gtt->usertask) put_task_struct(gtt->usertask); @@ -1121,6 +1111,8 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev, { struct amdgpu_device *adev = amdgpu_ttm_adev(bdev); struct amdgpu_ttm_tt *gtt = (void *)ttm; + pgoff_t i; + int ret; /* user pages are bound by amdgpu_ttm_tt_pin_userptr() */ if (gtt->userptr) { @@ -1130,10 +1122,17 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev, return 0; } - if (ttm->page_flags & TTM_PAGE_FLAG_SG) + if (ttm->page_flags & TTM_TT_FLAG_EXTERNAL) return 0; - return ttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx); + ret = ttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx); + if (ret) + return ret; + + for (i = 0; i < ttm->num_pages; ++i) + ttm->pages[i]->mapping = bdev->dev_mapping; + + return 0; } /* @@ -1147,6 +1146,9 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev, { struct amdgpu_ttm_tt *gtt = (void *)ttm; struct amdgpu_device *adev; + pgoff_t i; + + amdgpu_ttm_backend_unbind(bdev, ttm); if (gtt->userptr) { amdgpu_ttm_tt_set_user_pages(ttm, NULL); @@ -1155,9 +1157,12 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev, return; } - if (ttm->page_flags & TTM_PAGE_FLAG_SG) + if (ttm->page_flags & TTM_TT_FLAG_EXTERNAL) return; + for (i = 0; i < ttm->num_pages; ++i) + ttm->pages[i]->mapping = NULL; + adev = amdgpu_ttm_adev(bdev); return ttm_pool_free(&adev->mman.bdev.pool, ttm); } @@ -1185,8 +1190,8 @@ int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo, return -ENOMEM; } - /* Set TTM_PAGE_FLAG_SG before populate but after create. */ - bo->ttm->page_flags |= TTM_PAGE_FLAG_SG; + /* Set TTM_TT_FLAG_EXTERNAL before populate but after create. */ + bo->ttm->page_flags |= TTM_TT_FLAG_EXTERNAL; gtt = (void *)bo->ttm; gtt->userptr = addr; @@ -1222,7 +1227,7 @@ struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm) * */ bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start, - unsigned long end) + unsigned long end, unsigned long *userptr) { struct amdgpu_ttm_tt *gtt = (void *)ttm; unsigned long size; @@ -1237,6 +1242,8 @@ bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start, if (gtt->userptr > end || gtt->userptr + size <= start) return false; + if (userptr) + *userptr = gtt->userptr; return true; } @@ -1332,10 +1339,9 @@ static bool amdgpu_ttm_bo_eviction_valuable(struct ttm_buffer_object *bo, const struct ttm_place *place) { unsigned long num_pages = bo->resource->num_pages; + struct dma_resv_iter resv_cursor; struct amdgpu_res_cursor cursor; - struct dma_resv_list *flist; struct dma_fence *f; - int i; /* Swapout? */ if (bo->resource->mem_type == TTM_PL_SYSTEM) @@ -1349,14 +1355,9 @@ static bool amdgpu_ttm_bo_eviction_valuable(struct ttm_buffer_object *bo, * If true, then return false as any KFD process needs all its BOs to * be resident to run successfully */ - flist = dma_resv_shared_list(bo->base.resv); - if (flist) { - for (i = 0; i < flist->shared_count; ++i) { - f = rcu_dereference_protected(flist->shared[i], - dma_resv_held(bo->base.resv)); - if (amdkfd_fence_check_mm(f, current->mm)) - return false; - } + dma_resv_for_each_fence(&resv_cursor, bo->base.resv, true, f) { + if (amdkfd_fence_check_mm(f, current->mm)) + return false; } switch (bo->resource->mem_type) { @@ -2036,6 +2037,36 @@ error_free: return r; } +/** + * amdgpu_ttm_evict_resources - evict memory buffers + * @adev: amdgpu device object + * @mem_type: evicted BO's memory type + * + * Evicts all @mem_type buffers on the lru list of the memory type. + * + * Returns: + * 0 for success or a negative error code on failure. + */ +int amdgpu_ttm_evict_resources(struct amdgpu_device *adev, int mem_type) +{ + struct ttm_resource_manager *man; + + switch (mem_type) { + case TTM_PL_VRAM: + case TTM_PL_TT: + case AMDGPU_PL_GWS: + case AMDGPU_PL_GDS: + case AMDGPU_PL_OA: + man = ttm_manager_type(&adev->mman.bdev, mem_type); + break; + default: + DRM_ERROR("Trying to evict invalid memory type\n"); + return -EINVAL; + } + + return ttm_resource_manager_evict_all(&adev->mman.bdev, man); +} + #if defined(CONFIG_DEBUG_FS) static int amdgpu_mm_vram_table_show(struct seq_file *m, void *unused) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h index 3205fd520060..7346ecff4438 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h @@ -182,7 +182,7 @@ int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo, bool amdgpu_ttm_tt_has_userptr(struct ttm_tt *ttm); struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm); bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start, - unsigned long end); + unsigned long end, unsigned long *userptr); bool amdgpu_ttm_tt_userptr_invalidated(struct ttm_tt *ttm, int *last_invalidated); bool amdgpu_ttm_tt_is_userptr(struct ttm_tt *ttm); @@ -190,6 +190,7 @@ bool amdgpu_ttm_tt_is_readonly(struct ttm_tt *ttm); uint64_t amdgpu_ttm_tt_pde_flags(struct ttm_tt *ttm, struct ttm_resource *mem); uint64_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm, struct ttm_resource *mem); +int amdgpu_ttm_evict_resources(struct amdgpu_device *adev, int mem_type); void amdgpu_ttm_debugfs_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c index abd8469380e5..ca3350502618 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c @@ -416,10 +416,11 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type) else return AMDGPU_FW_LOAD_PSP; default: - DRM_ERROR("Unknown firmware load type\n"); + if (!load_type) + return AMDGPU_FW_LOAD_DIRECT; + else + return AMDGPU_FW_LOAD_PSP; } - - return AMDGPU_FW_LOAD_DIRECT; } const char *amdgpu_ucode_name(enum AMDGPU_UCODE_ID ucode_id) @@ -508,7 +509,7 @@ static ssize_t show_##name(struct device *dev, \ struct drm_device *ddev = dev_get_drvdata(dev); \ struct amdgpu_device *adev = drm_to_adev(ddev); \ \ - return snprintf(buf, PAGE_SIZE, "0x%08x\n", adev->field); \ + return sysfs_emit(buf, "0x%08x\n", adev->field); \ } \ static DEVICE_ATTR(name, mode, show_##name, NULL) @@ -525,9 +526,9 @@ FW_VERSION_ATTR(rlc_srls_fw_version, 0444, gfx.rlc_srls_fw_version); FW_VERSION_ATTR(mec_fw_version, 0444, gfx.mec_fw_version); FW_VERSION_ATTR(mec2_fw_version, 0444, gfx.mec2_fw_version); FW_VERSION_ATTR(sos_fw_version, 0444, psp.sos.fw_version); -FW_VERSION_ATTR(asd_fw_version, 0444, psp.asd.fw_version); -FW_VERSION_ATTR(ta_ras_fw_version, 0444, psp.ras.feature_version); -FW_VERSION_ATTR(ta_xgmi_fw_version, 0444, psp.xgmi.feature_version); +FW_VERSION_ATTR(asd_fw_version, 0444, psp.asd_context.bin_desc.fw_version); +FW_VERSION_ATTR(ta_ras_fw_version, 0444, psp.ras_context.context.bin_desc.fw_version); +FW_VERSION_ATTR(ta_xgmi_fw_version, 0444, psp.xgmi_context.context.bin_desc.fw_version); FW_VERSION_ATTR(smc_fw_version, 0444, pm.fw_version); FW_VERSION_ATTR(sdma_fw_version, 0444, sdma.instance[0].fw_version); FW_VERSION_ATTR(sdma2_fw_version, 0444, sdma.instance[1].fw_version); @@ -572,6 +573,7 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev, const struct dmcu_firmware_header_v1_0 *dmcu_hdr = NULL; const struct dmcub_firmware_header_v1_0 *dmcub_hdr = NULL; const struct mes_firmware_header_v1_0 *mes_hdr = NULL; + u8 *ucode_addr; if (NULL == ucode->fw) return 0; @@ -588,94 +590,83 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev, dmcub_hdr = (const struct dmcub_firmware_header_v1_0 *)ucode->fw->data; mes_hdr = (const struct mes_firmware_header_v1_0 *)ucode->fw->data; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP || - (ucode->ucode_id != AMDGPU_UCODE_ID_CP_MEC1 && - ucode->ucode_id != AMDGPU_UCODE_ID_CP_MEC2 && - ucode->ucode_id != AMDGPU_UCODE_ID_CP_MEC1_JT && - ucode->ucode_id != AMDGPU_UCODE_ID_CP_MEC2_JT && - ucode->ucode_id != AMDGPU_UCODE_ID_CP_MES && - ucode->ucode_id != AMDGPU_UCODE_ID_CP_MES_DATA && - ucode->ucode_id != AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL && - ucode->ucode_id != AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM && - ucode->ucode_id != AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM && - ucode->ucode_id != AMDGPU_UCODE_ID_RLC_IRAM && - ucode->ucode_id != AMDGPU_UCODE_ID_RLC_DRAM && - ucode->ucode_id != AMDGPU_UCODE_ID_DMCU_ERAM && - ucode->ucode_id != AMDGPU_UCODE_ID_DMCU_INTV && - ucode->ucode_id != AMDGPU_UCODE_ID_DMCUB)) { - ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes); - - memcpy(ucode->kaddr, (void *)((uint8_t *)ucode->fw->data + - le32_to_cpu(header->ucode_array_offset_bytes)), - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_CP_MEC1 || - ucode->ucode_id == AMDGPU_UCODE_ID_CP_MEC2) { - ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes) - - le32_to_cpu(cp_hdr->jt_size) * 4; - - memcpy(ucode->kaddr, (void *)((uint8_t *)ucode->fw->data + - le32_to_cpu(header->ucode_array_offset_bytes)), - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_CP_MEC1_JT || - ucode->ucode_id == AMDGPU_UCODE_ID_CP_MEC2_JT) { - ucode->ucode_size = le32_to_cpu(cp_hdr->jt_size) * 4; - - memcpy(ucode->kaddr, (void *)((uint8_t *)ucode->fw->data + - le32_to_cpu(header->ucode_array_offset_bytes) + - le32_to_cpu(cp_hdr->jt_offset) * 4), - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_DMCU_ERAM) { - ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes) - + if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { + switch (ucode->ucode_id) { + case AMDGPU_UCODE_ID_CP_MEC1: + case AMDGPU_UCODE_ID_CP_MEC2: + ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes) - + le32_to_cpu(cp_hdr->jt_size) * 4; + ucode_addr = (u8 *)ucode->fw->data + + le32_to_cpu(header->ucode_array_offset_bytes); + break; + case AMDGPU_UCODE_ID_CP_MEC1_JT: + case AMDGPU_UCODE_ID_CP_MEC2_JT: + ucode->ucode_size = le32_to_cpu(cp_hdr->jt_size) * 4; + ucode_addr = (u8 *)ucode->fw->data + + le32_to_cpu(header->ucode_array_offset_bytes) + + le32_to_cpu(cp_hdr->jt_offset) * 4; + break; + case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL: + ucode->ucode_size = adev->gfx.rlc.save_restore_list_cntl_size_bytes; + ucode_addr = adev->gfx.rlc.save_restore_list_cntl; + break; + case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM: + ucode->ucode_size = adev->gfx.rlc.save_restore_list_gpm_size_bytes; + ucode_addr = adev->gfx.rlc.save_restore_list_gpm; + break; + case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM: + ucode->ucode_size = adev->gfx.rlc.save_restore_list_srm_size_bytes; + ucode_addr = adev->gfx.rlc.save_restore_list_srm; + break; + case AMDGPU_UCODE_ID_RLC_IRAM: + ucode->ucode_size = adev->gfx.rlc.rlc_iram_ucode_size_bytes; + ucode_addr = adev->gfx.rlc.rlc_iram_ucode; + break; + case AMDGPU_UCODE_ID_RLC_DRAM: + ucode->ucode_size = adev->gfx.rlc.rlc_dram_ucode_size_bytes; + ucode_addr = adev->gfx.rlc.rlc_dram_ucode; + break; + case AMDGPU_UCODE_ID_CP_MES: + ucode->ucode_size = le32_to_cpu(mes_hdr->mes_ucode_size_bytes); + ucode_addr = (u8 *)ucode->fw->data + + le32_to_cpu(mes_hdr->mes_ucode_offset_bytes); + break; + case AMDGPU_UCODE_ID_CP_MES_DATA: + ucode->ucode_size = le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes); + ucode_addr = (u8 *)ucode->fw->data + + le32_to_cpu(mes_hdr->mes_ucode_data_offset_bytes); + break; + case AMDGPU_UCODE_ID_DMCU_ERAM: + ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes) - le32_to_cpu(dmcu_hdr->intv_size_bytes); - - memcpy(ucode->kaddr, (void *)((uint8_t *)ucode->fw->data + - le32_to_cpu(header->ucode_array_offset_bytes)), - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_DMCU_INTV) { - ucode->ucode_size = le32_to_cpu(dmcu_hdr->intv_size_bytes); - - memcpy(ucode->kaddr, (void *)((uint8_t *)ucode->fw->data + - le32_to_cpu(header->ucode_array_offset_bytes) + - le32_to_cpu(dmcu_hdr->intv_offset_bytes)), - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_DMCUB) { - ucode->ucode_size = le32_to_cpu(dmcub_hdr->inst_const_bytes); - memcpy(ucode->kaddr, - (void *)((uint8_t *)ucode->fw->data + - le32_to_cpu(header->ucode_array_offset_bytes)), - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL) { - ucode->ucode_size = adev->gfx.rlc.save_restore_list_cntl_size_bytes; - memcpy(ucode->kaddr, adev->gfx.rlc.save_restore_list_cntl, - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM) { - ucode->ucode_size = adev->gfx.rlc.save_restore_list_gpm_size_bytes; - memcpy(ucode->kaddr, adev->gfx.rlc.save_restore_list_gpm, - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM) { - ucode->ucode_size = adev->gfx.rlc.save_restore_list_srm_size_bytes; - memcpy(ucode->kaddr, adev->gfx.rlc.save_restore_list_srm, - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_RLC_IRAM) { - ucode->ucode_size = adev->gfx.rlc.rlc_iram_ucode_size_bytes; - memcpy(ucode->kaddr, adev->gfx.rlc.rlc_iram_ucode, - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_RLC_DRAM) { - ucode->ucode_size = adev->gfx.rlc.rlc_dram_ucode_size_bytes; - memcpy(ucode->kaddr, adev->gfx.rlc.rlc_dram_ucode, - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_CP_MES) { - ucode->ucode_size = le32_to_cpu(mes_hdr->mes_ucode_size_bytes); - memcpy(ucode->kaddr, (void *)((uint8_t *)adev->mes.fw->data + - le32_to_cpu(mes_hdr->mes_ucode_offset_bytes)), - ucode->ucode_size); - } else if (ucode->ucode_id == AMDGPU_UCODE_ID_CP_MES_DATA) { - ucode->ucode_size = le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes); - memcpy(ucode->kaddr, (void *)((uint8_t *)adev->mes.fw->data + - le32_to_cpu(mes_hdr->mes_ucode_data_offset_bytes)), - ucode->ucode_size); + ucode_addr = (u8 *)ucode->fw->data + + le32_to_cpu(header->ucode_array_offset_bytes); + break; + case AMDGPU_UCODE_ID_DMCU_INTV: + ucode->ucode_size = le32_to_cpu(dmcu_hdr->intv_size_bytes); + ucode_addr = (u8 *)ucode->fw->data + + le32_to_cpu(header->ucode_array_offset_bytes) + + le32_to_cpu(dmcu_hdr->intv_offset_bytes); + break; + case AMDGPU_UCODE_ID_DMCUB: + ucode->ucode_size = le32_to_cpu(dmcub_hdr->inst_const_bytes); + ucode_addr = (u8 *)ucode->fw->data + + le32_to_cpu(header->ucode_array_offset_bytes); + break; + default: + ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes); + ucode_addr = (u8 *)ucode->fw->data + + le32_to_cpu(header->ucode_array_offset_bytes); + break; + } + } else { + ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes); + ucode_addr = (u8 *)ucode->fw->data + + le32_to_cpu(header->ucode_array_offset_bytes); } + memcpy(ucode->kaddr, ucode_addr, ucode->ucode_size); + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c index a90029ee9733..46264a4002f7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c @@ -23,6 +23,120 @@ #include "amdgpu_ras.h" +static int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev, + void *ras_error_status, + struct amdgpu_iv_entry *entry, + bool reset) +{ + struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; + struct amdgpu_ras *con = amdgpu_ras_get_context(adev); + int ret = 0; + + kgd2kfd_set_sram_ecc_flag(adev->kfd.dev); + ret = smu_get_ecc_info(&adev->smu, (void *)&(con->umc_ecc)); + if (ret == -EOPNOTSUPP) { + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->query_ras_error_count) + adev->umc.ras_funcs->query_ras_error_count(adev, ras_error_status); + + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->query_ras_error_address && + adev->umc.max_ras_err_cnt_per_query) { + err_data->err_addr = + kcalloc(adev->umc.max_ras_err_cnt_per_query, + sizeof(struct eeprom_table_record), GFP_KERNEL); + + /* still call query_ras_error_address to clear error status + * even NOMEM error is encountered + */ + if(!err_data->err_addr) + dev_warn(adev->dev, "Failed to alloc memory for " + "umc error address record!\n"); + + /* umc query_ras_error_address is also responsible for clearing + * error status + */ + adev->umc.ras_funcs->query_ras_error_address(adev, ras_error_status); + } + } else if (!ret) { + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->ecc_info_query_ras_error_count) + adev->umc.ras_funcs->ecc_info_query_ras_error_count(adev, ras_error_status); + + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->ecc_info_query_ras_error_address && + adev->umc.max_ras_err_cnt_per_query) { + err_data->err_addr = + kcalloc(adev->umc.max_ras_err_cnt_per_query, + sizeof(struct eeprom_table_record), GFP_KERNEL); + + /* still call query_ras_error_address to clear error status + * even NOMEM error is encountered + */ + if(!err_data->err_addr) + dev_warn(adev->dev, "Failed to alloc memory for " + "umc error address record!\n"); + + /* umc query_ras_error_address is also responsible for clearing + * error status + */ + adev->umc.ras_funcs->ecc_info_query_ras_error_address(adev, ras_error_status); + } + } + + /* only uncorrectable error needs gpu reset */ + if (err_data->ue_count) { + dev_info(adev->dev, "%ld uncorrectable hardware errors " + "detected in UMC block\n", + err_data->ue_count); + + if ((amdgpu_bad_page_threshold != 0) && + err_data->err_addr_cnt) { + amdgpu_ras_add_bad_pages(adev, err_data->err_addr, + err_data->err_addr_cnt); + amdgpu_ras_save_bad_pages(adev); + + if (adev->smu.ppt_funcs && adev->smu.ppt_funcs->send_hbm_bad_pages_num) + adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.ras_num_recs); + } + + if (reset) + amdgpu_ras_reset_gpu(adev); + } + + kfree(err_data->err_addr); + return AMDGPU_RAS_SUCCESS; +} + +int amdgpu_umc_poison_handler(struct amdgpu_device *adev, + void *ras_error_status, + bool reset) +{ + int ret; + struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; + struct ras_common_if head = { + .block = AMDGPU_RAS_BLOCK__UMC, + }; + struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head); + + ret = + amdgpu_umc_do_page_retirement(adev, ras_error_status, NULL, reset); + + if (ret == AMDGPU_RAS_SUCCESS && obj) { + obj->err_data.ue_count += err_data->ue_count; + obj->err_data.ce_count += err_data->ce_count; + } + + return ret; +} + +static int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev, + void *ras_error_status, + struct amdgpu_iv_entry *entry) +{ + return amdgpu_umc_do_page_retirement(adev, ras_error_status, entry, true); +} + int amdgpu_umc_ras_late_init(struct amdgpu_device *adev) { int r; @@ -88,61 +202,6 @@ void amdgpu_umc_ras_fini(struct amdgpu_device *adev) } } -int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev, - void *ras_error_status, - struct amdgpu_iv_entry *entry) -{ - struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; - struct amdgpu_ras *con = amdgpu_ras_get_context(adev); - - kgd2kfd_set_sram_ecc_flag(adev->kfd.dev); - if (adev->umc.ras_funcs && - adev->umc.ras_funcs->query_ras_error_count) - adev->umc.ras_funcs->query_ras_error_count(adev, ras_error_status); - - if (adev->umc.ras_funcs && - adev->umc.ras_funcs->query_ras_error_address && - adev->umc.max_ras_err_cnt_per_query) { - err_data->err_addr = - kcalloc(adev->umc.max_ras_err_cnt_per_query, - sizeof(struct eeprom_table_record), GFP_KERNEL); - - /* still call query_ras_error_address to clear error status - * even NOMEM error is encountered - */ - if(!err_data->err_addr) - dev_warn(adev->dev, "Failed to alloc memory for " - "umc error address record!\n"); - - /* umc query_ras_error_address is also responsible for clearing - * error status - */ - adev->umc.ras_funcs->query_ras_error_address(adev, ras_error_status); - } - - /* only uncorrectable error needs gpu reset */ - if (err_data->ue_count) { - dev_info(adev->dev, "%ld uncorrectable hardware errors " - "detected in UMC block\n", - err_data->ue_count); - - if ((amdgpu_bad_page_threshold != 0) && - err_data->err_addr_cnt) { - amdgpu_ras_add_bad_pages(adev, err_data->err_addr, - err_data->err_addr_cnt); - amdgpu_ras_save_bad_pages(adev); - - if (adev->smu.ppt_funcs && adev->smu.ppt_funcs->send_hbm_bad_pages_num) - adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.ras_num_recs); - } - - amdgpu_ras_reset_gpu(adev); - } - - kfree(err_data->err_addr); - return AMDGPU_RAS_SUCCESS; -} - int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h index e5a75fb788dd..b72194e8bfe5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h @@ -48,6 +48,11 @@ struct amdgpu_umc_ras_funcs { void *ras_error_status); void (*query_ras_error_address)(struct amdgpu_device *adev, void *ras_error_status); + bool (*query_ras_poison_mode)(struct amdgpu_device *adev); + void (*ecc_info_query_ras_error_count)(struct amdgpu_device *adev, + void *ras_error_status); + void (*ecc_info_query_ras_error_address)(struct amdgpu_device *adev, + void *ras_error_status); }; struct amdgpu_umc_funcs { @@ -73,9 +78,9 @@ struct amdgpu_umc { int amdgpu_umc_ras_late_init(struct amdgpu_device *adev); void amdgpu_umc_ras_fini(struct amdgpu_device *adev); -int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev, +int amdgpu_umc_poison_handler(struct amdgpu_device *adev, void *ras_error_status, - struct amdgpu_iv_entry *entry); + bool reset); int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h new file mode 100644 index 000000000000..919d9d401750 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h @@ -0,0 +1,51 @@ +/* + * Copyright 2021 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/ioctl.h> + +/* + * MMIO debugfs IOCTL structure + */ +struct amdgpu_debugfs_regs2_iocdata { + __u32 use_srbm, use_grbm, pg_lock; + struct { + __u32 se, sh, instance; + } grbm; + struct { + __u32 me, pipe, queue, vmid; + } srbm; +}; + +/* + * MMIO debugfs state data (per file* handle) + */ +struct amdgpu_debugfs_regs2_data { + struct amdgpu_device *adev; + struct mutex lock; + struct amdgpu_debugfs_regs2_iocdata id; +}; + +enum AMDGPU_DEBUGFS_REGS2_CMDS { + AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE=0, +}; + +#define AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE _IOWR(0x20, AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE, struct amdgpu_debugfs_regs2_iocdata) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c index d451c359606a..6f8de11a17f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c @@ -134,6 +134,51 @@ MODULE_FIRMWARE(FIRMWARE_VEGA12); MODULE_FIRMWARE(FIRMWARE_VEGA20); static void amdgpu_uvd_idle_work_handler(struct work_struct *work); +static void amdgpu_uvd_force_into_uvd_segment(struct amdgpu_bo *abo); + +static int amdgpu_uvd_create_msg_bo_helper(struct amdgpu_device *adev, + uint32_t size, + struct amdgpu_bo **bo_ptr) +{ + struct ttm_operation_ctx ctx = { true, false }; + struct amdgpu_bo *bo = NULL; + void *addr; + int r; + + r = amdgpu_bo_create_reserved(adev, size, PAGE_SIZE, + AMDGPU_GEM_DOMAIN_GTT, + &bo, NULL, &addr); + if (r) + return r; + + if (adev->uvd.address_64_bit) + goto succ; + + amdgpu_bo_kunmap(bo); + amdgpu_bo_unpin(bo); + amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_VRAM); + amdgpu_uvd_force_into_uvd_segment(bo); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); + if (r) + goto err; + r = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_VRAM); + if (r) + goto err_pin; + r = amdgpu_bo_kmap(bo, &addr); + if (r) + goto err_kmap; +succ: + amdgpu_bo_unreserve(bo); + *bo_ptr = bo; + return 0; +err_kmap: + amdgpu_bo_unpin(bo); +err_pin: +err: + amdgpu_bo_unreserve(bo); + amdgpu_bo_unref(&bo); + return r; +} int amdgpu_uvd_sw_init(struct amdgpu_device *adev) { @@ -302,6 +347,10 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev) if (!amdgpu_device_ip_block_version_cmp(adev, AMD_IP_BLOCK_TYPE_UVD, 5, 0)) adev->uvd.address_64_bit = true; + r = amdgpu_uvd_create_msg_bo_helper(adev, 128 << 10, &adev->uvd.ib_bo); + if (r) + return r; + switch (adev->asic_type) { case CHIP_TONGA: adev->uvd.use_ctx_buf = adev->uvd.fw_version >= FW_1_65_10; @@ -324,6 +373,7 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev) int amdgpu_uvd_sw_fini(struct amdgpu_device *adev) { + void *addr = amdgpu_bo_kptr(adev->uvd.ib_bo); int i, j; drm_sched_entity_destroy(&adev->uvd.entity); @@ -342,6 +392,7 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev) for (i = 0; i < AMDGPU_MAX_UVD_ENC_RINGS; ++i) amdgpu_ring_fini(&adev->uvd.inst[j].ring_enc[i]); } + amdgpu_bo_free_kernel(&adev->uvd.ib_bo, NULL, &addr); release_firmware(adev->uvd.fw); return 0; @@ -403,7 +454,7 @@ int amdgpu_uvd_suspend(struct amdgpu_device *adev) if (!adev->uvd.inst[j].saved_bo) return -ENOMEM; - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { /* re-write 0 since err_event_athub will corrupt VCPU buffer */ if (in_ras_intr) memset(adev->uvd.inst[j].saved_bo, 0, size); @@ -436,7 +487,7 @@ int amdgpu_uvd_resume(struct amdgpu_device *adev) ptr = adev->uvd.inst[i].cpu_addr; if (adev->uvd.inst[i].saved_bo != NULL) { - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { memcpy_toio(ptr, adev->uvd.inst[i].saved_bo, size); drm_dev_exit(idx); } @@ -449,7 +500,7 @@ int amdgpu_uvd_resume(struct amdgpu_device *adev) hdr = (const struct common_firmware_header *)adev->uvd.fw->data; if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { offset = le32_to_cpu(hdr->ucode_array_offset_bytes); - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { memcpy_toio(adev->uvd.inst[i].cpu_addr, adev->uvd.fw->data + offset, le32_to_cpu(hdr->ucode_size_bytes)); drm_dev_exit(idx); @@ -1080,23 +1131,10 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo, unsigned offset_idx = 0; unsigned offset[3] = { UVD_BASE_SI, 0, 0 }; - amdgpu_bo_kunmap(bo); - amdgpu_bo_unpin(bo); - - if (!ring->adev->uvd.address_64_bit) { - struct ttm_operation_ctx ctx = { true, false }; - - amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_VRAM); - amdgpu_uvd_force_into_uvd_segment(bo); - r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); - if (r) - goto err; - } - r = amdgpu_job_alloc_with_ib(adev, 64, direct ? AMDGPU_IB_POOL_DIRECT : AMDGPU_IB_POOL_DELAYED, &job); if (r) - goto err; + return r; if (adev->asic_type >= CHIP_VEGA10) { offset_idx = 1 + ring->me; @@ -1147,9 +1185,9 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo, goto err_free; } + amdgpu_bo_reserve(bo, true); amdgpu_bo_fence(bo, f, false); amdgpu_bo_unreserve(bo); - amdgpu_bo_unref(&bo); if (fence) *fence = dma_fence_get(f); @@ -1159,10 +1197,6 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo, err_free: amdgpu_job_free(job); - -err: - amdgpu_bo_unreserve(bo); - amdgpu_bo_unref(&bo); return r; } @@ -1173,16 +1207,11 @@ int amdgpu_uvd_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, struct dma_fence **fence) { struct amdgpu_device *adev = ring->adev; - struct amdgpu_bo *bo = NULL; + struct amdgpu_bo *bo = adev->uvd.ib_bo; uint32_t *msg; - int r, i; - - r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_GTT, - &bo, NULL, (void **)&msg); - if (r) - return r; + int i; + msg = amdgpu_bo_kptr(bo); /* stitch together an UVD create msg */ msg[0] = cpu_to_le32(0x00000de4); msg[1] = cpu_to_le32(0x00000000); @@ -1199,6 +1228,7 @@ int amdgpu_uvd_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, msg[i] = cpu_to_le32(0x0); return amdgpu_uvd_send_msg(ring, bo, true, fence); + } int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, @@ -1209,12 +1239,15 @@ int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, uint32_t *msg; int r, i; - r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_GTT, - &bo, NULL, (void **)&msg); - if (r) - return r; + if (direct) { + bo = adev->uvd.ib_bo; + } else { + r = amdgpu_uvd_create_msg_bo_helper(adev, 4096, &bo); + if (r) + return r; + } + msg = amdgpu_bo_kptr(bo); /* stitch together an UVD destroy msg */ msg[0] = cpu_to_le32(0x00000de4); msg[1] = cpu_to_le32(0x00000002); @@ -1223,7 +1256,12 @@ int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, for (i = 4; i < 1024; ++i) msg[i] = cpu_to_le32(0x0); - return amdgpu_uvd_send_msg(ring, bo, direct, fence); + r = amdgpu_uvd_send_msg(ring, bo, direct, fence); + + if (!direct) + amdgpu_bo_free_kernel(&bo, NULL, (void **)&msg); + + return r; } static void amdgpu_uvd_idle_work_handler(struct work_struct *work) @@ -1298,10 +1336,17 @@ int amdgpu_uvd_ring_test_ib(struct amdgpu_ring *ring, long timeout) struct dma_fence *fence; long r; - r = amdgpu_uvd_get_create_msg(ring, 1, NULL); + r = amdgpu_uvd_get_create_msg(ring, 1, &fence); if (r) goto error; + r = dma_fence_wait_timeout(fence, false, timeout); + dma_fence_put(fence); + if (r == 0) + r = -ETIMEDOUT; + if (r < 0) + goto error; + r = amdgpu_uvd_get_destroy_msg(ring, 1, true, &fence); if (r) goto error; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h index edbb8194ee81..76ac9699885d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h @@ -68,6 +68,7 @@ struct amdgpu_uvd { /* store image width to adjust nb memory state */ unsigned decode_image_width; uint32_t keyselect; + struct amdgpu_bo *ib_bo; }; int amdgpu_uvd_sw_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index 8e8dee9fac9f..344f711ad144 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -82,7 +82,6 @@ MODULE_FIRMWARE(FIRMWARE_VEGA20); static void amdgpu_vce_idle_work_handler(struct work_struct *work); static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, - struct amdgpu_bo *bo, struct dma_fence **fence); static int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, bool direct, struct dma_fence **fence); @@ -314,7 +313,7 @@ int amdgpu_vce_resume(struct amdgpu_device *adev) hdr = (const struct common_firmware_header *)adev->vce.fw->data; offset = le32_to_cpu(hdr->ucode_array_offset_bytes); - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { memcpy_toio(cpu_addr, adev->vce.fw->data + offset, adev->vce.fw->size - offset); drm_dev_exit(idx); @@ -435,18 +434,17 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp) * * @ring: ring we should submit the msg to * @handle: VCE session handle to use - * @bo: amdgpu object for which we query the offset * @fence: optional fence to return * * Open up a stream for HW test */ static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, - struct amdgpu_bo *bo, struct dma_fence **fence) { const unsigned ib_size_dw = 1024; struct amdgpu_job *job; struct amdgpu_ib *ib; + struct amdgpu_ib ib_msg; struct dma_fence *f = NULL; uint64_t addr; int i, r; @@ -456,9 +454,17 @@ static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, if (r) return r; - ib = &job->ibs[0]; + memset(&ib_msg, 0, sizeof(ib_msg)); + /* only one gpu page is needed, alloc +1 page to make addr aligned. */ + r = amdgpu_ib_get(ring->adev, NULL, AMDGPU_GPU_PAGE_SIZE * 2, + AMDGPU_IB_POOL_DIRECT, + &ib_msg); + if (r) + goto err; - addr = amdgpu_bo_gpu_offset(bo); + ib = &job->ibs[0]; + /* let addr point to page boundary */ + addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg.gpu_addr); /* stitch together an VCE create msg */ ib->length_dw = 0; @@ -498,6 +504,7 @@ static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, ib->ptr[i] = 0x0; r = amdgpu_job_submit_direct(job, ring, &f); + amdgpu_ib_free(ring->adev, &ib_msg, f); if (r) goto err; @@ -1134,20 +1141,13 @@ int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring) int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct dma_fence *fence = NULL; - struct amdgpu_bo *bo = NULL; long r; /* skip vce ring1/2 ib test for now, since it's not reliable */ if (ring != &ring->adev->vce.ring[0]) return 0; - r = amdgpu_bo_create_reserved(ring->adev, 512, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, - &bo, NULL, NULL); - if (r) - return r; - - r = amdgpu_vce_get_create_msg(ring, 1, bo, NULL); + r = amdgpu_vce_get_create_msg(ring, 1, NULL); if (r) goto error; @@ -1163,7 +1163,19 @@ int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring, long timeout) error: dma_fence_put(fence); - amdgpu_bo_unreserve(bo); - amdgpu_bo_free_kernel(&bo, NULL, NULL); return r; } + +enum amdgpu_ring_priority_level amdgpu_vce_get_ring_prio(int ring) +{ + switch(ring) { + case 0: + return AMDGPU_RING_PRIO_0; + case 1: + return AMDGPU_RING_PRIO_1; + case 2: + return AMDGPU_RING_PRIO_2; + default: + return AMDGPU_RING_PRIO_0; + } +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h index d6d83a3ec803..be4a6e773c5b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h @@ -71,5 +71,6 @@ void amdgpu_vce_ring_begin_use(struct amdgpu_ring *ring); void amdgpu_vce_ring_end_use(struct amdgpu_ring *ring); unsigned amdgpu_vce_ring_get_emit_ib_size(struct amdgpu_ring *ring); unsigned amdgpu_vce_ring_get_dma_frame_size(struct amdgpu_ring *ring); +enum amdgpu_ring_priority_level amdgpu_vce_get_ring_prio(int ring); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index 008a308a4eca..9a19a6a57b23 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -86,8 +86,9 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) for (i = 0; i < adev->vcn.num_vcn_inst; i++) atomic_set(&adev->vcn.inst[i].dpg_enc_submission_cnt, 0); - switch (adev->asic_type) { - case CHIP_RAVEN: + switch (adev->ip_versions[UVD_HWIP][0]) { + case IP_VERSION(1, 0, 0): + case IP_VERSION(1, 0, 1): if (adev->apu_flags & AMD_APU_IS_RAVEN2) fw_name = FIRMWARE_RAVEN2; else if (adev->apu_flags & AMD_APU_IS_PICASSO) @@ -95,13 +96,13 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) else fw_name = FIRMWARE_RAVEN; break; - case CHIP_ARCTURUS: + case IP_VERSION(2, 5, 0): fw_name = FIRMWARE_ARCTURUS; if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) adev->vcn.indirect_sram = true; break; - case CHIP_RENOIR: + case IP_VERSION(2, 2, 0): if (adev->apu_flags & AMD_APU_IS_RENOIR) fw_name = FIRMWARE_RENOIR; else @@ -111,58 +112,54 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) adev->vcn.indirect_sram = true; break; - case CHIP_ALDEBARAN: + case IP_VERSION(2, 6, 0): fw_name = FIRMWARE_ALDEBARAN; if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) adev->vcn.indirect_sram = true; break; - case CHIP_NAVI10: + case IP_VERSION(2, 0, 0): fw_name = FIRMWARE_NAVI10; if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) adev->vcn.indirect_sram = true; break; - case CHIP_NAVI14: - fw_name = FIRMWARE_NAVI14; - if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && - (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) - adev->vcn.indirect_sram = true; - break; - case CHIP_NAVI12: - fw_name = FIRMWARE_NAVI12; - if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && - (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) - adev->vcn.indirect_sram = true; - break; - case CHIP_SIENNA_CICHLID: - fw_name = FIRMWARE_SIENNA_CICHLID; + case IP_VERSION(2, 0, 2): + if (adev->asic_type == CHIP_NAVI12) + fw_name = FIRMWARE_NAVI12; + else + fw_name = FIRMWARE_NAVI14; if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) adev->vcn.indirect_sram = true; break; - case CHIP_NAVY_FLOUNDER: - fw_name = FIRMWARE_NAVY_FLOUNDER; + case IP_VERSION(3, 0, 0): + case IP_VERSION(3, 0, 64): + case IP_VERSION(3, 0, 192): + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0)) + fw_name = FIRMWARE_SIENNA_CICHLID; + else + fw_name = FIRMWARE_NAVY_FLOUNDER; if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) adev->vcn.indirect_sram = true; break; - case CHIP_VANGOGH: + case IP_VERSION(3, 0, 2): fw_name = FIRMWARE_VANGOGH; break; - case CHIP_DIMGREY_CAVEFISH: + case IP_VERSION(3, 0, 16): fw_name = FIRMWARE_DIMGREY_CAVEFISH; if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) adev->vcn.indirect_sram = true; break; - case CHIP_BEIGE_GOBY: + case IP_VERSION(3, 0, 33): fw_name = FIRMWARE_BEIGE_GOBY; if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) adev->vcn.indirect_sram = true; break; - case CHIP_YELLOW_CARP: + case IP_VERSION(3, 1, 1): fw_name = FIRMWARE_YELLOW_CARP; if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) @@ -289,20 +286,13 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev) bool amdgpu_vcn_is_disabled_vcn(struct amdgpu_device *adev, enum vcn_ring_type type, uint32_t vcn_instance) { bool ret = false; + int vcn_config = adev->vcn.vcn_config[vcn_instance]; - int major; - int minor; - int revision; - - /* if cannot find IP data, then this VCN does not exist */ - if (amdgpu_discovery_get_vcn_version(adev, vcn_instance, &major, &minor, &revision) != 0) - return true; - - if ((type == VCN_ENCODE_RING) && (revision & VCN_BLOCK_ENCODE_DISABLE_MASK)) { + if ((type == VCN_ENCODE_RING) && (vcn_config & VCN_BLOCK_ENCODE_DISABLE_MASK)) { ret = true; - } else if ((type == VCN_DECODE_RING) && (revision & VCN_BLOCK_DECODE_DISABLE_MASK)) { + } else if ((type == VCN_DECODE_RING) && (vcn_config & VCN_BLOCK_DECODE_DISABLE_MASK)) { ret = true; - } else if ((type == VCN_UNIFIED_RING) && (revision & VCN_BLOCK_QUEUE_DISABLE_MASK)) { + } else if ((type == VCN_UNIFIED_RING) && (vcn_config & VCN_BLOCK_QUEUE_DISABLE_MASK)) { ret = true; } @@ -330,7 +320,7 @@ int amdgpu_vcn_suspend(struct amdgpu_device *adev) if (!adev->vcn.inst[i].saved_bo) return -ENOMEM; - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { memcpy_fromio(adev->vcn.inst[i].saved_bo, ptr, size); drm_dev_exit(idx); } @@ -354,7 +344,7 @@ int amdgpu_vcn_resume(struct amdgpu_device *adev) ptr = adev->vcn.inst[i].cpu_addr; if (adev->vcn.inst[i].saved_bo != NULL) { - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { memcpy_toio(ptr, adev->vcn.inst[i].saved_bo, size); drm_dev_exit(idx); } @@ -367,7 +357,7 @@ int amdgpu_vcn_resume(struct amdgpu_device *adev) hdr = (const struct common_firmware_header *)adev->vcn.fw->data; if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { offset = le32_to_cpu(hdr->ucode_array_offset_bytes); - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { memcpy_toio(adev->vcn.inst[i].cpu_addr, adev->vcn.fw->data + offset, le32_to_cpu(hdr->ucode_size_bytes)); drm_dev_exit(idx); @@ -541,15 +531,14 @@ int amdgpu_vcn_dec_sw_ring_test_ring(struct amdgpu_ring *ring) } static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring, - struct amdgpu_bo *bo, + struct amdgpu_ib *ib_msg, struct dma_fence **fence) { struct amdgpu_device *adev = ring->adev; struct dma_fence *f = NULL; struct amdgpu_job *job; struct amdgpu_ib *ib; - uint64_t addr; - void *msg = NULL; + uint64_t addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg->gpu_addr); int i, r; r = amdgpu_job_alloc_with_ib(adev, 64, @@ -558,8 +547,6 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring, goto err; ib = &job->ibs[0]; - addr = amdgpu_bo_gpu_offset(bo); - msg = amdgpu_bo_kptr(bo); ib->ptr[0] = PACKET0(adev->vcn.internal.data0, 0); ib->ptr[1] = addr; ib->ptr[2] = PACKET0(adev->vcn.internal.data1, 0); @@ -576,9 +563,7 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring, if (r) goto err_free; - amdgpu_bo_fence(bo, f, false); - amdgpu_bo_unreserve(bo); - amdgpu_bo_free_kernel(&bo, NULL, (void **)&msg); + amdgpu_ib_free(adev, ib_msg, f); if (fence) *fence = dma_fence_get(f); @@ -588,27 +573,26 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring, err_free: amdgpu_job_free(job); - err: - amdgpu_bo_unreserve(bo); - amdgpu_bo_free_kernel(&bo, NULL, (void **)&msg); + amdgpu_ib_free(adev, ib_msg, f); return r; } static int amdgpu_vcn_dec_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, - struct amdgpu_bo **bo) + struct amdgpu_ib *ib) { struct amdgpu_device *adev = ring->adev; uint32_t *msg; int r, i; - *bo = NULL; - r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, - bo, NULL, (void **)&msg); + memset(ib, 0, sizeof(*ib)); + r = amdgpu_ib_get(adev, NULL, AMDGPU_GPU_PAGE_SIZE * 2, + AMDGPU_IB_POOL_DIRECT, + ib); if (r) return r; + msg = (uint32_t *)AMDGPU_GPU_PAGE_ALIGN((unsigned long)ib->ptr); msg[0] = cpu_to_le32(0x00000028); msg[1] = cpu_to_le32(0x00000038); msg[2] = cpu_to_le32(0x00000001); @@ -630,19 +614,20 @@ static int amdgpu_vcn_dec_get_create_msg(struct amdgpu_ring *ring, uint32_t hand } static int amdgpu_vcn_dec_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, - struct amdgpu_bo **bo) + struct amdgpu_ib *ib) { struct amdgpu_device *adev = ring->adev; uint32_t *msg; int r, i; - *bo = NULL; - r = amdgpu_bo_create_reserved(adev, 1024, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, - bo, NULL, (void **)&msg); + memset(ib, 0, sizeof(*ib)); + r = amdgpu_ib_get(adev, NULL, AMDGPU_GPU_PAGE_SIZE * 2, + AMDGPU_IB_POOL_DIRECT, + ib); if (r) return r; + msg = (uint32_t *)AMDGPU_GPU_PAGE_ALIGN((unsigned long)ib->ptr); msg[0] = cpu_to_le32(0x00000028); msg[1] = cpu_to_le32(0x00000018); msg[2] = cpu_to_le32(0x00000000); @@ -658,21 +643,21 @@ static int amdgpu_vcn_dec_get_destroy_msg(struct amdgpu_ring *ring, uint32_t han int amdgpu_vcn_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct dma_fence *fence = NULL; - struct amdgpu_bo *bo; + struct amdgpu_ib ib; long r; - r = amdgpu_vcn_dec_get_create_msg(ring, 1, &bo); + r = amdgpu_vcn_dec_get_create_msg(ring, 1, &ib); if (r) goto error; - r = amdgpu_vcn_dec_send_msg(ring, bo, NULL); + r = amdgpu_vcn_dec_send_msg(ring, &ib, NULL); if (r) goto error; - r = amdgpu_vcn_dec_get_destroy_msg(ring, 1, &bo); + r = amdgpu_vcn_dec_get_destroy_msg(ring, 1, &ib); if (r) goto error; - r = amdgpu_vcn_dec_send_msg(ring, bo, &fence); + r = amdgpu_vcn_dec_send_msg(ring, &ib, &fence); if (r) goto error; @@ -688,8 +673,8 @@ error: } static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring, - struct amdgpu_bo *bo, - struct dma_fence **fence) + struct amdgpu_ib *ib_msg, + struct dma_fence **fence) { struct amdgpu_vcn_decode_buffer *decode_buffer = NULL; const unsigned int ib_size_dw = 64; @@ -697,7 +682,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring, struct dma_fence *f = NULL; struct amdgpu_job *job; struct amdgpu_ib *ib; - uint64_t addr; + uint64_t addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg->gpu_addr); int i, r; r = amdgpu_job_alloc_with_ib(adev, ib_size_dw * 4, @@ -706,7 +691,6 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring, goto err; ib = &job->ibs[0]; - addr = amdgpu_bo_gpu_offset(bo); ib->length_dw = 0; ib->ptr[ib->length_dw++] = sizeof(struct amdgpu_vcn_decode_buffer) + 8; @@ -726,9 +710,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring, if (r) goto err_free; - amdgpu_bo_fence(bo, f, false); - amdgpu_bo_unreserve(bo); - amdgpu_bo_unref(&bo); + amdgpu_ib_free(adev, ib_msg, f); if (fence) *fence = dma_fence_get(f); @@ -738,31 +720,29 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring, err_free: amdgpu_job_free(job); - err: - amdgpu_bo_unreserve(bo); - amdgpu_bo_unref(&bo); + amdgpu_ib_free(adev, ib_msg, f); return r; } int amdgpu_vcn_dec_sw_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct dma_fence *fence = NULL; - struct amdgpu_bo *bo; + struct amdgpu_ib ib; long r; - r = amdgpu_vcn_dec_get_create_msg(ring, 1, &bo); + r = amdgpu_vcn_dec_get_create_msg(ring, 1, &ib); if (r) goto error; - r = amdgpu_vcn_dec_sw_send_msg(ring, bo, NULL); + r = amdgpu_vcn_dec_sw_send_msg(ring, &ib, NULL); if (r) goto error; - r = amdgpu_vcn_dec_get_destroy_msg(ring, 1, &bo); + r = amdgpu_vcn_dec_get_destroy_msg(ring, 1, &ib); if (r) goto error; - r = amdgpu_vcn_dec_sw_send_msg(ring, bo, &fence); + r = amdgpu_vcn_dec_sw_send_msg(ring, &ib, &fence); if (r) goto error; @@ -809,7 +789,7 @@ int amdgpu_vcn_enc_ring_test_ring(struct amdgpu_ring *ring) } static int amdgpu_vcn_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, - struct amdgpu_bo *bo, + struct amdgpu_ib *ib_msg, struct dma_fence **fence) { const unsigned ib_size_dw = 16; @@ -825,7 +805,7 @@ static int amdgpu_vcn_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t hand return r; ib = &job->ibs[0]; - addr = amdgpu_bo_gpu_offset(bo); + addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg->gpu_addr); ib->length_dw = 0; ib->ptr[ib->length_dw++] = 0x00000018; @@ -863,7 +843,7 @@ err: } static int amdgpu_vcn_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, - struct amdgpu_bo *bo, + struct amdgpu_ib *ib_msg, struct dma_fence **fence) { const unsigned ib_size_dw = 16; @@ -879,7 +859,7 @@ static int amdgpu_vcn_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t han return r; ib = &job->ibs[0]; - addr = amdgpu_bo_gpu_offset(bo); + addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg->gpu_addr); ib->length_dw = 0; ib->ptr[ib->length_dw++] = 0x00000018; @@ -918,21 +898,23 @@ err: int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) { + struct amdgpu_device *adev = ring->adev; struct dma_fence *fence = NULL; - struct amdgpu_bo *bo = NULL; + struct amdgpu_ib ib; long r; - r = amdgpu_bo_create_reserved(ring->adev, 128 * 1024, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, - &bo, NULL, NULL); + memset(&ib, 0, sizeof(ib)); + r = amdgpu_ib_get(adev, NULL, (128 << 10) + AMDGPU_GPU_PAGE_SIZE, + AMDGPU_IB_POOL_DIRECT, + &ib); if (r) return r; - r = amdgpu_vcn_enc_get_create_msg(ring, 1, bo, NULL); + r = amdgpu_vcn_enc_get_create_msg(ring, 1, &ib, NULL); if (r) goto error; - r = amdgpu_vcn_enc_get_destroy_msg(ring, 1, bo, &fence); + r = amdgpu_vcn_enc_get_destroy_msg(ring, 1, &ib, &fence); if (r) goto error; @@ -943,9 +925,49 @@ int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) r = 0; error: + amdgpu_ib_free(adev, &ib, fence); dma_fence_put(fence); - amdgpu_bo_unreserve(bo); - amdgpu_bo_free_kernel(&bo, NULL, NULL); return r; } + +enum amdgpu_ring_priority_level amdgpu_vcn_get_enc_ring_prio(int ring) +{ + switch(ring) { + case 0: + return AMDGPU_RING_PRIO_0; + case 1: + return AMDGPU_RING_PRIO_1; + case 2: + return AMDGPU_RING_PRIO_2; + default: + return AMDGPU_RING_PRIO_0; + } +} + +void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev) +{ + int i; + unsigned int idx; + + if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { + const struct common_firmware_header *hdr; + hdr = (const struct common_firmware_header *)adev->vcn.fw->data; + + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + if (adev->vcn.harvest_config & (1 << i)) + continue; + /* currently only support 2 FW instances */ + if (i >= 2) { + dev_info(adev->dev, "More then 2 VCN FW instances!\n"); + break; + } + idx = AMDGPU_UCODE_ID_VCN + i; + adev->firmware.ucode[idx].ucode_id = idx; + adev->firmware.ucode[idx].fw = adev->vcn.fw; + adev->firmware.fw_size += + ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); + } + dev_info(adev->dev, "Will use PSP to load VCN firmware\n"); + } +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index d74c62b49795..5d3728b027d3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -235,6 +235,7 @@ struct amdgpu_vcn { uint8_t num_vcn_inst; struct amdgpu_vcn_inst inst[AMDGPU_MAX_VCN_INSTANCES]; + uint8_t vcn_config[AMDGPU_MAX_VCN_INSTANCES]; struct amdgpu_vcn_reg internal; struct mutex vcn_pg_lock; struct mutex vcn1_jpeg1_workaround; @@ -308,4 +309,8 @@ int amdgpu_vcn_dec_sw_ring_test_ib(struct amdgpu_ring *ring, long timeout); int amdgpu_vcn_enc_ring_test_ring(struct amdgpu_ring *ring); int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout); +enum amdgpu_ring_priority_level amdgpu_vcn_get_enc_ring_prio(int ring); + +void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev); + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index ca058fbcccd4..f8e574cc0e22 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -283,17 +283,15 @@ static int amdgpu_virt_init_ras_err_handler_data(struct amdgpu_device *adev) *data = kmalloc(sizeof(struct amdgpu_virt_ras_err_handler_data), GFP_KERNEL); if (!*data) - return -ENOMEM; + goto data_failure; bps = kmalloc_array(align_space, sizeof((*data)->bps), GFP_KERNEL); - bps_bo = kmalloc_array(align_space, sizeof((*data)->bps_bo), GFP_KERNEL); + if (!bps) + goto bps_failure; - if (!bps || !bps_bo) { - kfree(bps); - kfree(bps_bo); - kfree(*data); - return -ENOMEM; - } + bps_bo = kmalloc_array(align_space, sizeof((*data)->bps_bo), GFP_KERNEL); + if (!bps_bo) + goto bps_bo_failure; (*data)->bps = bps; (*data)->bps_bo = bps_bo; @@ -303,6 +301,13 @@ static int amdgpu_virt_init_ras_err_handler_data(struct amdgpu_device *adev) virt->ras_init_done = true; return 0; + +bps_bo_failure: + kfree(bps); +bps_failure: + kfree(*data); +data_failure: + return -ENOMEM; } static void amdgpu_virt_ras_release_bp(struct amdgpu_device *adev) @@ -532,9 +537,12 @@ static void amdgpu_virt_populate_vf2pf_ucode_info(struct amdgpu_device *adev) POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_MEC, adev->gfx.mec_fw_version); POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_MEC2, adev->gfx.mec2_fw_version); POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_SOS, adev->psp.sos.fw_version); - POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_ASD, adev->psp.asd.fw_version); - POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_TA_RAS, adev->psp.ras.feature_version); - POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_TA_XGMI, adev->psp.xgmi.feature_version); + POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_ASD, + adev->psp.asd_context.bin_desc.fw_version); + POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_TA_RAS, + adev->psp.ras_context.context.bin_desc.fw_version); + POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_TA_XGMI, + adev->psp.xgmi_context.context.bin_desc.fw_version); POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_SMC, adev->pm.fw_version); POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_SDMA, adev->sdma.instance[0].fw_version); POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_SDMA2, adev->sdma.instance[1].fw_version); @@ -581,6 +589,7 @@ static int amdgpu_virt_write_vf2pf_data(struct amdgpu_device *adev) vf2pf_info->encode_usage = 0; vf2pf_info->decode_usage = 0; + vf2pf_info->dummy_page_addr = (uint64_t)adev->dummy_page_addr; vf2pf_info->checksum = amd_sriov_msg_checksum( vf2pf_info, vf2pf_info->header.size, 0, 0); @@ -613,19 +622,37 @@ void amdgpu_virt_fini_data_exchange(struct amdgpu_device *adev) void amdgpu_virt_init_data_exchange(struct amdgpu_device *adev) { - uint64_t bp_block_offset = 0; - uint32_t bp_block_size = 0; - struct amd_sriov_msg_pf2vf_info *pf2vf_v2 = NULL; - adev->virt.fw_reserve.p_pf2vf = NULL; adev->virt.fw_reserve.p_vf2pf = NULL; adev->virt.vf2pf_update_interval_ms = 0; - if (adev->mman.fw_vram_usage_va != NULL) { + if (adev->bios != NULL) { adev->virt.vf2pf_update_interval_ms = 2000; adev->virt.fw_reserve.p_pf2vf = (struct amd_sriov_msg_pf2vf_info_header *) + (adev->bios + (AMD_SRIOV_MSG_PF2VF_OFFSET_KB << 10)); + + amdgpu_virt_read_pf2vf_data(adev); + } + + if (adev->virt.vf2pf_update_interval_ms != 0) { + INIT_DELAYED_WORK(&adev->virt.vf2pf_work, amdgpu_virt_update_vf2pf_work_item); + schedule_delayed_work(&(adev->virt.vf2pf_work), msecs_to_jiffies(adev->virt.vf2pf_update_interval_ms)); + } +} + + +void amdgpu_virt_exchange_data(struct amdgpu_device *adev) +{ + uint64_t bp_block_offset = 0; + uint32_t bp_block_size = 0; + struct amd_sriov_msg_pf2vf_info *pf2vf_v2 = NULL; + + if (adev->mman.fw_vram_usage_va != NULL) { + + adev->virt.fw_reserve.p_pf2vf = + (struct amd_sriov_msg_pf2vf_info_header *) (adev->mman.fw_vram_usage_va + (AMD_SRIOV_MSG_PF2VF_OFFSET_KB << 10)); adev->virt.fw_reserve.p_vf2pf = (struct amd_sriov_msg_vf2pf_info_header *) @@ -654,16 +681,10 @@ void amdgpu_virt_init_data_exchange(struct amdgpu_device *adev) (adev->bios + (AMD_SRIOV_MSG_PF2VF_OFFSET_KB << 10)); amdgpu_virt_read_pf2vf_data(adev); - - return; - } - - if (adev->virt.vf2pf_update_interval_ms != 0) { - INIT_DELAYED_WORK(&adev->virt.vf2pf_work, amdgpu_virt_update_vf2pf_work_item); - schedule_delayed_work(&(adev->virt.vf2pf_work), adev->virt.vf2pf_update_interval_ms); } } + void amdgpu_detect_virtualization(struct amdgpu_device *adev) { uint32_t reg; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h index 8d4c20bb71c5..9adfb8d63280 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h @@ -308,6 +308,7 @@ int amdgpu_virt_alloc_mm_table(struct amdgpu_device *adev); void amdgpu_virt_free_mm_table(struct amdgpu_device *adev); void amdgpu_virt_release_ras_err_handler_data(struct amdgpu_device *adev); void amdgpu_virt_init_data_exchange(struct amdgpu_device *adev); +void amdgpu_virt_exchange_data(struct amdgpu_device *adev); void amdgpu_virt_fini_data_exchange(struct amdgpu_device *adev); void amdgpu_detect_virtualization(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c index ce982afeff91..2dcc68e04e84 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c @@ -16,6 +16,8 @@ #include "ivsrcid/ivsrcid_vislands30.h" #include "amdgpu_vkms.h" #include "amdgpu_display.h" +#include "atom.h" +#include "amdgpu_irq.h" /** * DOC: amdgpu_vkms @@ -41,16 +43,16 @@ static const u32 amdgpu_vkms_formats[] = { static enum hrtimer_restart amdgpu_vkms_vblank_simulate(struct hrtimer *timer) { - struct amdgpu_vkms_output *output = container_of(timer, - struct amdgpu_vkms_output, - vblank_hrtimer); - struct drm_crtc *crtc = &output->crtc; + struct amdgpu_crtc *amdgpu_crtc = container_of(timer, struct amdgpu_crtc, vblank_timer); + struct drm_crtc *crtc = &amdgpu_crtc->base; + struct amdgpu_vkms_output *output = drm_crtc_to_amdgpu_vkms_output(crtc); u64 ret_overrun; bool ret; - ret_overrun = hrtimer_forward_now(&output->vblank_hrtimer, + ret_overrun = hrtimer_forward_now(&amdgpu_crtc->vblank_timer, output->period_ns); - WARN_ON(ret_overrun != 1); + if (ret_overrun != 1) + DRM_WARN("%s: vblank timer overrun\n", __func__); ret = drm_crtc_handle_vblank(crtc); if (!ret) @@ -65,22 +67,21 @@ static int amdgpu_vkms_enable_vblank(struct drm_crtc *crtc) unsigned int pipe = drm_crtc_index(crtc); struct drm_vblank_crtc *vblank = &dev->vblank[pipe]; struct amdgpu_vkms_output *out = drm_crtc_to_amdgpu_vkms_output(crtc); + struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); drm_calc_timestamping_constants(crtc, &crtc->mode); - hrtimer_init(&out->vblank_hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); - out->vblank_hrtimer.function = &amdgpu_vkms_vblank_simulate; out->period_ns = ktime_set(0, vblank->framedur_ns); - hrtimer_start(&out->vblank_hrtimer, out->period_ns, HRTIMER_MODE_REL); + hrtimer_start(&amdgpu_crtc->vblank_timer, out->period_ns, HRTIMER_MODE_REL); return 0; } static void amdgpu_vkms_disable_vblank(struct drm_crtc *crtc) { - struct amdgpu_vkms_output *out = drm_crtc_to_amdgpu_vkms_output(crtc); + struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - hrtimer_cancel(&out->vblank_hrtimer); + hrtimer_cancel(&amdgpu_crtc->vblank_timer); } static bool amdgpu_vkms_get_vblank_timestamp(struct drm_crtc *crtc, @@ -92,13 +93,14 @@ static bool amdgpu_vkms_get_vblank_timestamp(struct drm_crtc *crtc, unsigned int pipe = crtc->index; struct amdgpu_vkms_output *output = drm_crtc_to_amdgpu_vkms_output(crtc); struct drm_vblank_crtc *vblank = &dev->vblank[pipe]; + struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); if (!READ_ONCE(vblank->enabled)) { *vblank_time = ktime_get(); return true; } - *vblank_time = READ_ONCE(output->vblank_hrtimer.node.expires); + *vblank_time = READ_ONCE(amdgpu_crtc->vblank_timer.node.expires); if (WARN_ON(*vblank_time == vblank->time)) return true; @@ -165,6 +167,8 @@ static const struct drm_crtc_helper_funcs amdgpu_vkms_crtc_helper_funcs = { static int amdgpu_vkms_crtc_init(struct drm_device *dev, struct drm_crtc *crtc, struct drm_plane *primary, struct drm_plane *cursor) { + struct amdgpu_device *adev = drm_to_adev(dev); + struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); int ret; ret = drm_crtc_init_with_planes(dev, crtc, primary, cursor, @@ -176,6 +180,17 @@ static int amdgpu_vkms_crtc_init(struct drm_device *dev, struct drm_crtc *crtc, drm_crtc_helper_add(crtc, &amdgpu_vkms_crtc_helper_funcs); + amdgpu_crtc->crtc_id = drm_crtc_index(crtc); + adev->mode_info.crtcs[drm_crtc_index(crtc)] = amdgpu_crtc; + + amdgpu_crtc->pll_id = ATOM_PPLL_INVALID; + amdgpu_crtc->encoder = NULL; + amdgpu_crtc->connector = NULL; + amdgpu_crtc->vsync_timer_enabled = AMDGPU_IRQ_STATE_DISABLE; + + hrtimer_init(&amdgpu_crtc->vblank_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + amdgpu_crtc->vblank_timer.function = &amdgpu_vkms_vblank_simulate; + return ret; } @@ -375,6 +390,7 @@ static struct drm_plane *amdgpu_vkms_plane_init(struct drm_device *dev, int index) { struct drm_plane *plane; + uint64_t modifiers[] = {DRM_FORMAT_MOD_LINEAR, DRM_FORMAT_MOD_INVALID}; int ret; plane = kzalloc(sizeof(*plane), GFP_KERNEL); @@ -385,7 +401,7 @@ static struct drm_plane *amdgpu_vkms_plane_init(struct drm_device *dev, &amdgpu_vkms_plane_funcs, amdgpu_vkms_formats, ARRAY_SIZE(amdgpu_vkms_formats), - NULL, type, NULL); + modifiers, type, NULL); if (ret) { kfree(plane); return ERR_PTR(ret); @@ -396,12 +412,12 @@ static struct drm_plane *amdgpu_vkms_plane_init(struct drm_device *dev, return plane; } -int amdgpu_vkms_output_init(struct drm_device *dev, - struct amdgpu_vkms_output *output, int index) +static int amdgpu_vkms_output_init(struct drm_device *dev, struct + amdgpu_vkms_output *output, int index) { struct drm_connector *connector = &output->connector; struct drm_encoder *encoder = &output->encoder; - struct drm_crtc *crtc = &output->crtc; + struct drm_crtc *crtc = &output->crtc.base; struct drm_plane *primary, *cursor = NULL; int ret; @@ -465,6 +481,11 @@ static int amdgpu_vkms_sw_init(void *handle) int r, i; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + adev->amdgpu_vkms_output = kcalloc(adev->mode_info.num_crtc, + sizeof(struct amdgpu_vkms_output), GFP_KERNEL); + if (!adev->amdgpu_vkms_output) + return -ENOMEM; + adev_to_drm(adev)->max_vblank_count = 0; adev_to_drm(adev)->mode_config.funcs = &amdgpu_vkms_mode_funcs; @@ -481,10 +502,6 @@ static int amdgpu_vkms_sw_init(void *handle) if (r) return r; - adev->amdgpu_vkms_output = kcalloc(adev->mode_info.num_crtc, sizeof(struct amdgpu_vkms_output), GFP_KERNEL); - if (!adev->amdgpu_vkms_output) - return -ENOMEM; - /* allocate crtcs, encoders, connectors */ for (i = 0; i < adev->mode_info.num_crtc; i++) { r = amdgpu_vkms_output_init(adev_to_drm(adev), &adev->amdgpu_vkms_output[i], i); @@ -507,12 +524,13 @@ static int amdgpu_vkms_sw_fini(void *handle) if (adev->mode_info.crtcs[i]) hrtimer_cancel(&adev->mode_info.crtcs[i]->vblank_timer); - kfree(adev->mode_info.bios_hardcoded_edid); - kfree(adev->amdgpu_vkms_output); - drm_kms_helper_poll_fini(adev_to_drm(adev)); + drm_mode_config_cleanup(adev_to_drm(adev)); adev->mode_info.mode_config_initialized = false; + + kfree(adev->mode_info.bios_hardcoded_edid); + kfree(adev->amdgpu_vkms_output); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.h index 97f1b79c0724..4f8722ff37c2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.h @@ -10,15 +10,14 @@ #define YRES_MAX 16384 #define drm_crtc_to_amdgpu_vkms_output(target) \ - container_of(target, struct amdgpu_vkms_output, crtc) + container_of(target, struct amdgpu_vkms_output, crtc.base) extern const struct amdgpu_ip_block_version amdgpu_vkms_ip_block; struct amdgpu_vkms_output { - struct drm_crtc crtc; + struct amdgpu_crtc crtc; struct drm_encoder encoder; struct drm_connector connector; - struct hrtimer vblank_hrtimer; ktime_t period_ns; struct drm_pending_vblank_event *event; }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 6b15cad78de9..b37fc7d7d2c7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -53,7 +53,7 @@ * can be mapped as snooped (cached system pages) or unsnooped * (uncached system pages). * Each VM has an ID associated with it and there is a page table - * associated with each VMID. When execting a command buffer, + * associated with each VMID. When executing a command buffer, * the kernel tells the the ring what VMID to use for that command * buffer. VMIDs are allocated dynamically as commands are submitted. * The userspace drivers maintain their own address space and the kernel @@ -800,7 +800,7 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, struct amdgpu_bo *bo = &vmbo->bo; unsigned entries, ats_entries; uint64_t addr; - int r; + int r, idx; /* Figure out our place in the hierarchy */ if (ancestor->parent) { @@ -845,9 +845,12 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, return r; } + if (!drm_dev_enter(adev_to_drm(adev), &idx)) + return -ENODEV; + r = vm->update_funcs->map_table(vmbo); if (r) - return r; + goto exit; memset(¶ms, 0, sizeof(params)); params.adev = adev; @@ -856,7 +859,7 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, r = vm->update_funcs->prepare(¶ms, NULL, AMDGPU_SYNC_EXPLICIT); if (r) - return r; + goto exit; addr = 0; if (ats_entries) { @@ -872,7 +875,7 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, r = vm->update_funcs->update(¶ms, vmbo, addr, 0, ats_entries, value, flags); if (r) - return r; + goto exit; addr += ats_entries * 8; } @@ -895,10 +898,13 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, r = vm->update_funcs->update(¶ms, vmbo, addr, 0, entries, value, flags); if (r) - return r; + goto exit; } - return vm->update_funcs->commit(¶ms, NULL); + r = vm->update_funcs->commit(¶ms, NULL); +exit: + drm_dev_exit(idx); + return r; } /** @@ -1384,11 +1390,14 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev, struct amdgpu_vm *vm, bool immediate) { struct amdgpu_vm_update_params params; - int r; + int r, idx; if (list_empty(&vm->relocated)) return 0; + if (!drm_dev_enter(adev_to_drm(adev), &idx)) + return -ENODEV; + memset(¶ms, 0, sizeof(params)); params.adev = adev; params.vm = vm; @@ -1396,7 +1405,7 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev, r = vm->update_funcs->prepare(¶ms, NULL, AMDGPU_SYNC_EXPLICIT); if (r) - return r; + goto exit; while (!list_empty(&vm->relocated)) { struct amdgpu_vm_bo_base *entry; @@ -1414,10 +1423,13 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev, r = vm->update_funcs->commit(¶ms, &vm->last_update); if (r) goto error; + drm_dev_exit(idx); return 0; error: amdgpu_vm_invalidate_pds(adev, vm); +exit: + drm_dev_exit(idx); return r; } @@ -1706,7 +1718,7 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, enum amdgpu_sync_mode sync_mode; int r, idx; - if (!drm_dev_enter(&adev->ddev, &idx)) + if (!drm_dev_enter(adev_to_drm(adev), &idx)) return -ENODEV; memset(¶ms, 0, sizeof(params)); @@ -2090,30 +2102,14 @@ static void amdgpu_vm_free_mapping(struct amdgpu_device *adev, static void amdgpu_vm_prt_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) { struct dma_resv *resv = vm->root.bo->tbo.base.resv; - struct dma_fence *excl, **shared; - unsigned i, shared_count; - int r; + struct dma_resv_iter cursor; + struct dma_fence *fence; - r = dma_resv_get_fences(resv, &excl, &shared_count, &shared); - if (r) { - /* Not enough memory to grab the fence list, as last resort - * block for all the fences to complete. - */ - dma_resv_wait_timeout(resv, true, false, - MAX_SCHEDULE_TIMEOUT); - return; - } - - /* Add a callback for each fence in the reservation object */ - amdgpu_vm_prt_get(adev); - amdgpu_vm_add_prt_cb(adev, excl); - - for (i = 0; i < shared_count; ++i) { + dma_resv_for_each_fence(&cursor, resv, true, fence) { + /* Add a callback for each fence in the reservation object */ amdgpu_vm_prt_get(adev); - amdgpu_vm_add_prt_cb(adev, shared[i]); + amdgpu_vm_add_prt_cb(adev, fence); } - - kfree(shared); } /** diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c index 978ac927ac11..a38c6a747fa4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c @@ -265,6 +265,11 @@ static ssize_t amdgpu_xgmi_show_error(struct device *dev, ficaa_pie_ctl_in = AMDGPU_XGMI_SET_FICAA(0x200); ficaa_pie_status_in = AMDGPU_XGMI_SET_FICAA(0x208); + if ((!adev->df.funcs) || + (!adev->df.funcs->get_fica) || + (!adev->df.funcs->set_fica)) + return -EINVAL; + fica_out = adev->df.funcs->get_fica(adev, ficaa_pie_ctl_in); if (fica_out != 0x1f) pr_err("xGMI error counters not enabled!\n"); @@ -386,6 +391,7 @@ struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev) "%s", "xgmi_hive_info"); if (ret) { dev_err(adev->dev, "XGMI: failed initializing kobject for xgmi hive\n"); + kobject_put(&hive->kobj); kfree(hive); hive = NULL; goto pro_end; @@ -806,9 +812,9 @@ static void amdgpu_xgmi_reset_ras_error_count(struct amdgpu_device *adev) for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++) pcs_clear_status(adev, xgmi23_pcs_err_status_reg_aldebaran[i]); - for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++) + for (i = 0; i < ARRAY_SIZE(xgmi3x16_pcs_err_status_reg_aldebaran); i++) pcs_clear_status(adev, - xgmi23_pcs_err_status_reg_aldebaran[i]); + xgmi3x16_pcs_err_status_reg_aldebaran[i]); for (i = 0; i < ARRAY_SIZE(walf_pcs_err_status_reg_aldebaran); i++) pcs_clear_status(adev, walf_pcs_err_status_reg_aldebaran[i]); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h index a434c71fde8e..7326b6c1b71c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h @@ -204,8 +204,10 @@ struct amd_sriov_msg_pf2vf_info { } mm_bw_management[AMD_SRIOV_MSG_RESERVE_VCN_INST]; /* UUID info */ struct amd_sriov_msg_uuid_info uuid_info; + /* pcie atomic Ops info */ + uint32_t pcie_atomic_ops_enabled_flags; /* reserved */ - uint32_t reserved[256 - 47]; + uint32_t reserved[256 - 48]; }; struct amd_sriov_msg_vf2pf_info_header { @@ -259,9 +261,10 @@ struct amd_sriov_msg_vf2pf_info { uint8_t id; uint32_t version; } ucode_info[AMD_SRIOV_MSG_RESERVE_UCODE]; + uint64_t dummy_page_addr; /* reserved */ - uint32_t reserved[256-68]; + uint32_t reserved[256-70]; }; /* mailbox message send from guest to host */ diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c index 3ac505d954c4..ab6a07e5e8c4 100644 --- a/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c @@ -77,10 +77,9 @@ int athub_v2_0_set_clockgating(struct amdgpu_device *adev, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: + switch (adev->ip_versions[ATHUB_HWIP][0]) { + case IP_VERSION(2, 0, 0): + case IP_VERSION(2, 0, 2): athub_v2_0_update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); athub_v2_0_update_medium_grain_light_sleep(adev, diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c index c12c2900732b..2edefd10e56c 100644 --- a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c +++ b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c @@ -70,11 +70,10 @@ int athub_v2_1_set_clockgating(struct amdgpu_device *adev, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[ATHUB_HWIP][0]) { + case IP_VERSION(2, 1, 0): + case IP_VERSION(2, 1, 1): + case IP_VERSION(2, 1, 2): athub_v2_1_update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); athub_v2_1_update_medium_grain_light_sleep(adev, state == AMD_CG_STATE_GATE); break; diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c b/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c index 6134ed964027..a92d86e12718 100644 --- a/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c +++ b/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c @@ -469,7 +469,7 @@ int amdgpu_atombios_encoder_get_encoder_mode(struct drm_encoder *encoder) if (amdgpu_connector->use_digital && (amdgpu_connector->audio == AMDGPU_AUDIO_ENABLE)) return ATOM_ENCODER_MODE_HDMI; - else if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector)) && + else if (connector->display_info.is_hdmi && (amdgpu_connector->audio == AMDGPU_AUDIO_AUTO)) return ATOM_ENCODER_MODE_HDMI; else if (amdgpu_connector->use_digital) @@ -488,7 +488,7 @@ int amdgpu_atombios_encoder_get_encoder_mode(struct drm_encoder *encoder) if (amdgpu_audio != 0) { if (amdgpu_connector->audio == AMDGPU_AUDIO_ENABLE) return ATOM_ENCODER_MODE_HDMI; - else if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector)) && + else if (connector->display_info.is_hdmi && (amdgpu_connector->audio == AMDGPU_AUDIO_AUTO)) return ATOM_ENCODER_MODE_HDMI; else @@ -506,7 +506,7 @@ int amdgpu_atombios_encoder_get_encoder_mode(struct drm_encoder *encoder) } else if (amdgpu_audio != 0) { if (amdgpu_connector->audio == AMDGPU_AUDIO_ENABLE) return ATOM_ENCODER_MODE_HDMI; - else if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector)) && + else if (connector->display_info.is_hdmi && (amdgpu_connector->audio == AMDGPU_AUDIO_AUTO)) return ATOM_ENCODER_MODE_HDMI; else diff --git a/drivers/gpu/drm/amd/amdgpu/beige_goby_reg_init.c b/drivers/gpu/drm/amd/amdgpu/beige_goby_reg_init.c deleted file mode 100644 index 608a113ce354..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/beige_goby_reg_init.c +++ /dev/null @@ -1,54 +0,0 @@ -/* - * Copyright 2020 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ -#include "amdgpu.h" -#include "nv.h" - -#include "soc15_common.h" -#include "soc15_hw_ip.h" -#include "beige_goby_ip_offset.h" - -int beige_goby_reg_base_init(struct amdgpu_device *adev) -{ - /* HW has more IP blocks, only initialize the block needed by driver */ - uint32_t i; - for (i = 0 ; i < MAX_INSTANCE ; ++i) { - adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i])); - adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i])); - adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i])); - adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIO_BASE.instance[i])); - adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i])); - adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i])); - adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(VCN0_BASE.instance[i])); - adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i])); - adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DCN_BASE.instance[i])); - adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i])); - adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA1_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA2_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA3_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i])); - adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i])); - } - return 0; -} diff --git a/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c b/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c deleted file mode 100644 index 58808814d8fb..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c +++ /dev/null @@ -1,51 +0,0 @@ -/* - * Copyright 2018 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ -#include "amdgpu.h" -#include "nv.h" - -#include "soc15_common.h" -#include "soc15_hw_ip.h" -#include "cyan_skillfish_ip_offset.h" - -int cyan_skillfish_reg_base_init(struct amdgpu_device *adev) -{ - /* HW has more IP blocks, only initialized the blocke needed by driver */ - uint32_t i; - for (i = 0 ; i < MAX_INSTANCE ; ++i) { - adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i])); - adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i])); - adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i])); - adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIO_BASE.instance[i])); - adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i])); - adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i])); - adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(UVD0_BASE.instance[i])); - adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i])); - adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DMU_BASE.instance[i])); - adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i])); - adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA1_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i])); - } - return 0; -} diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c index b200b9e722d9..8318ee8339f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c @@ -2092,22 +2092,18 @@ static int dce_v8_0_pick_dig_encoder(struct drm_encoder *encoder) return 1; else return 0; - break; case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: if (dig->linkb) return 3; else return 2; - break; case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: if (dig->linkb) return 5; else return 4; - break; case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3: return 6; - break; default: DRM_ERROR("invalid encoder_id: 0x%x\n", amdgpu_encoder->encoder_id); return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/df_v3_6.c b/drivers/gpu/drm/amd/amdgpu/df_v3_6.c index 14514a145c17..43c5e3ec9a39 100644 --- a/drivers/gpu/drm/amd/amdgpu/df_v3_6.c +++ b/drivers/gpu/drm/amd/amdgpu/df_v3_6.c @@ -637,6 +637,36 @@ static void df_v3_6_pmc_get_count(struct amdgpu_device *adev, } } +static bool df_v3_6_query_ras_poison_mode(struct amdgpu_device *adev) +{ + uint32_t hw_assert_msklo, hw_assert_mskhi; + uint32_t v0, v1, v28, v31; + + hw_assert_msklo = RREG32_SOC15(DF, 0, + mmDF_CS_UMC_AON0_HardwareAssertMaskLow); + hw_assert_mskhi = RREG32_SOC15(DF, 0, + mmDF_NCS_PG0_HardwareAssertMaskHigh); + + v0 = REG_GET_FIELD(hw_assert_msklo, + DF_CS_UMC_AON0_HardwareAssertMaskLow, HWAssertMsk0); + v1 = REG_GET_FIELD(hw_assert_msklo, + DF_CS_UMC_AON0_HardwareAssertMaskLow, HWAssertMsk1); + v28 = REG_GET_FIELD(hw_assert_mskhi, + DF_NCS_PG0_HardwareAssertMaskHigh, HWAssertMsk28); + v31 = REG_GET_FIELD(hw_assert_mskhi, + DF_NCS_PG0_HardwareAssertMaskHigh, HWAssertMsk31); + + if (v0 && v1 && v28 && v31) + return true; + else if (!v0 && !v1 && !v28 && !v31) + return false; + else { + dev_warn(adev->dev, "DF poison setting is inconsistent(%d:%d:%d:%d)!\n", + v0, v1, v28, v31); + return false; + } +} + const struct amdgpu_df_funcs df_v3_6_funcs = { .sw_init = df_v3_6_sw_init, .sw_fini = df_v3_6_sw_fini, @@ -651,4 +681,5 @@ const struct amdgpu_df_funcs df_v3_6_funcs = { .pmc_get_count = df_v3_6_pmc_get_count, .get_fica = df_v3_6_get_fica, .set_fica = df_v3_6_set_fica, + .query_ras_poison_mode = df_v3_6_query_ras_poison_mode, }; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c index 16dbe593cba2..dbe7442fb25c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c @@ -270,25 +270,6 @@ MODULE_FIRMWARE("amdgpu/cyan_skillfish2_mec.bin"); MODULE_FIRMWARE("amdgpu/cyan_skillfish2_mec2.bin"); MODULE_FIRMWARE("amdgpu/cyan_skillfish2_rlc.bin"); -static const struct soc15_reg_golden golden_settings_gc_10_0[] = -{ - SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_CAM_INDEX, 0xffffffff, 0x00000000), - /* TA_GRAD_ADJ_UCONFIG -> TA_GRAD_ADJ */ - SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_CAM_DATA, 0xffffffff, 0x2544c382), - /* VGT_TF_RING_SIZE_UMD -> VGT_TF_RING_SIZE */ - SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_CAM_DATA, 0xffffffff, 0x2262c24e), - /* VGT_HS_OFFCHIP_PARAM_UMD -> VGT_HS_OFFCHIP_PARAM */ - SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_CAM_DATA, 0xffffffff, 0x226cc24f), - /* VGT_TF_MEMORY_BASE_UMD -> VGT_TF_MEMORY_BASE */ - SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_CAM_DATA, 0xffffffff, 0x226ec250), - /* VGT_TF_MEMORY_BASE_HI_UMD -> VGT_TF_MEMORY_BASE_HI */ - SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_CAM_DATA, 0xffffffff, 0x2278c261), - /* VGT_ESGS_RING_SIZE_UMD -> VGT_ESGS_RING_SIZE */ - SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_CAM_DATA, 0xffffffff, 0x2232c240), - /* VGT_GSVS_RING_SIZE_UMD -> VGT_GSVS_RING_SIZE */ - SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_CAM_DATA, 0xffffffff, 0x2233c241), -}; - static const struct soc15_reg_golden golden_settings_gc_10_1[] = { SOC15_REG_GOLDEN_VALUE(GC, 0, mmCB_HW_CONTROL_4, 0xffffffff, 0x00400014), @@ -1537,7 +1518,7 @@ static u32 gfx_v10_rlcg_rw(struct amdgpu_device *adev, u32 offset, u32 v, uint32 scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3) * 4; - if (adev->asic_type >= CHIP_SIENNA_CICHLID) { + if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0)) { spare_int = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_0_Sienna_Cichlid_BASE_IDX] + mmRLC_SPARE_INT_0_Sienna_Cichlid) * 4; @@ -3727,18 +3708,18 @@ static void gfx_v10_0_set_kiq_pm4_funcs(struct amdgpu_device *adev) static void gfx_v10_0_init_spm_golden_registers(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_NAVI10: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): soc15_program_register_sequence(adev, golden_settings_gc_rlc_spm_10_0_nv10, (const u32)ARRAY_SIZE(golden_settings_gc_rlc_spm_10_0_nv10)); break; - case CHIP_NAVI14: + case IP_VERSION(10, 1, 1): soc15_program_register_sequence(adev, golden_settings_gc_rlc_spm_10_1_nv14, (const u32)ARRAY_SIZE(golden_settings_gc_rlc_spm_10_1_nv14)); break; - case CHIP_NAVI12: + case IP_VERSION(10, 1, 2): soc15_program_register_sequence(adev, golden_settings_gc_rlc_spm_10_1_2_nv12, (const u32)ARRAY_SIZE(golden_settings_gc_rlc_spm_10_1_2_nv12)); @@ -3750,8 +3731,8 @@ static void gfx_v10_0_init_spm_golden_registers(struct amdgpu_device *adev) static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_NAVI10: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): soc15_program_register_sequence(adev, golden_settings_gc_10_1, (const u32)ARRAY_SIZE(golden_settings_gc_10_1)); @@ -3759,7 +3740,7 @@ static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_gc_10_0_nv10, (const u32)ARRAY_SIZE(golden_settings_gc_10_0_nv10)); break; - case CHIP_NAVI14: + case IP_VERSION(10, 1, 1): soc15_program_register_sequence(adev, golden_settings_gc_10_1_1, (const u32)ARRAY_SIZE(golden_settings_gc_10_1_1)); @@ -3767,7 +3748,7 @@ static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_gc_10_1_nv14, (const u32)ARRAY_SIZE(golden_settings_gc_10_1_nv14)); break; - case CHIP_NAVI12: + case IP_VERSION(10, 1, 2): soc15_program_register_sequence(adev, golden_settings_gc_10_1_2, (const u32)ARRAY_SIZE(golden_settings_gc_10_1_2)); @@ -3775,7 +3756,7 @@ static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_gc_10_1_2_nv12, (const u32)ARRAY_SIZE(golden_settings_gc_10_1_2_nv12)); break; - case CHIP_SIENNA_CICHLID: + case IP_VERSION(10, 3, 0): soc15_program_register_sequence(adev, golden_settings_gc_10_3, (const u32)ARRAY_SIZE(golden_settings_gc_10_3)); @@ -3783,35 +3764,32 @@ static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_gc_10_3_sienna_cichlid, (const u32)ARRAY_SIZE(golden_settings_gc_10_3_sienna_cichlid)); break; - case CHIP_NAVY_FLOUNDER: + case IP_VERSION(10, 3, 2): soc15_program_register_sequence(adev, golden_settings_gc_10_3_2, (const u32)ARRAY_SIZE(golden_settings_gc_10_3_2)); break; - case CHIP_VANGOGH: + case IP_VERSION(10, 3, 1): soc15_program_register_sequence(adev, golden_settings_gc_10_3_vangogh, (const u32)ARRAY_SIZE(golden_settings_gc_10_3_vangogh)); break; - case CHIP_YELLOW_CARP: + case IP_VERSION(10, 3, 3): soc15_program_register_sequence(adev, golden_settings_gc_10_3_3, (const u32)ARRAY_SIZE(golden_settings_gc_10_3_3)); break; - case CHIP_DIMGREY_CAVEFISH: + case IP_VERSION(10, 3, 4): soc15_program_register_sequence(adev, golden_settings_gc_10_3_4, (const u32)ARRAY_SIZE(golden_settings_gc_10_3_4)); break; - case CHIP_BEIGE_GOBY: + case IP_VERSION(10, 3, 5): soc15_program_register_sequence(adev, golden_settings_gc_10_3_5, (const u32)ARRAY_SIZE(golden_settings_gc_10_3_5)); break; - case CHIP_CYAN_SKILLFISH: - soc15_program_register_sequence(adev, - golden_settings_gc_10_0, - (const u32)ARRAY_SIZE(golden_settings_gc_10_0)); + case IP_VERSION(10, 1, 3): soc15_program_register_sequence(adev, golden_settings_gc_10_0_cyan_skillfish, (const u32)ARRAY_SIZE(golden_settings_gc_10_0_cyan_skillfish)); @@ -3985,11 +3963,11 @@ static void gfx_v10_0_check_fw_write_wait(struct amdgpu_device *adev) { adev->gfx.cp_fw_write_wait = false; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI12: - case CHIP_NAVI14: - case CHIP_CYAN_SKILLFISH: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 3): if ((adev->gfx.me_fw_version >= 0x00000046) && (adev->gfx.me_feature_version >= 27) && (adev->gfx.pfp_fw_version >= 0x00000068) && @@ -3998,12 +3976,12 @@ static void gfx_v10_0_check_fw_write_wait(struct amdgpu_device *adev) (adev->gfx.mec_feature_version >= 27)) adev->gfx.cp_fw_write_wait = true; break; - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): adev->gfx.cp_fw_write_wait = true; break; default: @@ -4066,8 +4044,8 @@ static bool gfx_v10_0_navi10_gfxoff_should_enable(struct amdgpu_device *adev) static void gfx_v10_0_check_gfxoff_flag(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_NAVI10: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): if (!gfx_v10_0_navi10_gfxoff_should_enable(adev)) adev->pm.pp_feature &= ~PP_GFXOFF_MASK; break; @@ -4093,38 +4071,38 @@ static int gfx_v10_0_init_microcode(struct amdgpu_device *adev) DRM_DEBUG("\n"); - switch (adev->asic_type) { - case CHIP_NAVI10: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): chip_name = "navi10"; break; - case CHIP_NAVI14: + case IP_VERSION(10, 1, 1): chip_name = "navi14"; if (!(adev->pdev->device == 0x7340 && adev->pdev->revision != 0x00)) wks = "_wks"; break; - case CHIP_NAVI12: + case IP_VERSION(10, 1, 2): chip_name = "navi12"; break; - case CHIP_SIENNA_CICHLID: + case IP_VERSION(10, 3, 0): chip_name = "sienna_cichlid"; break; - case CHIP_NAVY_FLOUNDER: + case IP_VERSION(10, 3, 2): chip_name = "navy_flounder"; break; - case CHIP_VANGOGH: + case IP_VERSION(10, 3, 1): chip_name = "vangogh"; break; - case CHIP_DIMGREY_CAVEFISH: + case IP_VERSION(10, 3, 4): chip_name = "dimgrey_cavefish"; break; - case CHIP_BEIGE_GOBY: + case IP_VERSION(10, 3, 5): chip_name = "beige_goby"; break; - case CHIP_YELLOW_CARP: + case IP_VERSION(10, 3, 3): chip_name = "yellow_carp"; break; - case CHIP_CYAN_SKILLFISH: + case IP_VERSION(10, 1, 3): if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2) chip_name = "cyan_skillfish2"; else @@ -4684,10 +4662,10 @@ static void gfx_v10_0_gpu_early_init(struct amdgpu_device *adev) adev->gfx.funcs = &gfx_v10_0_gfx_funcs; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): 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; @@ -4695,12 +4673,12 @@ static void gfx_v10_0_gpu_early_init(struct amdgpu_device *adev) adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0; gb_addr_config = RREG32_SOC15(GC, 0, mmGB_ADDR_CONFIG); break; - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 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; @@ -4710,7 +4688,7 @@ static void gfx_v10_0_gpu_early_init(struct amdgpu_device *adev) adev->gfx.config.gb_addr_config_fields.num_pkrs = 1 << REG_GET_FIELD(gb_addr_config, GB_ADDR_CONFIG, NUM_PKRS); break; - case CHIP_CYAN_SKILLFISH: + case IP_VERSION(10, 1, 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; @@ -4818,11 +4796,11 @@ static int gfx_v10_0_sw_init(void *handle) struct amdgpu_kiq *kiq; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_CYAN_SKILLFISH: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 3): adev->gfx.me.num_me = 1; adev->gfx.me.num_pipe_per_me = 1; adev->gfx.me.num_queue_per_pipe = 1; @@ -4830,12 +4808,12 @@ static int gfx_v10_0_sw_init(void *handle) adev->gfx.mec.num_pipe_per_mec = 4; adev->gfx.mec.num_queue_per_pipe = 8; break; - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): adev->gfx.me.num_me = 1; adev->gfx.me.num_pipe_per_me = 1; adev->gfx.me.num_queue_per_pipe = 1; @@ -5068,8 +5046,8 @@ static void gfx_v10_0_setup_rb(struct amdgpu_device *adev) for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { bitmap = i * adev->gfx.config.max_sh_per_se + j; - if (((adev->asic_type == CHIP_SIENNA_CICHLID) || - (adev->asic_type == CHIP_YELLOW_CARP)) && + if (((adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0)) || + (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 3))) && ((gfx_v10_3_get_disabled_sa(adev) >> bitmap) & 1)) continue; gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff); @@ -5096,7 +5074,7 @@ static u32 gfx_v10_0_init_pa_sc_tile_steering_override(struct amdgpu_device *ade /* for ASICs that integrates GFX v10.3 * pa_sc_tile_steering_override should be set to 0 */ - if (adev->asic_type >= CHIP_SIENNA_CICHLID) + if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0)) return 0; /* init num_sc */ @@ -5249,7 +5227,7 @@ static void gfx_v10_0_get_tcc_info(struct amdgpu_device *adev) /* TCCs are global (not instanced). */ uint32_t tcc_disable; - if (adev->asic_type >= CHIP_SIENNA_CICHLID) { + if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0)) { tcc_disable = RREG32_SOC15(GC, 0, mmCGTS_TCC_DISABLE_gc_10_3) | RREG32_SOC15(GC, 0, mmCGTS_USER_TCC_DISABLE_gc_10_3); } else { @@ -5326,7 +5304,7 @@ static int gfx_v10_0_init_csb(struct amdgpu_device *adev) adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr); /* csib */ - if (adev->asic_type == CHIP_NAVI12) { + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2)) { WREG32_SOC15_RLC(GC, 0, mmRLC_CSIB_ADDR_HI, adev->gfx.rlc.clear_state_gpu_addr >> 32); WREG32_SOC15_RLC(GC, 0, mmRLC_CSIB_ADDR_LO, @@ -5948,7 +5926,7 @@ static int gfx_v10_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable) tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_HALT, enable ? 0 : 1); tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, CE_HALT, enable ? 0 : 1); - if (adev->asic_type == CHIP_NAVI12) { + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2)) { WREG32_SOC15_RLC(GC, 0, mmCP_ME_CNTL, tmp); } else { WREG32_SOC15(GC, 0, mmCP_ME_CNTL, tmp); @@ -6337,13 +6315,13 @@ static void gfx_v10_0_cp_gfx_set_doorbell(struct amdgpu_device *adev, } WREG32_SOC15(GC, 0, mmCP_RB_DOORBELL_CONTROL, tmp); } - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): tmp = REG_SET_FIELD(0, CP_RB_DOORBELL_RANGE_LOWER, DOORBELL_RANGE_LOWER_Sienna_Cichlid, ring->doorbell_index); WREG32_SOC15(GC, 0, mmCP_RB_DOORBELL_RANGE_LOWER, tmp); @@ -6474,13 +6452,13 @@ static int gfx_v10_0_cp_gfx_resume(struct amdgpu_device *adev) static void gfx_v10_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) { if (enable) { - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): WREG32_SOC15(GC, 0, mmCP_MEC_CNTL_Sienna_Cichlid, 0); break; default: @@ -6488,13 +6466,13 @@ static void gfx_v10_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) break; } } else { - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): WREG32_SOC15(GC, 0, mmCP_MEC_CNTL_Sienna_Cichlid, (CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK)); @@ -6586,13 +6564,13 @@ static void gfx_v10_0_kiq_setting(struct amdgpu_ring *ring) struct amdgpu_device *adev = ring->adev; /* tell RLC which is KIQ queue */ - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); @@ -7303,11 +7281,11 @@ static bool gfx_v10_0_check_grbm_cam_remapping(struct amdgpu_device *adev) /* check if mmVGT_ESGS_RING_SIZE_UMD * has been remapped to mmVGT_ESGS_RING_SIZE */ - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): data = RREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_Sienna_Cichlid); WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_Sienna_Cichlid, 0); WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_UMD, pattern); @@ -7320,8 +7298,8 @@ static bool gfx_v10_0_check_grbm_cam_remapping(struct amdgpu_device *adev) return false; } break; - case CHIP_VANGOGH: - case CHIP_YELLOW_CARP: + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 3): return true; default: data = RREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE); @@ -7350,13 +7328,13 @@ static void gfx_v10_0_setup_grbm_cam_remapping(struct amdgpu_device *adev) * index will auto-inc after each data writting */ WREG32_SOC15(GC, 0, mmGRBM_CAM_INDEX, 0); - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): /* mmVGT_TF_RING_SIZE_UMD -> mmVGT_TF_RING_SIZE */ data = (SOC15_REG_OFFSET(GC, 0, mmVGT_TF_RING_SIZE_UMD) << GRBM_CAM_DATA__CAM_ADDR__SHIFT) | @@ -7520,19 +7498,19 @@ static int gfx_v10_0_hw_init(void *handle) * init golden registers and rlc resume may override some registers, * reconfig them here */ - if (adev->asic_type == CHIP_NAVI10 || - adev->asic_type == CHIP_NAVI14 || - adev->asic_type == CHIP_NAVI12) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 10) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 1) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2)) gfx_v10_0_tcp_harvest(adev); r = gfx_v10_0_cp_resume(adev); if (r) return r; - if (adev->asic_type == CHIP_SIENNA_CICHLID) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0)) gfx_v10_3_program_pbb_mode(adev); - if (adev->asic_type >= CHIP_SIENNA_CICHLID) + if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0)) gfx_v10_3_set_power_brake_sequence(adev); return r; @@ -7584,7 +7562,7 @@ static int gfx_v10_0_hw_fini(void *handle) if (amdgpu_sriov_vf(adev)) { gfx_v10_0_cp_gfx_enable(adev, false); /* Program KIQ position of RLC_CP_SCHEDULERS during destroy */ - if (adev->asic_type >= CHIP_SIENNA_CICHLID) { + if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0)) { tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid); tmp &= 0xffffff00; WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp); @@ -7670,13 +7648,13 @@ static int gfx_v10_0_soft_reset(void *handle) /* GRBM_STATUS2 */ tmp = RREG32_SOC15(GC, 0, mmGRBM_STATUS2); - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): if (REG_GET_FIELD(tmp, GRBM_STATUS2, RLC_BUSY_Sienna_Cichlid)) grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, @@ -7726,11 +7704,22 @@ static uint64_t gfx_v10_0_get_gpu_clock_counter(struct amdgpu_device *adev) { uint64_t clock, clock_lo, clock_hi, hi_check; - switch (adev->asic_type) { - case CHIP_VANGOGH: - case CHIP_YELLOW_CARP: - clock = (uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Vangogh) | - ((uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Vangogh) << 32ULL); + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 3): + preempt_disable(); + clock_hi = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Vangogh); + clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Vangogh); + hi_check = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Vangogh); + /* The SMUIO TSC clock frequency is 100MHz, which sets 32-bit carry over + * roughly every 42 seconds. + */ + if (hi_check != clock_hi) { + clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Vangogh); + clock_hi = hi_check; + } + preempt_enable(); + clock = clock_lo | (clock_hi << 32ULL); break; default: preempt_disable(); @@ -7784,19 +7773,19 @@ static int gfx_v10_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_CYAN_SKILLFISH: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 3): adev->gfx.num_gfx_rings = GFX10_NUM_GFX_RINGS_NV1X; break; - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): adev->gfx.num_gfx_rings = GFX10_NUM_GFX_RINGS_Sienna_Cichlid; break; default: @@ -7848,13 +7837,13 @@ static void gfx_v10_0_set_safe_mode(struct amdgpu_device *adev) data = RLC_SAFE_MODE__CMD_MASK; data |= (1 << RLC_SAFE_MODE__MESSAGE__SHIFT); - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): WREG32_SOC15(GC, 0, mmRLC_SAFE_MODE_Sienna_Cichlid, data); /* wait for RLC_SAFE_MODE */ @@ -7884,13 +7873,13 @@ static void gfx_v10_0_unset_safe_mode(struct amdgpu_device *adev) uint32_t data; data = RLC_SAFE_MODE__CMD_MASK; - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): WREG32_SOC15(GC, 0, mmRLC_SAFE_MODE_Sienna_Cichlid, data); break; default: @@ -8193,7 +8182,7 @@ static void gfx_v10_0_apply_medium_grain_clock_gating_workaround(struct amdgpu_d mmCGTS_SA1_QUAD1_SM_CTRL_REG }; - if (adev->asic_type == CHIP_NAVI12) { + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2)) { for (i = 0; i < ARRAY_SIZE(tcp_ctrl_regs_nv12); i++) { reg_idx = adev->reg_offset[GC_HWIP][0][mmCGTS_SA0_WGP00_CU0_TCP_CTRL_REG_BASE_IDX] + tcp_ctrl_regs_nv12[i]; @@ -8238,8 +8227,9 @@ static int gfx_v10_0_update_gfx_clock_gating(struct amdgpu_device *adev, /* === CGCG + CGLS === */ gfx_v10_0_update_coarse_grain_clock_gating(adev, enable); - if ((adev->asic_type >= CHIP_NAVI10) && - (adev->asic_type <= CHIP_NAVI12)) + if ((adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 10)) || + (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 1)) || + (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2))) gfx_v10_0_apply_medium_grain_clock_gating_workaround(adev); } else { /* CGCG/CGLS should be disabled before MGCG/MGLS @@ -8270,6 +8260,9 @@ static int gfx_v10_0_update_gfx_clock_gating(struct amdgpu_device *adev, static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid) { u32 reg, data; + + amdgpu_gfx_off_ctrl(adev, false); + /* not for *_SOC15 */ reg = SOC15_REG_OFFSET(GC, 0, mmRLC_SPM_MC_CNTL); if (amdgpu_sriov_is_pp_one_vf(adev)) @@ -8284,6 +8277,8 @@ static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid) WREG32_SOC15_NO_KIQ(GC, 0, mmRLC_SPM_MC_CNTL, data); else WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data); + + amdgpu_gfx_off_ctrl(adev, true); } static bool gfx_v10_0_check_rlcg_range(struct amdgpu_device *adev, @@ -8335,15 +8330,12 @@ static void gfx_v10_cntl_power_gating(struct amdgpu_device *adev, bool enable) * Power/performance team will optimize it and might give a new value later. */ if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG)) { - switch (adev->asic_type) { - case CHIP_VANGOGH: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 3): data = 0x4E20 & RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK_Vangogh; WREG32_SOC15(GC, 0, mmRLC_PG_DELAY_3, data); break; - case CHIP_YELLOW_CARP: - data = 0x1388 & RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK_Vangogh; - WREG32_SOC15(GC, 0, mmRLC_PG_DELAY_3, data); - break; default: break; } @@ -8399,18 +8391,18 @@ static int gfx_v10_0_set_powergating_state(void *handle, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): amdgpu_gfx_off_ctrl(adev, enable); break; - case CHIP_VANGOGH: - case CHIP_YELLOW_CARP: + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 3): gfx_v10_cntl_pg(adev, enable); amdgpu_gfx_off_ctrl(adev, enable); break; @@ -8428,16 +8420,16 @@ static int gfx_v10_0_set_clockgating_state(void *handle, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): gfx_v10_0_update_gfx_clock_gating(adev, state == AMD_CG_STATE_GATE); break; @@ -9541,19 +9533,19 @@ static void gfx_v10_0_set_irq_funcs(struct amdgpu_device *adev) static void gfx_v10_0_set_rlc_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: - case CHIP_CYAN_SKILLFISH: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): adev->gfx.rlc.funcs = &gfx_v10_0_rlc_funcs; break; - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 3, 0): adev->gfx.rlc.funcs = &gfx_v10_0_rlc_funcs_sriov; break; default: @@ -9641,8 +9633,8 @@ static int gfx_v10_0_get_cu_info(struct amdgpu_device *adev, for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { bitmap = i * adev->gfx.config.max_sh_per_se + j; - if (((adev->asic_type == CHIP_SIENNA_CICHLID) || - (adev->asic_type == CHIP_YELLOW_CARP)) && + if (((adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0)) || + (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 3))) && ((gfx_v10_3_get_disabled_sa(adev) >> bitmap) & 1)) continue; mask = 1; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index 37b4a3db6360..d17a6f399347 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -3575,12 +3575,16 @@ static void gfx_v7_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid) { u32 data; + amdgpu_gfx_off_ctrl(adev, false); + data = RREG32(mmRLC_SPM_VMID); data &= ~RLC_SPM_VMID__RLC_SPM_VMID_MASK; data |= (vmid & RLC_SPM_VMID__RLC_SPM_VMID_MASK) << RLC_SPM_VMID__RLC_SPM_VMID__SHIFT; WREG32(mmRLC_SPM_VMID, data); + + amdgpu_gfx_off_ctrl(adev, true); } static void gfx_v7_0_enable_cgcg(struct amdgpu_device *adev, bool enable) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index e0302c23e9a7..5f112efda634 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -5624,6 +5624,8 @@ static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid) { u32 data; + amdgpu_gfx_off_ctrl(adev, false); + if (amdgpu_sriov_is_pp_one_vf(adev)) data = RREG32_NO_KIQ(mmRLC_SPM_VMID); else @@ -5636,6 +5638,8 @@ static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid) WREG32_NO_KIQ(mmRLC_SPM_VMID, data); else WREG32(mmRLC_SPM_VMID, data); + + amdgpu_gfx_off_ctrl(adev, true); } static const struct amdgpu_rlc_funcs iceland_rlc_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index 025184a556ee..9189fb85a4dd 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -63,6 +63,13 @@ #define mmGCEA_PROBE_MAP 0x070c #define mmGCEA_PROBE_MAP_BASE_IDX 0 +#define GFX9_RLCG_GC_WRITE_OLD (0x8 << 28) +#define GFX9_RLCG_GC_WRITE (0x0 << 28) +#define GFX9_RLCG_GC_READ (0x1 << 28) +#define GFX9_RLCG_VFGATE_DISABLED 0x4000000 +#define GFX9_RLCG_WRONG_OPERATION_TYPE 0x2000000 +#define GFX9_RLCG_NOT_IN_RANGE 0x1000000 + MODULE_FIRMWARE("amdgpu/vega10_ce.bin"); MODULE_FIRMWARE("amdgpu/vega10_pfp.bin"); MODULE_FIRMWARE("amdgpu/vega10_me.bin"); @@ -140,6 +147,11 @@ MODULE_FIRMWARE("amdgpu/aldebaran_rlc.bin"); #define mmTCP_CHAN_STEER_5_ARCT 0x0b0c #define mmTCP_CHAN_STEER_5_ARCT_BASE_IDX 0 +#define mmGOLDEN_TSC_COUNT_UPPER_Renoir 0x0025 +#define mmGOLDEN_TSC_COUNT_UPPER_Renoir_BASE_IDX 1 +#define mmGOLDEN_TSC_COUNT_LOWER_Renoir 0x0026 +#define mmGOLDEN_TSC_COUNT_LOWER_Renoir_BASE_IDX 1 + enum ta_ras_gfx_subblock { /*CPC*/ TA_RAS_BLOCK__GFX_CPC_INDEX_START = 0, @@ -734,7 +746,7 @@ static const u32 GFX_RLC_SRM_INDEX_CNTL_DATA_OFFSETS[] = mmRLC_SRM_INDEX_CNTL_DATA_7 - mmRLC_SRM_INDEX_CNTL_DATA_0, }; -static void gfx_v9_0_rlcg_w(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag) +static u32 gfx_v9_0_rlcg_rw(struct amdgpu_device *adev, u32 offset, u32 v, uint32_t flag) { static void *scratch_reg0; static void *scratch_reg1; @@ -743,21 +755,20 @@ static void gfx_v9_0_rlcg_w(struct amdgpu_device *adev, u32 offset, u32 v, u32 f static void *spare_int; static uint32_t grbm_cntl; static uint32_t grbm_idx; + uint32_t i = 0; + uint32_t retries = 50000; + u32 ret = 0; + u32 tmp; scratch_reg0 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0)*4; scratch_reg1 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1)*4; - scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG2)*4; - scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3)*4; + scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG2_BASE_IDX] + mmSCRATCH_REG2)*4; + scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG3_BASE_IDX] + mmSCRATCH_REG3)*4; spare_int = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] + mmRLC_SPARE_INT)*4; grbm_cntl = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_CNTL_BASE_IDX] + mmGRBM_GFX_CNTL; grbm_idx = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_INDEX_BASE_IDX] + mmGRBM_GFX_INDEX; - if (amdgpu_sriov_runtime(adev)) { - pr_err("shouldn't call rlcg write register during runtime\n"); - return; - } - if (offset == grbm_cntl || offset == grbm_idx) { if (offset == grbm_cntl) writel(v, scratch_reg2); @@ -766,41 +777,95 @@ static void gfx_v9_0_rlcg_w(struct amdgpu_device *adev, u32 offset, u32 v, u32 f writel(v, ((void __iomem *)adev->rmmio) + (offset * 4)); } else { - uint32_t i = 0; - uint32_t retries = 50000; - + /* + * SCRATCH_REG0 = read/write value + * SCRATCH_REG1[30:28] = command + * SCRATCH_REG1[19:0] = address in dword + * SCRATCH_REG1[26:24] = Error reporting + */ writel(v, scratch_reg0); - writel(offset | 0x80000000, scratch_reg1); + writel(offset | flag, scratch_reg1); writel(1, spare_int); - for (i = 0; i < retries; i++) { - u32 tmp; + for (i = 0; i < retries; i++) { tmp = readl(scratch_reg1); - if (!(tmp & 0x80000000)) + if (!(tmp & flag)) break; udelay(10); } - if (i >= retries) - pr_err("timeout: rlcg program reg:0x%05x failed !\n", offset); + + if (i >= retries) { + if (amdgpu_sriov_reg_indirect_gc(adev)) { + if (tmp & GFX9_RLCG_VFGATE_DISABLED) + pr_err("The vfgate is disabled, program reg:0x%05x failed!\n", offset); + else if (tmp & GFX9_RLCG_WRONG_OPERATION_TYPE) + pr_err("Wrong operation type, program reg:0x%05x failed!\n", offset); + else if (tmp & GFX9_RLCG_NOT_IN_RANGE) + pr_err("The register is not in range, program reg:0x%05x failed!\n", offset); + else + pr_err("Unknown error type, program reg:0x%05x failed!\n", offset); + } else + pr_err("timeout: rlcg program reg:0x%05x failed!\n", offset); + } + } + + ret = readl(scratch_reg0); + + return ret; +} + +static bool gfx_v9_0_get_rlcg_flag(struct amdgpu_device *adev, u32 acc_flags, u32 hwip, + int write, u32 *rlcg_flag) +{ + + switch (hwip) { + case GC_HWIP: + if (amdgpu_sriov_reg_indirect_gc(adev)) { + *rlcg_flag = write ? GFX9_RLCG_GC_WRITE : GFX9_RLCG_GC_READ; + + return true; + /* only in new version, AMDGPU_REGS_NO_KIQ and AMDGPU_REGS_RLC enabled simultaneously */ + } else if ((acc_flags & AMDGPU_REGS_RLC) && !(acc_flags & AMDGPU_REGS_NO_KIQ) && write) { + *rlcg_flag = GFX9_RLCG_GC_WRITE_OLD; + return true; + } + + break; + default: + return false; } + return false; +} + +static u32 gfx_v9_0_sriov_rreg(struct amdgpu_device *adev, u32 offset, u32 acc_flags, u32 hwip) +{ + u32 rlcg_flag; + + if (!amdgpu_sriov_runtime(adev) && gfx_v9_0_get_rlcg_flag(adev, acc_flags, hwip, 0, &rlcg_flag)) + return gfx_v9_0_rlcg_rw(adev, offset, 0, rlcg_flag); + + if (acc_flags & AMDGPU_REGS_NO_KIQ) + return RREG32_NO_KIQ(offset); + else + return RREG32(offset); } static void gfx_v9_0_sriov_wreg(struct amdgpu_device *adev, u32 offset, - u32 v, u32 acc_flags, u32 hwip) + u32 value, u32 acc_flags, u32 hwip) { - if ((acc_flags & AMDGPU_REGS_RLC) && - amdgpu_sriov_fullaccess(adev)) { - gfx_v9_0_rlcg_w(adev, offset, v, acc_flags); + u32 rlcg_flag; + if (!amdgpu_sriov_runtime(adev) && gfx_v9_0_get_rlcg_flag(adev, acc_flags, hwip, 1, &rlcg_flag)) { + gfx_v9_0_rlcg_rw(adev, offset, value, rlcg_flag); return; } if (acc_flags & AMDGPU_REGS_NO_KIQ) - WREG32_NO_KIQ(offset, v); + WREG32_NO_KIQ(offset, value); else - WREG32(offset, v); + WREG32(offset, value); } #define VEGA10_GB_ADDR_CONFIG_GOLDEN 0x2a114042 @@ -953,8 +1018,8 @@ static void gfx_v9_0_set_kiq_pm4_funcs(struct amdgpu_device *adev) static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): soc15_program_register_sequence(adev, golden_settings_gc_9_0, ARRAY_SIZE(golden_settings_gc_9_0)); @@ -962,7 +1027,7 @@ static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_gc_9_0_vg10, ARRAY_SIZE(golden_settings_gc_9_0_vg10)); break; - case CHIP_VEGA12: + case IP_VERSION(9, 2, 1): soc15_program_register_sequence(adev, golden_settings_gc_9_2_1, ARRAY_SIZE(golden_settings_gc_9_2_1)); @@ -970,7 +1035,7 @@ static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_gc_9_2_1_vg12, ARRAY_SIZE(golden_settings_gc_9_2_1_vg12)); break; - case CHIP_VEGA20: + case IP_VERSION(9, 4, 0): soc15_program_register_sequence(adev, golden_settings_gc_9_0, ARRAY_SIZE(golden_settings_gc_9_0)); @@ -978,12 +1043,13 @@ static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_gc_9_0_vg20, ARRAY_SIZE(golden_settings_gc_9_0_vg20)); break; - case CHIP_ARCTURUS: + case IP_VERSION(9, 4, 1): soc15_program_register_sequence(adev, golden_settings_gc_9_4_1_arct, ARRAY_SIZE(golden_settings_gc_9_4_1_arct)); break; - case CHIP_RAVEN: + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): soc15_program_register_sequence(adev, golden_settings_gc_9_1, ARRAY_SIZE(golden_settings_gc_9_1)); if (adev->apu_flags & AMD_APU_IS_RAVEN2) @@ -995,12 +1061,12 @@ static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_gc_9_1_rv1, ARRAY_SIZE(golden_settings_gc_9_1_rv1)); break; - case CHIP_RENOIR: + case IP_VERSION(9, 3, 0): soc15_program_register_sequence(adev, golden_settings_gc_9_1_rn, ARRAY_SIZE(golden_settings_gc_9_1_rn)); return; /* for renoir, don't need common goldensetting */ - case CHIP_ALDEBARAN: + case IP_VERSION(9, 4, 2): gfx_v9_4_2_init_golden_registers(adev, adev->smuio.funcs->get_die_id(adev)); break; @@ -1008,8 +1074,8 @@ static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev) break; } - if ((adev->asic_type != CHIP_ARCTURUS) && - (adev->asic_type != CHIP_ALDEBARAN)) + if ((adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 1)) && + (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 2))) soc15_program_register_sequence(adev, golden_settings_gc_9_x_common, (const u32)ARRAY_SIZE(golden_settings_gc_9_x_common)); } @@ -1193,15 +1259,15 @@ static void gfx_v9_0_check_fw_write_wait(struct amdgpu_device *adev) adev->gfx.me_fw_write_wait = false; adev->gfx.mec_fw_write_wait = false; - if ((adev->asic_type != CHIP_ARCTURUS) && + if ((adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 1)) && ((adev->gfx.mec_fw_version < 0x000001a5) || (adev->gfx.mec_feature_version < 46) || (adev->gfx.pfp_fw_version < 0x000000b7) || (adev->gfx.pfp_feature_version < 46))) DRM_WARN_ONCE("CP firmware version too old, please update!"); - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): if ((adev->gfx.me_fw_version >= 0x0000009c) && (adev->gfx.me_feature_version >= 42) && (adev->gfx.pfp_fw_version >= 0x000000b1) && @@ -1212,7 +1278,7 @@ static void gfx_v9_0_check_fw_write_wait(struct amdgpu_device *adev) (adev->gfx.mec_feature_version >= 42)) adev->gfx.mec_fw_write_wait = true; break; - case CHIP_VEGA12: + case IP_VERSION(9, 2, 1): if ((adev->gfx.me_fw_version >= 0x0000009c) && (adev->gfx.me_feature_version >= 44) && (adev->gfx.pfp_fw_version >= 0x000000b2) && @@ -1223,7 +1289,7 @@ static void gfx_v9_0_check_fw_write_wait(struct amdgpu_device *adev) (adev->gfx.mec_feature_version >= 44)) adev->gfx.mec_fw_write_wait = true; break; - case CHIP_VEGA20: + case IP_VERSION(9, 4, 0): if ((adev->gfx.me_fw_version >= 0x0000009c) && (adev->gfx.me_feature_version >= 44) && (adev->gfx.pfp_fw_version >= 0x000000b2) && @@ -1234,7 +1300,8 @@ static void gfx_v9_0_check_fw_write_wait(struct amdgpu_device *adev) (adev->gfx.mec_feature_version >= 44)) adev->gfx.mec_fw_write_wait = true; break; - case CHIP_RAVEN: + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 2): if ((adev->gfx.me_fw_version >= 0x0000009c) && (adev->gfx.me_feature_version >= 42) && (adev->gfx.pfp_fw_version >= 0x000000b1) && @@ -1297,7 +1364,7 @@ static bool is_raven_kicker(struct amdgpu_device *adev) static bool check_if_enlarge_doorbell_range(struct amdgpu_device *adev) { - if ((adev->asic_type == CHIP_RENOIR) && + if ((adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 3, 0)) && (adev->gfx.me_fw_version >= 0x000000a5) && (adev->gfx.me_feature_version >= 52)) return true; @@ -1310,12 +1377,13 @@ static void gfx_v9_0_check_if_need_gfxoff(struct amdgpu_device *adev) if (gfx_v9_0_should_disable_gfxoff(adev->pdev)) adev->pm.pp_feature &= ~PP_GFXOFF_MASK; - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 4, 0): break; - case CHIP_RAVEN: + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): if (!((adev->apu_flags & AMD_APU_IS_RAVEN2) || (adev->apu_flags & AMD_APU_IS_PICASSO)) && ((!is_raven_kicker(adev) && @@ -1329,7 +1397,7 @@ static void gfx_v9_0_check_if_need_gfxoff(struct amdgpu_device *adev) AMD_PG_SUPPORT_CP | AMD_PG_SUPPORT_RLC_SMU_HS; break; - case CHIP_RENOIR: + case IP_VERSION(9, 3, 0): if (adev->pm.pp_feature & PP_GFXOFF_MASK) adev->pg_flags |= AMD_PG_SUPPORT_GFX_PG | AMD_PG_SUPPORT_CP | @@ -1553,9 +1621,9 @@ out: static bool gfx_v9_0_load_mec2_fw_bin_support(struct amdgpu_device *adev) { - if (adev->asic_type == CHIP_ALDEBARAN || - adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_RENOIR) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 3, 0)) return false; return true; @@ -1663,17 +1731,18 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev) DRM_DEBUG("\n"); - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): chip_name = "vega10"; break; - case CHIP_VEGA12: + case IP_VERSION(9, 2, 1): chip_name = "vega12"; break; - case CHIP_VEGA20: + case IP_VERSION(9, 4, 0): chip_name = "vega20"; break; - case CHIP_RAVEN: + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): if (adev->apu_flags & AMD_APU_IS_RAVEN2) chip_name = "raven2"; else if (adev->apu_flags & AMD_APU_IS_PICASSO) @@ -1681,16 +1750,16 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev) else chip_name = "raven"; break; - case CHIP_ARCTURUS: + case IP_VERSION(9, 4, 1): chip_name = "arcturus"; break; - case CHIP_RENOIR: + case IP_VERSION(9, 3, 0): if (adev->apu_flags & AMD_APU_IS_RENOIR) chip_name = "renoir"; else chip_name = "green_sardine"; break; - case CHIP_ALDEBARAN: + case IP_VERSION(9, 4, 2): chip_name = "aldebaran"; break; default: @@ -1794,7 +1863,7 @@ static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev) if (adev->flags & AMD_IS_APU) always_on_cu_num = 4; - else if (adev->asic_type == CHIP_VEGA12) + else if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 2, 1)) always_on_cu_num = 8; else always_on_cu_num = 12; @@ -1963,11 +2032,12 @@ static int gfx_v9_0_rlc_init(struct amdgpu_device *adev) return r; } - switch (adev->asic_type) { - case CHIP_RAVEN: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): gfx_v9_0_init_lbpw(adev); break; - case CHIP_VEGA20: + case IP_VERSION(9, 4, 0): gfx_v9_4_init_lbpw(adev); break; default: @@ -2142,8 +2212,8 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev) adev->gfx.funcs = &gfx_v9_0_gfx_funcs; - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): 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; @@ -2151,7 +2221,7 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev) adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0; gb_addr_config = VEGA10_GB_ADDR_CONFIG_GOLDEN; break; - case CHIP_VEGA12: + case IP_VERSION(9, 2, 1): 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; @@ -2160,7 +2230,7 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev) gb_addr_config = VEGA12_GB_ADDR_CONFIG_GOLDEN; DRM_INFO("fix gfx.config for vega12\n"); break; - case CHIP_VEGA20: + case IP_VERSION(9, 4, 0): adev->gfx.ras_funcs = &gfx_v9_0_ras_funcs; adev->gfx.config.max_hw_contexts = 8; adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; @@ -2175,7 +2245,8 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev) if (err) return err; break; - case CHIP_RAVEN: + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 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; @@ -2186,7 +2257,7 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev) else gb_addr_config = RAVEN_GB_ADDR_CONFIG_GOLDEN; break; - case CHIP_ARCTURUS: + case IP_VERSION(9, 4, 1): adev->gfx.ras_funcs = &gfx_v9_4_ras_funcs; adev->gfx.config.max_hw_contexts = 8; adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; @@ -2197,7 +2268,7 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev) gb_addr_config &= ~0xf3e777ff; gb_addr_config |= 0x22014042; break; - case CHIP_RENOIR: + case IP_VERSION(9, 3, 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; @@ -2207,7 +2278,7 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev) gb_addr_config &= ~0xf3e777ff; gb_addr_config |= 0x22010042; break; - case CHIP_ALDEBARAN: + case IP_VERSION(9, 4, 2): adev->gfx.ras_funcs = &gfx_v9_4_2_ras_funcs; adev->gfx.config.max_hw_contexts = 8; adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; @@ -2305,14 +2376,15 @@ static int gfx_v9_0_sw_init(void *handle) struct amdgpu_kiq *kiq; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: - case CHIP_RAVEN: - case CHIP_ARCTURUS: - case CHIP_RENOIR: - case CHIP_ALDEBARAN: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 4, 0): + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 4, 1): + case IP_VERSION(9, 3, 0): + case IP_VERSION(9, 4, 2): adev->gfx.mec.num_mec = 2; break; default: @@ -2455,7 +2527,9 @@ static int gfx_v9_0_sw_fini(void *handle) amdgpu_gfx_kiq_fini(adev); gfx_v9_0_mec_fini(adev); - amdgpu_bo_unref(&adev->gfx.rlc.clear_state_obj); + amdgpu_bo_free_kernel(&adev->gfx.rlc.clear_state_obj, + &adev->gfx.rlc.clear_state_gpu_addr, + (void **)&adev->gfx.rlc.cs_ptr); if (adev->flags & AMD_IS_APU) { amdgpu_bo_free_kernel(&adev->gfx.rlc.cp_table_obj, &adev->gfx.rlc.cp_table_gpu_addr, @@ -2596,8 +2670,8 @@ static void gfx_v9_0_init_sq_config(struct amdgpu_device *adev) { uint32_t tmp; - switch (adev->asic_type) { - case CHIP_ARCTURUS: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 4, 1): tmp = RREG32_SOC15(GC, 0, mmSQ_CONFIG); tmp = REG_SET_FIELD(tmp, SQ_CONFIG, DISABLE_BARRIER_WAITCNT, 1); @@ -2932,7 +3006,7 @@ static void gfx_v9_0_init_gfx_power_gating(struct amdgpu_device *adev) /* program GRBM_REG_SAVE_GFX_IDLE_THRESHOLD to 0x55f0 */ data |= (0x55f0 << RLC_AUTO_PG_CTRL__GRBM_REG_SAVE_GFX_IDLE_THRESHOLD__SHIFT); WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_AUTO_PG_CTRL), data); - if (adev->asic_type != CHIP_RENOIR) + if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 3, 0)) pwr_10_0_gfxip_control_over_cgpg(adev, true); } } @@ -3044,7 +3118,7 @@ static void gfx_v9_0_init_pg(struct amdgpu_device *adev) * And it's needed by gfxoff feature. */ if (adev->gfx.rlc.is_rlc_v2_1) { - if (adev->asic_type == CHIP_VEGA12 || + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 2, 1) || (adev->apu_flags & AMD_APU_IS_RAVEN2)) gfx_v9_1_init_rlc_save_restore_list(adev); gfx_v9_0_enable_save_restore_machine(adev); @@ -3056,8 +3130,8 @@ static void gfx_v9_0_init_pg(struct amdgpu_device *adev) AMD_PG_SUPPORT_CP | AMD_PG_SUPPORT_GDS | AMD_PG_SUPPORT_RLC_SMU_HS)) { - WREG32(mmRLC_JUMP_TABLE_RESTORE, - adev->gfx.rlc.cp_table_gpu_addr >> 8); + WREG32_SOC15(GC, 0, mmRLC_JUMP_TABLE_RESTORE, + adev->gfx.rlc.cp_table_gpu_addr >> 8); gfx_v9_0_init_gfx_power_gating(adev); } } @@ -3157,14 +3231,15 @@ static int gfx_v9_0_rlc_resume(struct amdgpu_device *adev) return r; } - switch (adev->asic_type) { - case CHIP_RAVEN: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): if (amdgpu_lbpw == 0) gfx_v9_0_enable_lbpw(adev, false); else gfx_v9_0_enable_lbpw(adev, true); break; - case CHIP_VEGA20: + case IP_VERSION(9, 4, 0): if (amdgpu_lbpw > 0) gfx_v9_0_enable_lbpw(adev, true); else @@ -3959,8 +4034,8 @@ static void gfx_v9_0_init_tcp_config(struct amdgpu_device *adev) { u32 tmp; - if (adev->asic_type != CHIP_ARCTURUS && - adev->asic_type != CHIP_ALDEBARAN) + if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 1) && + adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 2)) return; tmp = RREG32_SOC15(GC, 0, mmTCP_ADDR_CONFIG); @@ -4000,7 +4075,7 @@ static int gfx_v9_0_hw_init(void *handle) if (r) return r; - if (adev->asic_type == CHIP_ALDEBARAN) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) gfx_v9_4_2_set_power_brake_sequence(adev); return r; @@ -4045,9 +4120,10 @@ static int gfx_v9_0_hw_fini(void *handle) gfx_v9_0_cp_enable(adev, false); - /* Skip suspend with A+A reset */ - if (adev->gmc.xgmi.connected_to_cpu && amdgpu_in_reset(adev)) { - dev_dbg(adev->dev, "Device in reset. Skipping RLC halt\n"); + /* Skip stopping RLC with A+A reset or when RLC controls GFX clock */ + if ((adev->gmc.xgmi.connected_to_cpu && amdgpu_in_reset(adev)) || + (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(9, 4, 2))) { + dev_dbg(adev->dev, "Skipping RLC halt\n"); return 0; } @@ -4228,19 +4304,38 @@ failed_kiq_read: static uint64_t gfx_v9_0_get_gpu_clock_counter(struct amdgpu_device *adev) { - uint64_t clock; + uint64_t clock, clock_lo, clock_hi, hi_check; - amdgpu_gfx_off_ctrl(adev, false); - mutex_lock(&adev->gfx.gpu_clock_mutex); - if (adev->asic_type == CHIP_VEGA10 && amdgpu_sriov_runtime(adev)) { - clock = gfx_v9_0_kiq_read_clock(adev); - } else { - WREG32_SOC15(GC, 0, mmRLC_CAPTURE_GPU_CLOCK_COUNT, 1); - clock = (uint64_t)RREG32_SOC15(GC, 0, mmRLC_GPU_CLOCK_COUNT_LSB) | - ((uint64_t)RREG32_SOC15(GC, 0, mmRLC_GPU_CLOCK_COUNT_MSB) << 32ULL); + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 3, 0): + preempt_disable(); + clock_hi = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Renoir); + clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Renoir); + hi_check = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Renoir); + /* The SMUIO TSC clock frequency is 100MHz, which sets 32-bit carry over + * roughly every 42 seconds. + */ + if (hi_check != clock_hi) { + clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Renoir); + clock_hi = hi_check; + } + preempt_enable(); + clock = clock_lo | (clock_hi << 32ULL); + break; + default: + amdgpu_gfx_off_ctrl(adev, false); + mutex_lock(&adev->gfx.gpu_clock_mutex); + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 0, 1) && amdgpu_sriov_runtime(adev)) { + clock = gfx_v9_0_kiq_read_clock(adev); + } else { + WREG32_SOC15(GC, 0, mmRLC_CAPTURE_GPU_CLOCK_COUNT, 1); + clock = (uint64_t)RREG32_SOC15(GC, 0, mmRLC_GPU_CLOCK_COUNT_LSB) | + ((uint64_t)RREG32_SOC15(GC, 0, mmRLC_GPU_CLOCK_COUNT_MSB) << 32ULL); + } + mutex_unlock(&adev->gfx.gpu_clock_mutex); + amdgpu_gfx_off_ctrl(adev, true); + break; } - mutex_unlock(&adev->gfx.gpu_clock_mutex); - amdgpu_gfx_off_ctrl(adev, true); return clock; } @@ -4582,7 +4677,7 @@ static int gfx_v9_0_do_edc_gpr_workarounds(struct amdgpu_device *adev) if (!ring->sched.ready) return 0; - if (adev->asic_type == CHIP_ARCTURUS) { + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1)) { vgpr_init_shader_ptr = vgpr_init_compute_shader_arcturus; vgpr_init_shader_size = sizeof(vgpr_init_compute_shader_arcturus); vgpr_init_regs_ptr = vgpr_init_regs_arcturus; @@ -4732,8 +4827,8 @@ static int gfx_v9_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - if (adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_ALDEBARAN) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) adev->gfx.num_gfx_rings = 0; else adev->gfx.num_gfx_rings = GFX9_NUM_GFX_RINGS; @@ -4767,7 +4862,7 @@ static int gfx_v9_0_ecc_late_init(void *handle) } /* requires IBs so do in late init after IB pool is initialized */ - if (adev->asic_type == CHIP_ALDEBARAN) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) r = gfx_v9_4_2_do_edc_gpr_workarounds(adev); else r = gfx_v9_0_do_edc_gpr_workarounds(adev); @@ -4895,7 +4990,7 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev /* 1 - RLC_CGTT_MGCG_OVERRIDE */ def = data = RREG32_SOC15(GC, 0, mmRLC_CGTT_MGCG_OVERRIDE); - if (adev->asic_type != CHIP_VEGA12) + if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 2, 1)) data &= ~RLC_CGTT_MGCG_OVERRIDE__CPF_CGTT_SCLK_OVERRIDE_MASK; data &= ~(RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK | @@ -4929,7 +5024,7 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev /* 1 - MGCG_OVERRIDE */ def = data = RREG32_SOC15(GC, 0, mmRLC_CGTT_MGCG_OVERRIDE); - if (adev->asic_type != CHIP_VEGA12) + if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 2, 1)) data |= RLC_CGTT_MGCG_OVERRIDE__CPF_CGTT_SCLK_OVERRIDE_MASK; data |= (RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK | @@ -5035,7 +5130,7 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev /* enable cgcg FSM(0x0000363F) */ def = RREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL); - if (adev->asic_type == CHIP_ARCTURUS) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1)) data = (0x2000 << RLC_CGCG_CGLS_CTRL__CGCG_GFX_IDLE_THRESHOLD__SHIFT) | RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK; else @@ -5094,11 +5189,13 @@ static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid) { u32 reg, data; + amdgpu_gfx_off_ctrl(adev, false); + reg = SOC15_REG_OFFSET(GC, 0, mmRLC_SPM_MC_CNTL); if (amdgpu_sriov_is_pp_one_vf(adev)) data = RREG32_NO_KIQ(reg); else - data = RREG32(reg); + data = RREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL); data &= ~RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK; data |= (vmid & RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK) << RLC_SPM_MC_CNTL__RLC_SPM_VMID__SHIFT; @@ -5107,6 +5204,8 @@ static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid) WREG32_SOC15_NO_KIQ(GC, 0, mmRLC_SPM_MC_CNTL, data); else WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data); + + amdgpu_gfx_off_ctrl(adev, true); } static bool gfx_v9_0_check_rlcg_range(struct amdgpu_device *adev, @@ -5152,6 +5251,7 @@ static const struct amdgpu_rlc_funcs gfx_v9_0_rlc_funcs = { .start = gfx_v9_0_rlc_start, .update_spm_vmid = gfx_v9_0_update_spm_vmid, .sriov_wreg = gfx_v9_0_sriov_wreg, + .sriov_rreg = gfx_v9_0_sriov_rreg, .is_rlcg_access_range = gfx_v9_0_is_rlcg_access_range, }; @@ -5161,9 +5261,10 @@ static int gfx_v9_0_set_powergating_state(void *handle, struct amdgpu_device *adev = (struct amdgpu_device *)handle; bool enable = (state == AMD_PG_STATE_GATE); - switch (adev->asic_type) { - case CHIP_RAVEN: - case CHIP_RENOIR: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 3, 0): if (!enable) amdgpu_gfx_off_ctrl(adev, false); @@ -5189,7 +5290,7 @@ static int gfx_v9_0_set_powergating_state(void *handle, if (enable) amdgpu_gfx_off_ctrl(adev, true); break; - case CHIP_VEGA12: + case IP_VERSION(9, 2, 1): amdgpu_gfx_off_ctrl(adev, enable); break; default: @@ -5207,14 +5308,15 @@ static int gfx_v9_0_set_clockgating_state(void *handle, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: - case CHIP_RAVEN: - case CHIP_ARCTURUS: - case CHIP_RENOIR: - case CHIP_ALDEBARAN: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 4, 0): + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 4, 1): + case IP_VERSION(9, 3, 0): + case IP_VERSION(9, 4, 2): gfx_v9_0_update_gfx_clock_gating(adev, state == AMD_CG_STATE_GATE); break; @@ -5256,7 +5358,7 @@ static void gfx_v9_0_get_clockgating_state(void *handle, u32 *flags) if (data & CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK) *flags |= AMD_CG_SUPPORT_GFX_CP_LS | AMD_CG_SUPPORT_GFX_MGLS; - if (adev->asic_type != CHIP_ARCTURUS) { + if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 1)) { /* AMD_CG_SUPPORT_GFX_3D_CGCG */ data = RREG32_KIQ(SOC15_REG_OFFSET(GC, 0, mmRLC_CGCG_CGLS_CTRL_3D)); if (data & RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK) @@ -5755,16 +5857,16 @@ static void gfx_v9_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev, switch (state) { case AMDGPU_IRQ_STATE_DISABLE: - mec_int_cntl = RREG32(mec_int_cntl_reg); + mec_int_cntl = RREG32_SOC15_IP(GC,mec_int_cntl_reg); mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL, TIME_STAMP_INT_ENABLE, 0); - WREG32(mec_int_cntl_reg, mec_int_cntl); + WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl); break; case AMDGPU_IRQ_STATE_ENABLE: - mec_int_cntl = RREG32(mec_int_cntl_reg); + mec_int_cntl = RREG32_SOC15_IP(GC, mec_int_cntl_reg); mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL, TIME_STAMP_INT_ENABLE, 1); - WREG32(mec_int_cntl_reg, mec_int_cntl); + WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl); break; default: break; @@ -7027,14 +7129,15 @@ static void gfx_v9_0_set_irq_funcs(struct amdgpu_device *adev) static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: - case CHIP_RAVEN: - case CHIP_ARCTURUS: - case CHIP_RENOIR: - case CHIP_ALDEBARAN: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 4, 0): + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 4, 1): + case IP_VERSION(9, 3, 0): + case IP_VERSION(9, 4, 2): adev->gfx.rlc.funcs = &gfx_v9_0_rlc_funcs; break; default: @@ -7045,17 +7148,18 @@ static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev) static void gfx_v9_0_set_gds_init(struct amdgpu_device *adev) { /* init asci gds info */ - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 4, 0): adev->gds.gds_size = 0x10000; break; - case CHIP_RAVEN: - case CHIP_ARCTURUS: + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 4, 1): adev->gds.gds_size = 0x1000; break; - case CHIP_ALDEBARAN: + case IP_VERSION(9, 4, 2): /* aldebaran removed all the GDS internal memory, * only support GWS opcode in kernel, like barrier * semaphore.etc */ @@ -7066,24 +7170,25 @@ static void gfx_v9_0_set_gds_init(struct amdgpu_device *adev) break; } - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA20: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 4, 0): adev->gds.gds_compute_max_wave_id = 0x7ff; break; - case CHIP_VEGA12: + case IP_VERSION(9, 2, 1): adev->gds.gds_compute_max_wave_id = 0x27f; break; - case CHIP_RAVEN: + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 1, 0): if (adev->apu_flags & AMD_APU_IS_RAVEN2) adev->gds.gds_compute_max_wave_id = 0x77; /* raven2 */ else adev->gds.gds_compute_max_wave_id = 0x15f; /* raven1 */ break; - case CHIP_ARCTURUS: + case IP_VERSION(9, 4, 1): adev->gds.gds_compute_max_wave_id = 0xfff; break; - case CHIP_ALDEBARAN: + case IP_VERSION(9, 4, 2): /* deprecated for Aldebaran, no usage at all */ adev->gds.gds_compute_max_wave_id = 0; break; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c index 00a2b36a24b3..c4f37a161875 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c @@ -706,6 +706,11 @@ int gfx_v9_4_2_do_edc_gpr_workarounds(struct amdgpu_device *adev) if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX)) return 0; + /* Workaround for ALDEBARAN, skip GPRs init in GPU reset. + Will remove it once GPRs init algorithm works for all CU settings. */ + if (amdgpu_in_reset(adev)) + return 0; + gfx_v9_4_2_do_sgprs_init(adev); gfx_v9_4_2_do_vgprs_init(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c index bda1542ef1dd..ec4d5e15b766 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c @@ -162,7 +162,6 @@ static void gfxhub_v1_0_init_tlb_regs(struct amdgpu_device *adev) ENABLE_ADVANCED_DRIVER_MODEL, 1); tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0); tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, MTYPE, MTYPE_UC);/* XXX for emulation. */ tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); @@ -348,6 +347,10 @@ static void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev) WREG32_SOC15_OFFSET(GC, 0, mmVM_CONTEXT0_CNTL, i * hub->ctx_distance, 0); + if (amdgpu_sriov_vf(adev)) + /* Avoid write to GMC registers */ + return; + /* Setup TLB control */ tmp = RREG32_SOC15(GC, 0, mmMC_VM_MX_L1_TLB_CNTL); tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_1.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_1.c index 497b86c376c6..90f0aefbdb39 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_1.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_1.c @@ -54,15 +54,17 @@ int gfxhub_v1_1_get_xgmi_info(struct amdgpu_device *adev) seg_size = REG_GET_FIELD( RREG32_SOC15(GC, 0, mmMC_VM_XGMI_LFB_SIZE_ALDE), MC_VM_XGMI_LFB_SIZE, PF_LFB_SIZE) << 24; + max_region = + REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL_ALDE, PF_MAX_REGION); } else { xgmi_lfb_cntl = RREG32_SOC15(GC, 0, mmMC_VM_XGMI_LFB_CNTL); seg_size = REG_GET_FIELD( RREG32_SOC15(GC, 0, mmMC_VM_XGMI_LFB_SIZE), MC_VM_XGMI_LFB_SIZE, PF_LFB_SIZE) << 24; + max_region = + REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL, PF_MAX_REGION); } - max_region = - REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL, PF_MAX_REGION); switch (adev->asic_type) { @@ -89,9 +91,15 @@ int gfxhub_v1_1_get_xgmi_info(struct amdgpu_device *adev) if (adev->gmc.xgmi.num_physical_nodes > max_num_physical_nodes) return -EINVAL; - adev->gmc.xgmi.physical_node_id = - REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL, - PF_LFB_REGION); + if (adev->asic_type == CHIP_ALDEBARAN) { + adev->gmc.xgmi.physical_node_id = + REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL_ALDE, + PF_LFB_REGION); + } else { + adev->gmc.xgmi.physical_node_id = + REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL, + PF_LFB_REGION); + } if (adev->gmc.xgmi.physical_node_id > max_physical_node_id) return -EINVAL; diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c index 14c1c1a297dd..6e0ace2fbfab 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c @@ -196,7 +196,6 @@ static void gfxhub_v2_0_init_tlb_regs(struct amdgpu_device *adev) ENABLE_ADVANCED_DRIVER_MODEL, 1); tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); - tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0); tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, MTYPE, MTYPE_UC); /* UC, uncached */ diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c index 1a374ec0514a..b4eddf6e98a6 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c @@ -197,7 +197,6 @@ static void gfxhub_v2_1_init_tlb_regs(struct amdgpu_device *adev) ENABLE_ADVANCED_DRIVER_MODEL, 1); tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); - tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0); tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, MTYPE, MTYPE_UC); /* UC, uncached */ @@ -506,8 +505,8 @@ static int gfxhub_v2_1_get_xgmi_info(struct amdgpu_device *adev) u32 max_num_physical_nodes = 0; u32 max_physical_node_id = 0; - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: + switch (adev->ip_versions[XGMI_HWIP][0]) { + case IP_VERSION(4, 8, 0): max_num_physical_nodes = 4; max_physical_node_id = 3; break; @@ -544,7 +543,7 @@ static void gfxhub_v2_1_utcl2_harvest(struct amdgpu_device *adev) adev->gfx.config.max_sh_per_se * adev->gfx.config.max_shader_engines); - if (adev->asic_type == CHIP_YELLOW_CARP) { + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 3)) { /* Get SA disabled bitmap from eFuse setting */ efuse_setting = RREG32_SOC15(GC, 0, mmCC_GC_SA_UNIT_DISABLE); efuse_setting &= CC_GC_SA_UNIT_DISABLE__SA_DISABLE_MASK; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c index e47104a1f559..3d5d47a799e3 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c @@ -107,7 +107,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev, /* Process it onyl if it's the first fault for this address */ if (entry->ih != &adev->irq.ih_soft && - amdgpu_gmc_filter_faults(adev, addr, entry->pasid, + amdgpu_gmc_filter_faults(adev, entry->ih, addr, entry->pasid, entry->timestamp)) return 1; @@ -133,7 +133,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev, * the new fast GRBM interface. */ if ((entry->vmid_src == AMDGPU_GFXHUB_0) && - (adev->asic_type < CHIP_SIENNA_CICHLID)) + (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 3, 0))) RREG32(hub->vm_l2_pro_fault_status); status = RREG32(hub->vm_l2_pro_fault_status); @@ -268,7 +268,7 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid, * to avoid a false ACK due to the new fast GRBM interface. */ if ((vmhub == AMDGPU_GFXHUB_0) && - (adev->asic_type < CHIP_SIENNA_CICHLID)) + (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 3, 0))) RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_req + hub->eng_distance * eng, hub_ip); @@ -657,8 +657,8 @@ static void gmc_v10_0_set_gmc_funcs(struct amdgpu_device *adev) static void gmc_v10_0_set_umc_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: + switch (adev->ip_versions[UMC_HWIP][0]) { + case IP_VERSION(8, 7, 0): adev->umc.max_ras_err_cnt_per_query = UMC_V8_7_TOTAL_CHANNEL_NUM; adev->umc.channel_inst_num = UMC_V8_7_CHANNEL_INSTANCE_NUM; adev->umc.umc_inst_num = UMC_V8_7_UMC_INSTANCE_NUM; @@ -674,9 +674,9 @@ static void gmc_v10_0_set_umc_funcs(struct amdgpu_device *adev) static void gmc_v10_0_set_mmhub_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_VANGOGH: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(2, 3, 0): + case IP_VERSION(2, 4, 0): adev->mmhub.funcs = &mmhub_v2_3_funcs; break; default: @@ -687,13 +687,13 @@ static void gmc_v10_0_set_mmhub_funcs(struct amdgpu_device *adev) static void gmc_v10_0_set_gfxhub_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): adev->gfxhub.funcs = &gfxhub_v2_1_funcs; break; default: @@ -800,23 +800,9 @@ static int gmc_v10_0_mc_init(struct amdgpu_device *adev) adev->gmc.visible_vram_size = adev->gmc.real_vram_size; /* set the gart size */ - if (amdgpu_gart_size == -1) { - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: - case CHIP_CYAN_SKILLFISH: - default: - adev->gmc.gart_size = 512ULL << 20; - break; - } - } else + if (amdgpu_gart_size == -1) + adev->gmc.gart_size = 512ULL << 20; + else adev->gmc.gart_size = (u64)amdgpu_gart_size << 20; gmc_v10_0_vram_gtt_location(adev, &adev->gmc); @@ -871,17 +857,17 @@ static int gmc_v10_0_sw_init(void *handle) adev->gmc.vram_vendor = vram_vendor; } - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: - case CHIP_CYAN_SKILLFISH: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 1): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 3): adev->num_vmhubs = 2; /* * To fulfill 4-level page support, @@ -928,12 +914,6 @@ static int gmc_v10_0_sw_init(void *handle) return r; } - if (adev->gmc.xgmi.supported) { - r = adev->gfxhub.funcs->get_xgmi_info(adev); - if (r) - return r; - } - r = gmc_v10_0_mc_init(adev); if (r) return r; @@ -989,21 +969,6 @@ static int gmc_v10_0_sw_fini(void *handle) static void gmc_v10_0_init_golden_registers(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: - case CHIP_CYAN_SKILLFISH: - break; - default: - break; - } } /** @@ -1021,10 +986,14 @@ static int gmc_v10_0_gart_enable(struct amdgpu_device *adev) return -EINVAL; } + if (amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev)) + goto skip_pin_bo; + r = amdgpu_gart_table_vram_pin(adev); if (r) return r; +skip_pin_bo: r = adev->gfxhub.funcs->gart_enable(adev); if (r) return r; @@ -1162,8 +1131,7 @@ static int gmc_v10_0_set_clockgating_state(void *handle, if (r) return r; - if (adev->asic_type >= CHIP_SIENNA_CICHLID && - adev->asic_type <= CHIP_YELLOW_CARP) + if (adev->ip_versions[ATHUB_HWIP][0] >= IP_VERSION(2, 1, 0)) return athub_v2_1_set_clockgating(adev, state); else return athub_v2_0_set_clockgating(adev, state); @@ -1175,8 +1143,7 @@ static void gmc_v10_0_get_clockgating_state(void *handle, u32 *flags) adev->mmhub.funcs->get_clockgating(adev, flags); - if (adev->asic_type >= CHIP_SIENNA_CICHLID && - adev->asic_type <= CHIP_YELLOW_CARP) + if (adev->ip_versions[ATHUB_HWIP][0] >= IP_VERSION(2, 1, 0)) athub_v2_1_get_clockgating(adev, flags); else athub_v2_0_get_clockgating(adev, flags); diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c index 0e81e03e9b49..0fe714f54cca 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c @@ -841,12 +841,12 @@ static int gmc_v6_0_sw_init(void *handle) adev->gmc.mc_mask = 0xffffffffffULL; - r = dma_set_mask_and_coherent(adev->dev, DMA_BIT_MASK(44)); + r = dma_set_mask_and_coherent(adev->dev, DMA_BIT_MASK(40)); if (r) { dev_warn(adev->dev, "No suitable DMA available.\n"); return r; } - adev->need_swiotlb = drm_need_swiotlb(44); + adev->need_swiotlb = drm_need_swiotlb(40); r = gmc_v6_0_init_microcode(adev); if (r) { diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 492ebed2915b..63b890f1e8af 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -515,10 +515,10 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev) static int gmc_v8_0_mc_init(struct amdgpu_device *adev) { int r; + u32 tmp; adev->gmc.vram_width = amdgpu_atombios_get_vram_width(adev); if (!adev->gmc.vram_width) { - u32 tmp; int chansize, numchan; /* Get VRAM informations */ @@ -562,8 +562,15 @@ static int gmc_v8_0_mc_init(struct amdgpu_device *adev) adev->gmc.vram_width = numchan * chansize; } /* size in MB on si */ - adev->gmc.mc_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL; - adev->gmc.real_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL; + tmp = RREG32(mmCONFIG_MEMSIZE); + /* some boards may have garbage in the upper 16 bits */ + if (tmp & 0xffff0000) { + DRM_INFO("Probable bad vram size: 0x%08x\n", tmp); + if (tmp & 0xffff) + tmp &= 0xffff; + } + adev->gmc.mc_vram_size = tmp * 1024ULL * 1024ULL; + adev->gmc.real_vram_size = adev->gmc.mc_vram_size; if (!(adev->flags & AMD_IS_APU)) { r = amdgpu_device_resize_fb_bar(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index 5551359d5dfd..57f2729a7bd0 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -478,9 +478,18 @@ static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev, hub = &adev->vmhub[j]; for (i = 0; i < 16; i++) { reg = hub->vm_context0_cntl + i; - tmp = RREG32(reg); + + if (j == AMDGPU_GFXHUB_0) + tmp = RREG32_SOC15_IP(GC, reg); + else + tmp = RREG32_SOC15_IP(MMHUB, reg); + tmp &= ~bits; - WREG32(reg, tmp); + + if (j == AMDGPU_GFXHUB_0) + WREG32_SOC15_IP(GC, reg, tmp); + else + WREG32_SOC15_IP(MMHUB, reg, tmp); } } break; @@ -489,9 +498,18 @@ static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev, hub = &adev->vmhub[j]; for (i = 0; i < 16; i++) { reg = hub->vm_context0_cntl + i; - tmp = RREG32(reg); + + if (j == AMDGPU_GFXHUB_0) + tmp = RREG32_SOC15_IP(GC, reg); + else + tmp = RREG32_SOC15_IP(MMHUB, reg); + tmp |= bits; - WREG32(reg, tmp); + + if (j == AMDGPU_GFXHUB_0) + WREG32_SOC15_IP(GC, reg, tmp); + else + WREG32_SOC15_IP(MMHUB, reg, tmp); } } break; @@ -523,7 +541,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, /* Process it onyl if it's the first fault for this address */ if (entry->ih != &adev->irq.ih_soft && - amdgpu_gmc_filter_faults(adev, addr, entry->pasid, + amdgpu_gmc_filter_faults(adev, entry->ih, addr, entry->pasid, entry->timestamp)) return 1; @@ -579,7 +597,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, * the new fast GRBM interface. */ if ((entry->vmid_src == AMDGPU_GFXHUB_0) && - (adev->asic_type < CHIP_ALDEBARAN)) + (adev->ip_versions[GC_HWIP][0] < IP_VERSION(9, 4, 2))) RREG32(hub->vm_l2_pro_fault_status); status = RREG32(hub->vm_l2_pro_fault_status); @@ -597,26 +615,28 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, gfxhub_client_ids[cid], cid); } else { - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(9, 0, 0): mmhub_cid = mmhub_client_ids_vega10[cid][rw]; break; - case CHIP_VEGA12: + case IP_VERSION(9, 3, 0): mmhub_cid = mmhub_client_ids_vega12[cid][rw]; break; - case CHIP_VEGA20: + case IP_VERSION(9, 4, 0): mmhub_cid = mmhub_client_ids_vega20[cid][rw]; break; - case CHIP_ARCTURUS: + case IP_VERSION(9, 4, 1): mmhub_cid = mmhub_client_ids_arcturus[cid][rw]; break; - case CHIP_RAVEN: + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 0): mmhub_cid = mmhub_client_ids_raven[cid][rw]; break; - case CHIP_RENOIR: + case IP_VERSION(1, 5, 0): + case IP_VERSION(2, 4, 0): mmhub_cid = mmhub_client_ids_renoir[cid][rw]; break; - case CHIP_ALDEBARAN: + case IP_VERSION(9, 4, 2): mmhub_cid = mmhub_client_ids_aldebaran[cid][rw]; break; default: @@ -694,7 +714,7 @@ static uint32_t gmc_v9_0_get_invalidate_req(unsigned int vmid, static bool gmc_v9_0_use_invalidate_semaphore(struct amdgpu_device *adev, uint32_t vmhub) { - if (adev->asic_type == CHIP_ALDEBARAN) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) return false; return ((vmhub == AMDGPU_MMHUB_0 || @@ -745,7 +765,7 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, hub = &adev->vmhub[vmhub]; if (adev->gmc.xgmi.num_physical_nodes && - adev->asic_type == CHIP_VEGA20) { + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 0)) { /* Vega20+XGMI caches PTEs in TC and TLB. Add a * heavy-weight TLB flush (type 2), which flushes * both. Due to a race condition with concurrent @@ -786,9 +806,12 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, /* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */ if (use_semaphore) { for (j = 0; j < adev->usec_timeout; j++) { - /* a read return value of 1 means semaphore acuqire */ - tmp = RREG32_NO_KIQ(hub->vm_inv_eng0_sem + - hub->eng_distance * eng); + /* a read return value of 1 means semaphore acquire */ + if (vmhub == AMDGPU_GFXHUB_0) + tmp = RREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_sem + hub->eng_distance * eng); + else + tmp = RREG32_SOC15_IP_NO_KIQ(MMHUB, hub->vm_inv_eng0_sem + hub->eng_distance * eng); + if (tmp & 0x1) break; udelay(1); @@ -799,8 +822,10 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, } do { - WREG32_NO_KIQ(hub->vm_inv_eng0_req + - hub->eng_distance * eng, inv_req); + if (vmhub == AMDGPU_GFXHUB_0) + WREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_req + hub->eng_distance * eng, inv_req); + else + WREG32_SOC15_IP_NO_KIQ(MMHUB, hub->vm_inv_eng0_req + hub->eng_distance * eng, inv_req); /* * Issue a dummy read to wait for the ACK register to @@ -808,13 +833,16 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, * GRBM interface. */ if ((vmhub == AMDGPU_GFXHUB_0) && - (adev->asic_type < CHIP_ALDEBARAN)) + (adev->ip_versions[GC_HWIP][0] < IP_VERSION(9, 4, 2))) RREG32_NO_KIQ(hub->vm_inv_eng0_req + hub->eng_distance * eng); for (j = 0; j < adev->usec_timeout; j++) { - tmp = RREG32_NO_KIQ(hub->vm_inv_eng0_ack + - hub->eng_distance * eng); + if (vmhub == AMDGPU_GFXHUB_0) + tmp = RREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_ack + hub->eng_distance * eng); + else + tmp = RREG32_SOC15_IP_NO_KIQ(MMHUB, hub->vm_inv_eng0_ack + hub->eng_distance * eng); + if (tmp & (1 << vmid)) break; udelay(1); @@ -825,13 +853,16 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, } while (inv_req); /* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */ - if (use_semaphore) + if (use_semaphore) { /* * add semaphore release after invalidation, * write with 0 means semaphore release */ - WREG32_NO_KIQ(hub->vm_inv_eng0_sem + - hub->eng_distance * eng, 0); + if (vmhub == AMDGPU_GFXHUB_0) + WREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_sem + hub->eng_distance * eng, 0); + else + WREG32_SOC15_IP_NO_KIQ(MMHUB, hub->vm_inv_eng0_sem + hub->eng_distance * eng, 0); + } spin_unlock(&adev->gmc.invalidate_lock); @@ -874,7 +905,7 @@ static int gmc_v9_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, * still need a second TLB flush after this. */ bool vega20_xgmi_wa = (adev->gmc.xgmi.num_physical_nodes && - adev->asic_type == CHIP_VEGA20); + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 0)); /* 2 dwords flush + 8 dwords fence */ unsigned int ndw = kiq->pmf->invalidate_tlbs_size + 8; @@ -1088,13 +1119,13 @@ static void gmc_v9_0_get_vm_pte(struct amdgpu_device *adev, *flags &= ~AMDGPU_PTE_VALID; } - if ((adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_ALDEBARAN) && + if ((adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) && !(*flags & AMDGPU_PTE_SYSTEM) && mapping->bo_va->is_xgmi) *flags |= AMDGPU_PTE_SNOOPED; - if (adev->asic_type == CHIP_ALDEBARAN) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) *flags |= mapping->flags & AMDGPU_PTE_SNOOPED; } @@ -1108,9 +1139,10 @@ static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev) } else { u32 viewport; - switch (adev->asic_type) { - case CHIP_RAVEN: - case CHIP_RENOIR: + switch (adev->ip_versions[DCE_HWIP][0]) { + case IP_VERSION(1, 0, 0): + case IP_VERSION(1, 0, 1): + case IP_VERSION(2, 1, 0): viewport = RREG32_SOC15(DCE, 0, mmHUBP0_DCSURF_PRI_VIEWPORT_DIMENSION); size = (REG_GET_FIELD(viewport, HUBP0_DCSURF_PRI_VIEWPORT_DIMENSION, PRI_VIEWPORT_HEIGHT) * @@ -1118,9 +1150,6 @@ static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev) HUBP0_DCSURF_PRI_VIEWPORT_DIMENSION, PRI_VIEWPORT_WIDTH) * 4); break; - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: default: viewport = RREG32_SOC15(DCE, 0, mmSCL0_VIEWPORT_SIZE); size = (REG_GET_FIELD(viewport, SCL0_VIEWPORT_SIZE, VIEWPORT_HEIGHT) * @@ -1151,11 +1180,11 @@ static void gmc_v9_0_set_gmc_funcs(struct amdgpu_device *adev) static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[UMC_HWIP][0]) { + case IP_VERSION(6, 0, 0): adev->umc.funcs = &umc_v6_0_funcs; break; - case CHIP_VEGA20: + case IP_VERSION(6, 1, 1): adev->umc.max_ras_err_cnt_per_query = UMC_V6_1_TOTAL_CHANNEL_NUM; adev->umc.channel_inst_num = UMC_V6_1_CHANNEL_INSTANCE_NUM; adev->umc.umc_inst_num = UMC_V6_1_UMC_INSTANCE_NUM; @@ -1163,7 +1192,7 @@ static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev) adev->umc.channel_idx_tbl = &umc_v6_1_channel_idx_tbl[0][0]; adev->umc.ras_funcs = &umc_v6_1_ras_funcs; break; - case CHIP_ARCTURUS: + case IP_VERSION(6, 1, 2): adev->umc.max_ras_err_cnt_per_query = UMC_V6_1_TOTAL_CHANNEL_NUM; adev->umc.channel_inst_num = UMC_V6_1_CHANNEL_INSTANCE_NUM; adev->umc.umc_inst_num = UMC_V6_1_UMC_INSTANCE_NUM; @@ -1171,7 +1200,7 @@ static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev) adev->umc.channel_idx_tbl = &umc_v6_1_channel_idx_tbl[0][0]; adev->umc.ras_funcs = &umc_v6_1_ras_funcs; break; - case CHIP_ALDEBARAN: + case IP_VERSION(6, 7, 0): adev->umc.max_ras_err_cnt_per_query = UMC_V6_7_TOTAL_CHANNEL_NUM; adev->umc.channel_inst_num = UMC_V6_7_CHANNEL_INSTANCE_NUM; adev->umc.umc_inst_num = UMC_V6_7_UMC_INSTANCE_NUM; @@ -1190,11 +1219,11 @@ static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev) static void gmc_v9_0_set_mmhub_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_ARCTURUS: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(9, 4, 1): adev->mmhub.funcs = &mmhub_v9_4_funcs; break; - case CHIP_ALDEBARAN: + case IP_VERSION(9, 4, 2): adev->mmhub.funcs = &mmhub_v1_7_funcs; break; default: @@ -1205,14 +1234,14 @@ static void gmc_v9_0_set_mmhub_funcs(struct amdgpu_device *adev) static void gmc_v9_0_set_mmhub_ras_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_VEGA20: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(9, 4, 0): adev->mmhub.ras_funcs = &mmhub_v1_0_ras_funcs; break; - case CHIP_ARCTURUS: + case IP_VERSION(9, 4, 1): adev->mmhub.ras_funcs = &mmhub_v9_4_ras_funcs; break; - case CHIP_ALDEBARAN: + case IP_VERSION(9, 4, 2): adev->mmhub.ras_funcs = &mmhub_v1_7_ras_funcs; break; default: @@ -1233,8 +1262,9 @@ static void gmc_v9_0_set_hdp_ras_funcs(struct amdgpu_device *adev) static void gmc_v9_0_set_mca_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_ALDEBARAN: + /* is UMC the right IP to check for MCA? Maybe DF? */ + switch (adev->ip_versions[UMC_HWIP][0]) { + case IP_VERSION(6, 7, 0): if (!adev->gmc.xgmi.connected_to_cpu) adev->mca.funcs = &mca_v3_0_funcs; break; @@ -1247,11 +1277,12 @@ static int gmc_v9_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + /* ARCT and VEGA20 don't have XGMI defined in their IP discovery tables */ if (adev->asic_type == CHIP_VEGA20 || adev->asic_type == CHIP_ARCTURUS) adev->gmc.xgmi.supported = true; - if (adev->asic_type == CHIP_ALDEBARAN) { + if (adev->ip_versions[XGMI_HWIP][0] == IP_VERSION(6, 1, 0)) { adev->gmc.xgmi.supported = true; adev->gmc.xgmi.connected_to_cpu = adev->smuio.funcs->is_host_gpu_xgmi_supported(adev); @@ -1289,9 +1320,11 @@ static int gmc_v9_0_late_init(void *handle) * Workaround performance drop issue with VBIOS enables partial * writes, while disables HBM ECC for vega10. */ - if (!amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_VEGA10)) { + if (!amdgpu_sriov_vf(adev) && + (adev->ip_versions[UMC_HWIP][0] == IP_VERSION(6, 0, 0))) { if (!(adev->ras_enabled & (1 << AMDGPU_RAS_BLOCK__UMC))) { - if (adev->df.funcs->enable_ecc_force_par_wr_rmw) + if (adev->df.funcs && + adev->df.funcs->enable_ecc_force_par_wr_rmw) adev->df.funcs->enable_ecc_force_par_wr_rmw(adev, false); } } @@ -1393,17 +1426,18 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev) /* set the gart size */ if (amdgpu_gart_size == -1) { - switch (adev->asic_type) { - case CHIP_VEGA10: /* all engines support GPUVM */ - case CHIP_VEGA12: /* all engines support GPUVM */ - case CHIP_VEGA20: - case CHIP_ARCTURUS: - case CHIP_ALDEBARAN: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): /* all engines support GPUVM */ + case IP_VERSION(9, 2, 1): /* all engines support GPUVM */ + case IP_VERSION(9, 4, 0): + case IP_VERSION(9, 4, 1): + case IP_VERSION(9, 4, 2): default: adev->gmc.gart_size = 512ULL << 20; break; - case CHIP_RAVEN: /* DCE SG support */ - case CHIP_RENOIR: + case IP_VERSION(9, 1, 0): /* DCE SG support */ + case IP_VERSION(9, 2, 2): /* DCE SG support */ + case IP_VERSION(9, 3, 0): adev->gmc.gart_size = 1024ULL << 20; break; } @@ -1464,7 +1498,8 @@ static int gmc_v9_0_gart_init(struct amdgpu_device *adev) */ static void gmc_v9_0_save_registers(struct amdgpu_device *adev) { - if (adev->asic_type == CHIP_RAVEN) + if ((adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 0)) || + (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 1))) adev->gmc.sdpif_register = RREG32_SOC15(DCE, 0, mmDCHUBBUB_SDPIF_MMIO_CNTRL_0); } @@ -1500,15 +1535,18 @@ static int gmc_v9_0_sw_init(void *handle) chansize = 64; else chansize = 128; - - numchan = adev->df.funcs->get_hbm_channel_number(adev); - adev->gmc.vram_width = numchan * chansize; + if (adev->df.funcs && + adev->df.funcs->get_hbm_channel_number) { + numchan = adev->df.funcs->get_hbm_channel_number(adev); + adev->gmc.vram_width = numchan * chansize; + } } adev->gmc.vram_type = vram_type; adev->gmc.vram_vendor = vram_vendor; - switch (adev->asic_type) { - case CHIP_RAVEN: + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 2): adev->num_vmhubs = 2; if (adev->rev_id == 0x0 || adev->rev_id == 0x1) { @@ -1520,11 +1558,11 @@ static int gmc_v9_0_sw_init(void *handle) adev->vm_manager.num_level > 1; } break; - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: - case CHIP_RENOIR: - case CHIP_ALDEBARAN: + case IP_VERSION(9, 0, 1): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 4, 0): + case IP_VERSION(9, 3, 0): + case IP_VERSION(9, 4, 2): adev->num_vmhubs = 2; @@ -1539,7 +1577,7 @@ static int gmc_v9_0_sw_init(void *handle) else amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48); break; - case CHIP_ARCTURUS: + case IP_VERSION(9, 4, 1): adev->num_vmhubs = 3; /* Keep the vm size same with Vega20 */ @@ -1555,7 +1593,7 @@ static int gmc_v9_0_sw_init(void *handle) if (r) return r; - if (adev->asic_type == CHIP_ARCTURUS) { + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1)) { r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VMC1, VMC_1_0__SRCID__VM_FAULT, &adev->gmc.vm_fault); if (r) @@ -1590,12 +1628,6 @@ static int gmc_v9_0_sw_init(void *handle) } adev->need_swiotlb = drm_need_swiotlb(44); - if (adev->gmc.xgmi.supported) { - r = adev->gfxhub.funcs->get_xgmi_info(adev); - if (r) - return r; - } - r = gmc_v9_0_mc_init(adev); if (r) return r; @@ -1622,8 +1654,8 @@ static int gmc_v9_0_sw_init(void *handle) * for video processing. */ adev->vm_manager.first_kfd_vmid = - (adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_ALDEBARAN) ? 3 : 8; + (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) ? 3 : 8; amdgpu_vm_manager_init(adev); @@ -1649,12 +1681,12 @@ static int gmc_v9_0_sw_fini(void *handle) static void gmc_v9_0_init_golden_registers(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(9, 0, 0): if (amdgpu_sriov_vf(adev)) break; fallthrough; - case CHIP_VEGA20: + case IP_VERSION(9, 4, 0): soc15_program_register_sequence(adev, golden_settings_mmhub_1_0_0, ARRAY_SIZE(golden_settings_mmhub_1_0_0)); @@ -1662,9 +1694,8 @@ static void gmc_v9_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_athub_1_0_0, ARRAY_SIZE(golden_settings_athub_1_0_0)); break; - case CHIP_VEGA12: - break; - case CHIP_RAVEN: + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 0): /* TODO for renoir */ soc15_program_register_sequence(adev, golden_settings_athub_1_0_0, @@ -1684,7 +1715,8 @@ static void gmc_v9_0_init_golden_registers(struct amdgpu_device *adev) */ void gmc_v9_0_restore_registers(struct amdgpu_device *adev) { - if (adev->asic_type == CHIP_RAVEN) { + if ((adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 0)) || + (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 1))) { WREG32_SOC15(DCE, 0, mmDCHUBBUB_SDPIF_MMIO_CNTRL_0, adev->gmc.sdpif_register); WARN_ON(adev->gmc.sdpif_register != RREG32_SOC15(DCE, 0, mmDCHUBBUB_SDPIF_MMIO_CNTRL_0)); @@ -1708,10 +1740,14 @@ static int gmc_v9_0_gart_enable(struct amdgpu_device *adev) return -EINVAL; } + if (amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev)) + goto skip_pin_bo; + r = amdgpu_gart_table_vram_pin(adev); if (r) return r; +skip_pin_bo: r = adev->gfxhub.funcs->gart_enable(adev); if (r) return r; @@ -1736,7 +1772,7 @@ static int gmc_v9_0_hw_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; bool value; - int r, i; + int i; /* The sequence of these two function calls matters.*/ gmc_v9_0_init_golden_registers(adev); @@ -1771,9 +1807,7 @@ static int gmc_v9_0_hw_init(void *handle) if (adev->umc.funcs && adev->umc.funcs->init_registers) adev->umc.funcs->init_registers(adev); - r = gmc_v9_0_gart_enable(adev); - - return r; + return gmc_v9_0_gart_enable(adev); } /** @@ -1802,6 +1836,14 @@ static int gmc_v9_0_hw_fini(void *handle) return 0; } + /* + * Pair the operations did in gmc_v9_0_hw_init and thus maintain + * a correct cached state for GMC. Otherwise, the "gate" again + * operation on S3 resuming will fail due to wrong cached state. + */ + if (adev->mmhub.funcs->update_power_gating) + adev->mmhub.funcs->update_power_gating(adev, false); + amdgpu_irq_put(adev, &adev->gmc.ecc_irq, 0); amdgpu_irq_put(adev, &adev->gmc.vm_fault, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c index 74b90cc2bf48..eecfb1545c1e 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c @@ -49,7 +49,7 @@ static void hdp_v4_0_flush_hdp(struct amdgpu_device *adev, static void hdp_v4_0_invalidate_hdp(struct amdgpu_device *adev, struct amdgpu_ring *ring) { - if (adev->asic_type == CHIP_ALDEBARAN) + if (adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 4, 0)) return; if (!ring || !ring->funcs->emit_wreg) @@ -79,7 +79,7 @@ static void hdp_v4_0_reset_ras_error_count(struct amdgpu_device *adev) if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__HDP)) return; - if (adev->asic_type >= CHIP_ALDEBARAN) + if (adev->ip_versions[HDP_HWIP][0] >= IP_VERSION(4, 4, 0)) WREG32_SOC15(HDP, 0, mmHDP_EDC_CNT, 0); else /*read back hdp ras counter to reset it to 0 */ @@ -91,9 +91,10 @@ static void hdp_v4_0_update_clock_gating(struct amdgpu_device *adev, { uint32_t def, data; - if (adev->asic_type == CHIP_VEGA10 || - adev->asic_type == CHIP_VEGA12 || - adev->asic_type == CHIP_RAVEN) { + if (adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 0, 0) || + adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 0, 1) || + adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 1, 1) || + adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 1, 0)) { def = data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS)); if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS)) @@ -135,8 +136,8 @@ static void hdp_v4_0_get_clockgating_state(struct amdgpu_device *adev, static void hdp_v4_0_init_registers(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_ARCTURUS: + switch (adev->ip_versions[HDP_HWIP][0]) { + case IP_VERSION(4, 2, 1): WREG32_FIELD15(HDP, 0, HDP_MMHUB_CNTL, HDP_MMHUB_GCC, 1); break; default: diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c index 85967a5570cb..299de1d131d8 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c @@ -32,26 +32,6 @@ #include "vcn/vcn_2_0_0_sh_mask.h" #include "ivsrcid/vcn/irqsrcs_vcn_2_0.h" -#define mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET 0x1bfff -#define mmUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET 0x4029 -#define mmUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET 0x402a -#define mmUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET 0x402b -#define mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40ea -#define mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40eb -#define mmUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET 0x40cf -#define mmUVD_LMI_JPEG_VMID_INTERNAL_OFFSET 0x40d1 -#define mmUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40e8 -#define mmUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40e9 -#define mmUVD_JRBC_IB_SIZE_INTERNAL_OFFSET 0x4082 -#define mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40ec -#define mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40ed -#define mmUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET 0x4085 -#define mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET 0x4084 -#define mmUVD_JRBC_STATUS_INTERNAL_OFFSET 0x4089 -#define mmUVD_JPEG_PITCH_INTERNAL_OFFSET 0x401f - -#define JRBC_DEC_EXTERNAL_REG_WRITE_ADDR 0x18000 - 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(void *handle, diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h index 15a344ed340f..1a03baa59755 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h @@ -24,6 +24,26 @@ #ifndef __JPEG_V2_0_H__ #define __JPEG_V2_0_H__ +#define mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET 0x1bfff +#define mmUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET 0x4029 +#define mmUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET 0x402a +#define mmUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET 0x402b +#define mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40ea +#define mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40eb +#define mmUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET 0x40cf +#define mmUVD_LMI_JPEG_VMID_INTERNAL_OFFSET 0x40d1 +#define mmUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40e8 +#define mmUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40e9 +#define mmUVD_JRBC_IB_SIZE_INTERNAL_OFFSET 0x4082 +#define mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40ec +#define mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40ed +#define mmUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET 0x4085 +#define mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET 0x4084 +#define mmUVD_JRBC_STATUS_INTERNAL_OFFSET 0x4089 +#define mmUVD_JPEG_PITCH_INTERNAL_OFFSET 0x401f + +#define JRBC_DEC_EXTERNAL_REG_WRITE_ADDR 0x18000 + void jpeg_v2_0_dec_ring_insert_start(struct amdgpu_ring *ring); void jpeg_v2_0_dec_ring_insert_end(struct amdgpu_ring *ring); void jpeg_v2_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq, diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c index 46096ad7f0d9..a29c86617fb5 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c @@ -423,6 +423,42 @@ static void jpeg_v2_5_dec_ring_set_wptr(struct amdgpu_ring *ring) } } +/** + * jpeg_v2_6_dec_ring_insert_start - insert a start command + * + * @ring: amdgpu_ring pointer + * + * Write a start command to the ring. + */ +static void jpeg_v2_6_dec_ring_insert_start(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x6aa04); /* PCTL0_MMHUB_DEEPSLEEP_IB */ + + amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x80000000 | (1 << (ring->me * 2 + 14))); +} + +/** + * jpeg_v2_6_dec_ring_insert_end - insert a end command + * + * @ring: amdgpu_ring pointer + * + * Write a end command to the ring. + */ +static void jpeg_v2_6_dec_ring_insert_end(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x6aa04); /* PCTL0_MMHUB_DEEPSLEEP_IB */ + + amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, (1 << (ring->me * 2 + 14))); +} + static bool jpeg_v2_5_is_idle(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -633,8 +669,8 @@ static const struct amdgpu_ring_funcs jpeg_v2_6_dec_ring_vm_funcs = { .test_ring = amdgpu_jpeg_dec_ring_test_ring, .test_ib = amdgpu_jpeg_dec_ring_test_ib, .insert_nop = jpeg_v2_0_dec_ring_nop, - .insert_start = jpeg_v2_0_dec_ring_insert_start, - .insert_end = jpeg_v2_0_dec_ring_insert_end, + .insert_start = jpeg_v2_6_dec_ring_insert_start, + .insert_end = jpeg_v2_6_dec_ring_insert_end, .pad_ib = amdgpu_ring_generic_pad_ib, .begin_use = amdgpu_jpeg_ring_begin_use, .end_use = amdgpu_jpeg_ring_end_use, diff --git a/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c index 058b65730a84..8f7107d392af 100644 --- a/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c @@ -52,7 +52,8 @@ const struct amdgpu_mca_ras_funcs mca_v3_0_mp0_ras_funcs = { .ras_fini = mca_v3_0_mp0_ras_fini, .query_ras_error_count = mca_v3_0_mp0_query_ras_error_count, .query_ras_error_address = NULL, - .ras_block = AMDGPU_RAS_BLOCK__MP0, + .ras_block = AMDGPU_RAS_BLOCK__MCA, + .ras_sub_block = AMDGPU_RAS_MCA_BLOCK__MP0, .sysfs_name = "mp0_err_count", }; @@ -79,7 +80,8 @@ const struct amdgpu_mca_ras_funcs mca_v3_0_mp1_ras_funcs = { .ras_fini = mca_v3_0_mp1_ras_fini, .query_ras_error_count = mca_v3_0_mp1_query_ras_error_count, .query_ras_error_address = NULL, - .ras_block = AMDGPU_RAS_BLOCK__MP1, + .ras_block = AMDGPU_RAS_BLOCK__MCA, + .ras_sub_block = AMDGPU_RAS_MCA_BLOCK__MP1, .sysfs_name = "mp1_err_count", }; @@ -106,7 +108,8 @@ const struct amdgpu_mca_ras_funcs mca_v3_0_mpio_ras_funcs = { .ras_fini = mca_v3_0_mpio_ras_fini, .query_ras_error_count = mca_v3_0_mpio_query_ras_error_count, .query_ras_error_address = NULL, - .ras_block = AMDGPU_RAS_BLOCK__MPIO, + .ras_block = AMDGPU_RAS_BLOCK__MCA, + .ras_sub_block = AMDGPU_RAS_MCA_BLOCK__MPIO, .sysfs_name = "mpio_err_count", }; diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c index a99953833820..1da2ec692057 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c @@ -145,7 +145,6 @@ static void mmhub_v1_0_init_tlb_regs(struct amdgpu_device *adev) ENABLE_ADVANCED_DRIVER_MODEL, 1); tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0); tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, MTYPE, MTYPE_UC);/* XXX for emulation. */ tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); @@ -302,10 +301,10 @@ static void mmhub_v1_0_update_power_gating(struct amdgpu_device *adev, if (amdgpu_sriov_vf(adev)) return; - if (enable && adev->pg_flags & AMD_PG_SUPPORT_MMHUB) { - amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GMC, true); - - } + if (adev->pg_flags & AMD_PG_SUPPORT_MMHUB) + amdgpu_dpm_set_powergating_by_smu(adev, + AMD_IP_BLOCK_TYPE_GMC, + enable); } static int mmhub_v1_0_gart_enable(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c index f80a14a1b82d..f5f7181f9af5 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c @@ -165,7 +165,6 @@ static void mmhub_v1_7_init_tlb_regs(struct amdgpu_device *adev) ENABLE_ADVANCED_DRIVER_MODEL, 1); tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0); tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, MTYPE, MTYPE_UC);/* XXX for emulation. */ tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c index 7ded6b2f058e..3718ff610ab2 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c @@ -153,18 +153,16 @@ mmhub_v2_0_print_l2_protection_fault_status(struct amdgpu_device *adev, dev_err(adev->dev, "MMVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n", status); - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI12: - case CHIP_NAVI14: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(2, 0, 0): + case IP_VERSION(2, 0, 2): mmhub_cid = mmhub_client_ids_navi1x[cid][rw]; break; - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: + case IP_VERSION(2, 1, 0): + case IP_VERSION(2, 1, 1): mmhub_cid = mmhub_client_ids_sienna_cichlid[cid][rw]; break; - case CHIP_BEIGE_GOBY: + case IP_VERSION(2, 1, 2): mmhub_cid = mmhub_client_ids_beige_goby[cid][rw]; break; default: @@ -269,7 +267,6 @@ static void mmhub_v2_0_init_tlb_regs(struct amdgpu_device *adev) ENABLE_ADVANCED_DRIVER_MODEL, 1); tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); - tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0); tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, MTYPE, MTYPE_UC); /* UC, uncached */ @@ -571,11 +568,10 @@ static void mmhub_v2_0_update_medium_grain_clock_gating(struct amdgpu_device *ad if (!(adev->cg_flags & AMD_CG_SUPPORT_MC_MGCG)) return; - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(2, 1, 0): + case IP_VERSION(2, 1, 1): + case IP_VERSION(2, 1, 2): def = data = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid); def1 = data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid); break; @@ -606,11 +602,10 @@ static void mmhub_v2_0_update_medium_grain_clock_gating(struct amdgpu_device *ad DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK); } - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(2, 1, 0): + case IP_VERSION(2, 1, 1): + case IP_VERSION(2, 1, 2): if (def != data) WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid, data); if (def1 != data1) @@ -633,11 +628,10 @@ static void mmhub_v2_0_update_medium_grain_light_sleep(struct amdgpu_device *ade if (!(adev->cg_flags & AMD_CG_SUPPORT_MC_LS)) return; - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(2, 1, 0): + case IP_VERSION(2, 1, 1): + case IP_VERSION(2, 1, 2): def = data = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid); break; default: @@ -651,11 +645,10 @@ static void mmhub_v2_0_update_medium_grain_light_sleep(struct amdgpu_device *ade data &= ~MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK; if (def != data) { - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(2, 1, 0): + case IP_VERSION(2, 1, 1): + case IP_VERSION(2, 1, 2): WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid, data); break; default: @@ -671,14 +664,12 @@ static int mmhub_v2_0_set_clockgating(struct amdgpu_device *adev, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(2, 0, 0): + case IP_VERSION(2, 0, 2): + case IP_VERSION(2, 1, 0): + case IP_VERSION(2, 1, 1): + case IP_VERSION(2, 1, 2): mmhub_v2_0_update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); mmhub_v2_0_update_medium_grain_light_sleep(adev, @@ -698,11 +689,10 @@ static void mmhub_v2_0_get_clockgating(struct amdgpu_device *adev, u32 *flags) if (amdgpu_sriov_vf(adev)) *flags = 0; - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(2, 1, 0): + case IP_VERSION(2, 1, 1): + case IP_VERSION(2, 1, 2): data = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid); data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid); break; diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c index 88e457a150e0..9e16da28505a 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c @@ -90,9 +90,9 @@ mmhub_v2_3_print_l2_protection_fault_status(struct amdgpu_device *adev, dev_err(adev->dev, "MMVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n", status); - switch (adev->asic_type) { - case CHIP_VANGOGH: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[MMHUB_HWIP][0]) { + case IP_VERSION(2, 3, 0): + case IP_VERSION(2, 4, 0): mmhub_cid = mmhub_client_ids_vangogh[cid][rw]; break; default: @@ -194,7 +194,6 @@ static void mmhub_v2_3_init_tlb_regs(struct amdgpu_device *adev) ENABLE_ADVANCED_DRIVER_MODEL, 1); tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); - tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0); tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, MTYPE, MTYPE_UC); /* UC, uncached */ diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c index c4ef822bbe8c..ff49eeaf7882 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c @@ -190,8 +190,6 @@ static void mmhub_v9_4_init_tlb_regs(struct amdgpu_device *adev, int hubid) tmp = REG_SET_FIELD(tmp, VMSHAREDVC0_MC_VM_MX_L1_TLB_CNTL, SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); tmp = REG_SET_FIELD(tmp, VMSHAREDVC0_MC_VM_MX_L1_TLB_CNTL, - ECO_BITS, 0); - tmp = REG_SET_FIELD(tmp, VMSHAREDVC0_MC_VM_MX_L1_TLB_CNTL, MTYPE, MTYPE_UC);/* XXX for emulation. */ tmp = REG_SET_FIELD(tmp, VMSHAREDVC0_MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c index 23b066bcffb2..0077e738db31 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c @@ -252,11 +252,12 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work) * otherwise the mailbox msg will be ruined/reseted by * the VF FLR. */ - if (!down_write_trylock(&adev->reset_sem)) + if (atomic_cmpxchg(&adev->in_gpu_reset, 0, 1) != 0) return; + down_write(&adev->reset_sem); + amdgpu_virt_fini_data_exchange(adev); - atomic_set(&adev->in_gpu_reset, 1); xgpu_ai_mailbox_trans_msg(adev, IDH_READY_TO_RESET, 0, 0, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h index bd3b23171579..f9aa4d0bb638 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h @@ -26,7 +26,7 @@ #define AI_MAILBOX_POLL_ACK_TIMEDOUT 500 #define AI_MAILBOX_POLL_MSG_TIMEDOUT 6000 -#define AI_MAILBOX_POLL_FLR_TIMEDOUT 5000 +#define AI_MAILBOX_POLL_FLR_TIMEDOUT 10000 #define AI_MAILBOX_POLL_MSG_REP_MAX 11 enum idh_request { diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c index a35e6d87e537..477d0dde19c5 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c @@ -281,11 +281,12 @@ static void xgpu_nv_mailbox_flr_work(struct work_struct *work) * otherwise the mailbox msg will be ruined/reseted by * the VF FLR. */ - if (!down_write_trylock(&adev->reset_sem)) + if (atomic_cmpxchg(&adev->in_gpu_reset, 0, 1) != 0) return; + down_write(&adev->reset_sem); + amdgpu_virt_fini_data_exchange(adev); - atomic_set(&adev->in_gpu_reset, 1); xgpu_nv_mailbox_trans_msg(adev, IDH_READY_TO_RESET, 0, 0, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c index 530011622801..8ce5b8ca1fd7 100644 --- a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c @@ -107,7 +107,7 @@ force_update_wptr_for_self_int(struct amdgpu_device *adev, { u32 ih_cntl, ih_rb_cntl; - if (adev->asic_type < CHIP_SIENNA_CICHLID) + if (adev->ip_versions[OSSSYS_HWIP][0] < IP_VERSION(5, 0, 3)) return; ih_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_CNTL2); @@ -160,6 +160,7 @@ static int navi10_ih_toggle_ring_interrupts(struct amdgpu_device *adev, tmp = RREG32(ih_regs->ih_rb_cntl); tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0)); + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_GPU_TS_ENABLE, 1); /* enable_intr field is only valid in ring0 */ if (ih == &adev->irq.ih) tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0)); @@ -275,10 +276,8 @@ static int navi10_ih_enable_ring(struct amdgpu_device *adev, tmp = navi10_ih_rb_cntl(ih, tmp); if (ih == &adev->irq.ih) tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RPTR_REARM, !!adev->irq.msi_enabled); - if (ih == &adev->irq.ih1) { - tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0); + if (ih == &adev->irq.ih1) tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1); - } if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) { if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) { @@ -319,7 +318,6 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev) { struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1, &adev->irq.ih2}; u32 ih_chicken; - u32 tmp; int ret; int i; @@ -332,13 +330,10 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev) if (unlikely(adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT)) { if (ih[0]->use_bus_addr) { - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[OSSSYS_HWIP][0]) { + case IP_VERSION(5, 0, 3): + case IP_VERSION(5, 2, 0): + case IP_VERSION(5, 2, 1): ih_chicken = RREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN_Sienna_Cichlid); ih_chicken = REG_SET_FIELD(ih_chicken, IH_CHICKEN, MC_SPACE_GPA_ENABLE, 1); @@ -366,15 +361,6 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev) adev->nbio.funcs->ih_doorbell_range(adev, ih[0]->use_doorbell, ih[0]->doorbell_index); - tmp = RREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL); - tmp = REG_SET_FIELD(tmp, IH_STORM_CLIENT_LIST_CNTL, - CLIENT18_IS_STORM_CLIENT, 1); - WREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL, tmp); - - tmp = RREG32_SOC15(OSSSYS, 0, mmIH_INT_FLOOD_CNTL); - tmp = REG_SET_FIELD(tmp, IH_INT_FLOOD_CNTL, FLOOD_CNTL_ENABLE, 1); - WREG32_SOC15(OSSSYS, 0, mmIH_INT_FLOOD_CNTL, tmp); - pci_set_master(adev->pdev); /* enable interrupts */ @@ -423,12 +409,19 @@ static u32 navi10_ih_get_wptr(struct amdgpu_device *adev, u32 wptr, tmp; struct amdgpu_ih_regs *ih_regs; - wptr = le32_to_cpu(*ih->wptr_cpu); - ih_regs = &ih->ih_regs; + if (ih == &adev->irq.ih) { + /* Only ring0 supports writeback. On other rings fall back + * to register-based code with overflow checking below. + */ + wptr = le32_to_cpu(*ih->wptr_cpu); - if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) - goto out; + if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) + goto out; + } + ih_regs = &ih->ih_regs; + + /* Double check that the overflow wasn't already cleared. */ wptr = RREG32_NO_KIQ(ih_regs->ih_rb_wptr); if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) goto out; @@ -516,15 +509,11 @@ static int navi10_ih_self_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) { - uint32_t wptr = cpu_to_le32(entry->src_data[0]); - switch (entry->ring_id) { case 1: - *adev->irq.ih1.wptr_cpu = wptr; schedule_work(&adev->irq.ih1_work); break; case 2: - *adev->irq.ih2.wptr_cpu = wptr; schedule_work(&adev->irq.ih2_work); break; default: break; @@ -727,6 +716,7 @@ static const struct amd_ip_funcs navi10_ih_ip_funcs = { static const struct amdgpu_ih_funcs navi10_ih_funcs = { .get_wptr = navi10_ih_get_wptr, .decode_iv = amdgpu_ih_decode_iv_helper, + .decode_iv_ts = amdgpu_ih_decode_iv_ts_helper, .set_rptr = navi10_ih_set_rptr }; diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_reg_init.c b/drivers/gpu/drm/amd/amdgpu/navi10_reg_init.c deleted file mode 100644 index 88efaecf9f70..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/navi10_reg_init.c +++ /dev/null @@ -1,55 +0,0 @@ -/* - * Copyright 2018 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ -#include "amdgpu.h" -#include "nv.h" - -#include "soc15_common.h" -#include "navi10_ip_offset.h" - -int navi10_reg_base_init(struct amdgpu_device *adev) -{ - int i; - - for (i = 0 ; i < MAX_INSTANCE ; ++i) { - adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i])); - adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i])); - adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i])); - adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIO_BASE.instance[i])); - adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i])); - adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i])); - adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(VCN_BASE.instance[i])); - adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i])); - adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DCN_BASE.instance[i])); - adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i])); - adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA1_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i])); - adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i])); - adev->reg_offset[CLK_HWIP][i] = (uint32_t *)(&(CLK_BASE.instance[i])); - } - - return 0; -} - - diff --git a/drivers/gpu/drm/amd/amdgpu/navi12_reg_init.c b/drivers/gpu/drm/amd/amdgpu/navi12_reg_init.c deleted file mode 100644 index a786d159e5e9..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/navi12_reg_init.c +++ /dev/null @@ -1,52 +0,0 @@ -/* - * Copyright 2018 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ -#include "amdgpu.h" -#include "nv.h" - -#include "soc15_common.h" -#include "navi12_ip_offset.h" - -int navi12_reg_base_init(struct amdgpu_device *adev) -{ - /* HW has more IP blocks, only initialized the blocks needed by driver */ - uint32_t i; - for (i = 0 ; i < MAX_INSTANCE ; ++i) { - adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i])); - adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i])); - adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i])); - adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIF0_BASE.instance[i])); - adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i])); - adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i])); - adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(UVD0_BASE.instance[i])); - adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i])); - adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DMU_BASE.instance[i])); - adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i])); - adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA1_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i])); - adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i])); - adev->reg_offset[CLK_HWIP][i] = (uint32_t *)(&(CLK_BASE.instance[i])); - } - return 0; -} diff --git a/drivers/gpu/drm/amd/amdgpu/navi14_reg_init.c b/drivers/gpu/drm/amd/amdgpu/navi14_reg_init.c deleted file mode 100644 index 4ea1e8fbb601..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/navi14_reg_init.c +++ /dev/null @@ -1,53 +0,0 @@ -/* - * Copyright 2018 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ -#include "amdgpu.h" -#include "nv.h" - -#include "soc15_common.h" -#include "navi14_ip_offset.h" - -int navi14_reg_base_init(struct amdgpu_device *adev) -{ - int i; - - for (i = 0 ; i < MAX_INSTANCE ; ++i) { - adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i])); - adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i])); - adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i])); - adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIF0_BASE.instance[i])); - adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i])); - adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i])); - adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(UVD0_BASE.instance[i])); - adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i])); - adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DMU_BASE.instance[i])); - adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i])); - adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA1_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i])); - adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i])); - adev->reg_offset[CLK_HWIP][i] = (uint32_t *)(&(CLK_BASE.instance[i])); - } - - return 0; -} diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c index b184b656b9b6..ee7cab37dfd5 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c @@ -53,6 +53,16 @@ #define smnPCIE_LC_LINK_WIDTH_CNTL 0x11140288 +#define GPU_HDP_FLUSH_DONE__RSVD_ENG0_MASK 0x00001000L /* Don't use. Firmware uses this bit internally */ +#define GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK 0x00002000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK 0x00004000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK 0x00008000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK 0x00010000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK 0x00020000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG6_MASK 0x00040000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG7_MASK 0x00080000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG8_MASK 0x00100000L + static void nbio_v2_3_remap_hdp_registers(struct amdgpu_device *adev) { WREG32_SOC15(NBIO, 0, mmREMAP_HDP_MEM_FLUSH_CNTL, @@ -318,6 +328,27 @@ const struct nbio_hdp_flush_reg nbio_v2_3_hdp_flush_reg = { .ref_and_mask_sdma1 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__SDMA1_MASK, }; +const struct nbio_hdp_flush_reg nbio_v2_3_hdp_flush_reg_sc = { + .ref_and_mask_cp0 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP0_MASK, + .ref_and_mask_cp1 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP1_MASK, + .ref_and_mask_cp2 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP2_MASK, + .ref_and_mask_cp3 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP3_MASK, + .ref_and_mask_cp4 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP4_MASK, + .ref_and_mask_cp5 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP5_MASK, + .ref_and_mask_cp6 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP6_MASK, + .ref_and_mask_cp7 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP7_MASK, + .ref_and_mask_cp8 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP8_MASK, + .ref_and_mask_cp9 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP9_MASK, + .ref_and_mask_sdma0 = GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK, + .ref_and_mask_sdma1 = GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK, + .ref_and_mask_sdma2 = GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK, + .ref_and_mask_sdma3 = GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK, + .ref_and_mask_sdma4 = GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK, + .ref_and_mask_sdma5 = GPU_HDP_FLUSH_DONE__RSVD_ENG6_MASK, + .ref_and_mask_sdma6 = GPU_HDP_FLUSH_DONE__RSVD_ENG7_MASK, + .ref_and_mask_sdma7 = GPU_HDP_FLUSH_DONE__RSVD_ENG8_MASK, +}; + static void nbio_v2_3_init_registers(struct amdgpu_device *adev) { uint32_t def, data; @@ -328,6 +359,10 @@ static void nbio_v2_3_init_registers(struct amdgpu_device *adev) if (def != data) WREG32_PCIE(smnPCIE_CONFIG_CNTL, data); + + if (amdgpu_sriov_vf(adev)) + adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0, + mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2; } #define NAVI10_PCIE__LC_L0S_INACTIVITY_DEFAULT 0x00000000 // off by default, no gains over L1 diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.h b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.h index a43b60acf7f6..6074dd3a1ed8 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.h +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.h @@ -27,6 +27,7 @@ #include "soc15_common.h" extern const struct nbio_hdp_flush_reg nbio_v2_3_hdp_flush_reg; +extern const struct nbio_hdp_flush_reg nbio_v2_3_hdp_flush_reg_sc; extern const struct amdgpu_nbio_funcs nbio_v2_3_funcs; #endif diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c index 0d2d629e2d6a..4bbacf1be25a 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c @@ -276,6 +276,10 @@ static void nbio_v6_1_init_registers(struct amdgpu_device *adev) if (def != data) WREG32_PCIE(smnPCIE_CI_CNTL, data); + + if (amdgpu_sriov_vf(adev)) + adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0, + mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2; } static void nbio_v6_1_program_ltr(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c index 3c00666a13e1..37a4039fdfc5 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c @@ -273,7 +273,9 @@ const struct nbio_hdp_flush_reg nbio_v7_0_hdp_flush_reg = { static void nbio_v7_0_init_registers(struct amdgpu_device *adev) { - + if (amdgpu_sriov_vf(adev)) + adev->rmmio_remap.reg_offset = + SOC15_REG_OFFSET(NBIO, 0, mmHDP_MEM_COHERENCY_FLUSH_CNTL) << 2; } const struct amdgpu_nbio_funcs nbio_v7_0_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c index 8f2a315e7c73..3444332ea110 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c @@ -371,6 +371,10 @@ static void nbio_v7_2_init_registers(struct amdgpu_device *adev) if (def != data) WREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regPCIE_CONFIG_CNTL), data); } + + if (amdgpu_sriov_vf(adev)) + adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0, + regBIF_BX_PF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2; } const struct amdgpu_nbio_funcs nbio_v7_2_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c index f50045cebd44..dc5e93756fea 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c @@ -56,12 +56,15 @@ * These are nbio v7_4_1 registers mask. Temporarily define these here since * nbio v7_4_1 header is incomplete. */ -#define GPU_HDP_FLUSH_DONE__RSVD_ENG0_MASK 0x00001000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG0_MASK 0x00001000L /* Don't use. Firmware uses this bit internally */ #define GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK 0x00002000L #define GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK 0x00004000L #define GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK 0x00008000L #define GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK 0x00010000L #define GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK 0x00020000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG6_MASK 0x00040000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG7_MASK 0x00080000L +#define GPU_HDP_FLUSH_DONE__RSVD_ENG8_MASK 0x00100000L #define mmBIF_MMSCH1_DOORBELL_RANGE 0x01dc #define mmBIF_MMSCH1_DOORBELL_RANGE_BASE_IDX 2 @@ -334,17 +337,34 @@ const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg = { .ref_and_mask_cp9 = GPU_HDP_FLUSH_DONE__CP9_MASK, .ref_and_mask_sdma0 = GPU_HDP_FLUSH_DONE__SDMA0_MASK, .ref_and_mask_sdma1 = GPU_HDP_FLUSH_DONE__SDMA1_MASK, - .ref_and_mask_sdma2 = GPU_HDP_FLUSH_DONE__RSVD_ENG0_MASK, - .ref_and_mask_sdma3 = GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK, - .ref_and_mask_sdma4 = GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK, - .ref_and_mask_sdma5 = GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK, - .ref_and_mask_sdma6 = GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK, - .ref_and_mask_sdma7 = GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK, +}; + +const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg_ald = { + .ref_and_mask_cp0 = GPU_HDP_FLUSH_DONE__CP0_MASK, + .ref_and_mask_cp1 = GPU_HDP_FLUSH_DONE__CP1_MASK, + .ref_and_mask_cp2 = GPU_HDP_FLUSH_DONE__CP2_MASK, + .ref_and_mask_cp3 = GPU_HDP_FLUSH_DONE__CP3_MASK, + .ref_and_mask_cp4 = GPU_HDP_FLUSH_DONE__CP4_MASK, + .ref_and_mask_cp5 = GPU_HDP_FLUSH_DONE__CP5_MASK, + .ref_and_mask_cp6 = GPU_HDP_FLUSH_DONE__CP6_MASK, + .ref_and_mask_cp7 = GPU_HDP_FLUSH_DONE__CP7_MASK, + .ref_and_mask_cp8 = GPU_HDP_FLUSH_DONE__CP8_MASK, + .ref_and_mask_cp9 = GPU_HDP_FLUSH_DONE__CP9_MASK, + .ref_and_mask_sdma0 = GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK, + .ref_and_mask_sdma1 = GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK, + .ref_and_mask_sdma2 = GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK, + .ref_and_mask_sdma3 = GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK, + .ref_and_mask_sdma4 = GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK, + .ref_and_mask_sdma5 = GPU_HDP_FLUSH_DONE__RSVD_ENG6_MASK, + .ref_and_mask_sdma6 = GPU_HDP_FLUSH_DONE__RSVD_ENG7_MASK, + .ref_and_mask_sdma7 = GPU_HDP_FLUSH_DONE__RSVD_ENG8_MASK, }; static void nbio_v7_4_init_registers(struct amdgpu_device *adev) { - + if (amdgpu_sriov_vf(adev)) + adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0, + mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2; } static void nbio_v7_4_handle_ras_controller_intr_no_bifring(struct amdgpu_device *adev) @@ -387,13 +407,13 @@ static void nbio_v7_4_handle_ras_controller_intr_no_bifring(struct amdgpu_device "errors detected in %s block, " "no user action is needed.\n", obj->err_data.ce_count, - ras_block_str(adev->nbio.ras_if->block)); + get_ras_block_str(adev->nbio.ras_if)); if (err_data.ue_count) dev_info(adev->dev, "%ld uncorrectable hardware " "errors detected in %s block\n", obj->err_data.ue_count, - ras_block_str(adev->nbio.ras_if->block)); + get_ras_block_str(adev->nbio.ras_if)); } dev_info(adev->dev, "RAS controller interrupt triggered " @@ -566,7 +586,9 @@ static int nbio_v7_4_init_ras_err_event_athub_interrupt (struct amdgpu_device *a return r; } -#define smnPARITY_ERROR_STATUS_UNCORR_GRP2 0x13a20030 +#define smnPARITY_ERROR_STATUS_UNCORR_GRP2 0x13a20030 +#define smnPARITY_ERROR_STATUS_UNCORR_GRP2_ALDE 0x13b20030 +#define smnRAS_GLOBAL_STATUS_LO_ALDE 0x13b20020 static void nbio_v7_4_query_ras_error_count(struct amdgpu_device *adev, void *ras_error_status) @@ -575,12 +597,20 @@ static void nbio_v7_4_query_ras_error_count(struct amdgpu_device *adev, uint32_t corr, fatal, non_fatal; struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; - global_sts = RREG32_PCIE(smnRAS_GLOBAL_STATUS_LO); + if (adev->asic_type == CHIP_ALDEBARAN) + global_sts = RREG32_PCIE(smnRAS_GLOBAL_STATUS_LO_ALDE); + else + global_sts = RREG32_PCIE(smnRAS_GLOBAL_STATUS_LO); + corr = REG_GET_FIELD(global_sts, RAS_GLOBAL_STATUS_LO, ParityErrCorr); fatal = REG_GET_FIELD(global_sts, RAS_GLOBAL_STATUS_LO, ParityErrFatal); non_fatal = REG_GET_FIELD(global_sts, RAS_GLOBAL_STATUS_LO, ParityErrNonFatal); - parity_sts = RREG32_PCIE(smnPARITY_ERROR_STATUS_UNCORR_GRP2); + + if (adev->asic_type == CHIP_ALDEBARAN) + parity_sts = RREG32_PCIE(smnPARITY_ERROR_STATUS_UNCORR_GRP2_ALDE); + else + parity_sts = RREG32_PCIE(smnPARITY_ERROR_STATUS_UNCORR_GRP2); if (corr) err_data->ce_count++; @@ -589,13 +619,21 @@ static void nbio_v7_4_query_ras_error_count(struct amdgpu_device *adev, if (corr || fatal || non_fatal) { central_sts = RREG32_PCIE(smnBIFL_RAS_CENTRAL_STATUS); + /* clear error status register */ - WREG32_PCIE(smnRAS_GLOBAL_STATUS_LO, global_sts); + if (adev->asic_type == CHIP_ALDEBARAN) + WREG32_PCIE(smnRAS_GLOBAL_STATUS_LO_ALDE, global_sts); + else + WREG32_PCIE(smnRAS_GLOBAL_STATUS_LO, global_sts); if (fatal) + { /* clear parity fatal error indication field */ - WREG32_PCIE(smnPARITY_ERROR_STATUS_UNCORR_GRP2, - parity_sts); + if (adev->asic_type == CHIP_ALDEBARAN) + WREG32_PCIE(smnPARITY_ERROR_STATUS_UNCORR_GRP2_ALDE, parity_sts); + else + WREG32_PCIE(smnPARITY_ERROR_STATUS_UNCORR_GRP2, parity_sts); + } if (REG_GET_FIELD(central_sts, BIFL_RAS_CENTRAL_STATUS, BIFL_RasContller_Intr_Recv)) { @@ -656,6 +694,9 @@ static void nbio_v7_4_program_aspm(struct amdgpu_device *adev) { uint32_t def, data; + if (adev->ip_versions[NBIO_HWIP][0] == IP_VERSION(7, 4, 4)) + return; + def = data = RREG32_PCIE(smnPCIE_LC_CNTL); data &= ~PCIE_LC_CNTL__LC_L1_INACTIVITY_MASK; data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK; diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.h b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.h index b8216581ec8d..cc5692db6f98 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.h +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.h @@ -27,6 +27,7 @@ #include "soc15_common.h" extern const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg; +extern const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg_ald; extern const struct amdgpu_nbio_funcs nbio_v7_4_funcs; extern const struct amdgpu_nbio_ras_funcs nbio_v7_4_ras_funcs; diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c index ff80786e3918..2ec1ffb36b1f 100644 --- a/drivers/gpu/drm/amd/amdgpu/nv.c +++ b/drivers/gpu/drm/amd/amdgpu/nv.c @@ -180,8 +180,10 @@ static const struct amdgpu_video_codecs yc_video_codecs_decode = { static int nv_query_video_codecs(struct amdgpu_device *adev, bool encode, const struct amdgpu_video_codecs **codecs) { - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: + switch (adev->ip_versions[UVD_HWIP][0]) { + case IP_VERSION(3, 0, 0): + case IP_VERSION(3, 0, 64): + case IP_VERSION(3, 0, 192): if (amdgpu_sriov_vf(adev)) { if (encode) *codecs = &sriov_sc_video_codecs_encode; @@ -194,29 +196,27 @@ static int nv_query_video_codecs(struct amdgpu_device *adev, bool encode, *codecs = &sc_video_codecs_decode; } return 0; - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_VANGOGH: + case IP_VERSION(3, 0, 16): + case IP_VERSION(3, 0, 2): if (encode) *codecs = &nv_video_codecs_encode; else *codecs = &sc_video_codecs_decode; return 0; - case CHIP_YELLOW_CARP: + case IP_VERSION(3, 1, 1): if (encode) *codecs = &nv_video_codecs_encode; else *codecs = &yc_video_codecs_decode; return 0; - case CHIP_BEIGE_GOBY: + case IP_VERSION(3, 0, 33): if (encode) *codecs = &bg_video_codecs_encode; else *codecs = &bg_video_codecs_decode; return 0; - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: + case IP_VERSION(2, 0, 0): + case IP_VERSION(2, 0, 2): if (encode) *codecs = &nv_video_codecs_encode; else @@ -511,14 +511,15 @@ nv_asic_reset_method(struct amdgpu_device *adev) dev_warn(adev->dev, "Specified reset method:%d isn't supported, using AUTO instead.\n", amdgpu_reset_method); - switch (adev->asic_type) { - case CHIP_VANGOGH: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[MP1_HWIP][0]) { + case IP_VERSION(11, 5, 0): + case IP_VERSION(13, 0, 1): + case IP_VERSION(13, 0, 3): return AMD_RESET_METHOD_MODE2; - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + case IP_VERSION(11, 0, 7): + case IP_VERSION(11, 0, 11): + case IP_VERSION(11, 0, 12): + case IP_VERSION(11, 0, 13): return AMD_RESET_METHOD_MODE1; default: if (amdgpu_dpm_is_baco_supported(adev)) @@ -599,7 +600,7 @@ static void nv_enable_doorbell_aperture(struct amdgpu_device *adev, adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, enable); } -static const struct amdgpu_ip_block_version nv_common_ip_block = +const struct amdgpu_ip_block_version nv_common_ip_block = { .type = AMD_IP_BLOCK_TYPE_COMMON, .major = 1, @@ -608,314 +609,11 @@ static const struct amdgpu_ip_block_version nv_common_ip_block = .funcs = &nv_common_ip_funcs, }; -static bool nv_is_headless_sku(struct pci_dev *pdev) -{ - if ((pdev->device == 0x731E && - (pdev->revision == 0xC6 || pdev->revision == 0xC7)) || - (pdev->device == 0x7340 && pdev->revision == 0xC9) || - (pdev->device == 0x7360 && pdev->revision == 0xC7)) - return true; - return false; -} - -static int nv_reg_base_init(struct amdgpu_device *adev) -{ - int r; - - if (amdgpu_discovery) { - r = amdgpu_discovery_reg_base_init(adev); - if (r) { - DRM_WARN("failed to init reg base from ip discovery table, " - "fallback to legacy init method\n"); - goto legacy_init; - } - - amdgpu_discovery_harvest_ip(adev); - if (nv_is_headless_sku(adev->pdev)) { - adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK; - adev->harvest_ip_mask |= AMD_HARVEST_IP_JPEG_MASK; - } - - return 0; - } - -legacy_init: - switch (adev->asic_type) { - case CHIP_NAVI10: - navi10_reg_base_init(adev); - break; - case CHIP_NAVI14: - navi14_reg_base_init(adev); - break; - case CHIP_NAVI12: - navi12_reg_base_init(adev); - break; - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - sienna_cichlid_reg_base_init(adev); - break; - case CHIP_VANGOGH: - vangogh_reg_base_init(adev); - break; - case CHIP_DIMGREY_CAVEFISH: - dimgrey_cavefish_reg_base_init(adev); - break; - case CHIP_BEIGE_GOBY: - beige_goby_reg_base_init(adev); - break; - case CHIP_YELLOW_CARP: - yellow_carp_reg_base_init(adev); - break; - case CHIP_CYAN_SKILLFISH: - cyan_skillfish_reg_base_init(adev); - break; - default: - return -EINVAL; - } - - return 0; -} - void nv_set_virt_ops(struct amdgpu_device *adev) { adev->virt.ops = &xgpu_nv_virt_ops; } -int nv_set_ip_blocks(struct amdgpu_device *adev) -{ - int r; - - if (adev->asic_type == CHIP_CYAN_SKILLFISH) { - adev->nbio.funcs = &nbio_v2_3_funcs; - adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg; - } else if (adev->flags & AMD_IS_APU) { - adev->nbio.funcs = &nbio_v7_2_funcs; - adev->nbio.hdp_flush_reg = &nbio_v7_2_hdp_flush_reg; - } else { - adev->nbio.funcs = &nbio_v2_3_funcs; - adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg; - } - adev->hdp.funcs = &hdp_v5_0_funcs; - - if (adev->asic_type >= CHIP_SIENNA_CICHLID) - adev->smuio.funcs = &smuio_v11_0_6_funcs; - else - adev->smuio.funcs = &smuio_v11_0_funcs; - - if (adev->asic_type == CHIP_SIENNA_CICHLID) - adev->gmc.xgmi.supported = true; - - /* Set IP register base before any HW register access */ - r = nv_reg_base_init(adev); - if (r) - return r; - - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - amdgpu_device_ip_block_add(adev, &nv_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP && - !amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v5_0_ip_block); - if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT && - !amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &vcn_v2_0_ip_block); - amdgpu_device_ip_block_add(adev, &jpeg_v2_0_ip_block); - if (adev->enable_mes) - amdgpu_device_ip_block_add(adev, &mes_v10_1_ip_block); - break; - case CHIP_NAVI12: - amdgpu_device_ip_block_add(adev, &nv_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); - if (!amdgpu_sriov_vf(adev)) { - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - } else { - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - } - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v5_0_ip_block); - if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT && - !amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &vcn_v2_0_ip_block); - if (!amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &jpeg_v2_0_ip_block); - break; - case CHIP_SIENNA_CICHLID: - amdgpu_device_ip_block_add(adev, &nv_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); - if (!amdgpu_sriov_vf(adev)) { - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - } else { - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - } - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP && - is_support_sw_smu(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block); - amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); - if (!amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block); - if (adev->enable_mes) - amdgpu_device_ip_block_add(adev, &mes_v10_1_ip_block); - break; - case CHIP_NAVY_FLOUNDER: - amdgpu_device_ip_block_add(adev, &nv_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP && - is_support_sw_smu(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block); - amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); - amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block); - if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT && - is_support_sw_smu(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - break; - case CHIP_VANGOGH: - amdgpu_device_ip_block_add(adev, &nv_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block); - amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); - amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block); - break; - case CHIP_DIMGREY_CAVEFISH: - amdgpu_device_ip_block_add(adev, &nv_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP && - is_support_sw_smu(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block); - amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); - amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block); - break; - case CHIP_BEIGE_GOBY: - amdgpu_device_ip_block_add(adev, &nv_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP && - is_support_sw_smu(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT && - is_support_sw_smu(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); - break; - case CHIP_YELLOW_CARP: - amdgpu_device_ip_block_add(adev, &nv_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v13_0_ip_block); - amdgpu_device_ip_block_add(adev, &smu_v13_0_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); - amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block); - break; - case CHIP_CYAN_SKILLFISH: - amdgpu_device_ip_block_add(adev, &nv_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); - if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2) { - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v11_0_8_ip_block); - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - } - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v5_0_ip_block); - break; - default: - return -EINVAL; - } - - return 0; -} - static uint32_t nv_get_rev_id(struct amdgpu_device *adev) { return adev->nbio.funcs->get_rev_id(adev); @@ -1034,8 +732,10 @@ static int nv_common_early_init(void *handle) #define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE) struct amdgpu_device *adev = (struct amdgpu_device *)handle; - adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET; - adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET; + if (!amdgpu_sriov_vf(adev)) { + adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET; + adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET; + } adev->smc_rreg = NULL; adev->smc_wreg = NULL; adev->pcie_rreg = &nv_pcie_rreg; @@ -1056,8 +756,11 @@ static int nv_common_early_init(void *handle) adev->rev_id = nv_get_rev_id(adev); adev->external_rev_id = 0xff; - switch (adev->asic_type) { - case CHIP_NAVI10: + /* TODO: split the GC and PG flags based on the relevant IP version for which + * they are relevant. + */ + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(10, 1, 10): adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_IH_CG | @@ -1079,7 +782,7 @@ static int nv_common_early_init(void *handle) AMD_PG_SUPPORT_ATHUB; adev->external_rev_id = adev->rev_id + 0x1; break; - case CHIP_NAVI14: + case IP_VERSION(10, 1, 1): adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_IH_CG | @@ -1100,7 +803,7 @@ static int nv_common_early_init(void *handle) AMD_PG_SUPPORT_VCN_DPG; adev->external_rev_id = adev->rev_id + 20; break; - case CHIP_NAVI12: + case IP_VERSION(10, 1, 2): adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | AMD_CG_SUPPORT_GFX_CGCG | @@ -1129,7 +832,7 @@ static int nv_common_early_init(void *handle) adev->rev_id = 0; adev->external_rev_id = adev->rev_id + 0xa; break; - case CHIP_SIENNA_CICHLID: + case IP_VERSION(10, 3, 0): adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_GFX_CGLS | @@ -1153,7 +856,7 @@ static int nv_common_early_init(void *handle) } adev->external_rev_id = adev->rev_id + 0x28; break; - case CHIP_NAVY_FLOUNDER: + case IP_VERSION(10, 3, 2): adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_GFX_CGLS | @@ -1172,8 +875,7 @@ static int nv_common_early_init(void *handle) AMD_PG_SUPPORT_MMHUB; adev->external_rev_id = adev->rev_id + 0x32; break; - - case CHIP_VANGOGH: + case IP_VERSION(10, 3, 1): adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | AMD_CG_SUPPORT_GFX_CP_LS | @@ -1196,7 +898,7 @@ static int nv_common_early_init(void *handle) if (adev->apu_flags & AMD_APU_IS_VANGOGH) adev->external_rev_id = adev->rev_id + 0x01; break; - case CHIP_DIMGREY_CAVEFISH: + case IP_VERSION(10, 3, 4): adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_GFX_CGLS | @@ -1215,7 +917,7 @@ static int nv_common_early_init(void *handle) AMD_PG_SUPPORT_MMHUB; adev->external_rev_id = adev->rev_id + 0x3c; break; - case CHIP_BEIGE_GOBY: + case IP_VERSION(10, 3, 5): adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_GFX_CGLS | @@ -1232,7 +934,7 @@ static int nv_common_early_init(void *handle) AMD_PG_SUPPORT_MMHUB; adev->external_rev_id = adev->rev_id + 0x46; break; - case CHIP_YELLOW_CARP: + case IP_VERSION(10, 3, 3): adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | AMD_CG_SUPPORT_GFX_CGCG | @@ -1257,11 +959,11 @@ static int nv_common_early_init(void *handle) AMD_PG_SUPPORT_VCN_DPG | AMD_PG_SUPPORT_JPEG; if (adev->pdev->device == 0x1681) - adev->external_rev_id = adev->rev_id + 0x19; + adev->external_rev_id = 0x20; else adev->external_rev_id = adev->rev_id + 0x01; break; - case CHIP_CYAN_SKILLFISH: + case IP_VERSION(10, 1, 3): adev->cg_flags = 0; adev->pg_flags = 0; adev->external_rev_id = adev->rev_id + 0x82; @@ -1333,7 +1035,7 @@ static int nv_common_hw_init(void *handle) * for the purpose of expose those registers * to process space */ - if (adev->nbio.funcs->remap_hdp_registers) + if (adev->nbio.funcs->remap_hdp_registers && !amdgpu_sriov_vf(adev)) adev->nbio.funcs->remap_hdp_registers(adev); /* enable the doorbell aperture */ nv_enable_doorbell_aperture(adev, true); @@ -1388,14 +1090,14 @@ static int nv_common_set_clockgating_state(void *handle, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: + switch (adev->ip_versions[NBIO_HWIP][0]) { + case IP_VERSION(2, 3, 0): + case IP_VERSION(2, 3, 1): + case IP_VERSION(2, 3, 2): + case IP_VERSION(3, 3, 0): + case IP_VERSION(3, 3, 1): + case IP_VERSION(3, 3, 2): + case IP_VERSION(3, 3, 3): adev->nbio.funcs->update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); adev->nbio.funcs->update_medium_grain_light_sleep(adev, diff --git a/drivers/gpu/drm/amd/amdgpu/nv.h b/drivers/gpu/drm/amd/amdgpu/nv.h index 1f40ba3b0460..83e9782aef39 100644 --- a/drivers/gpu/drm/amd/amdgpu/nv.h +++ b/drivers/gpu/drm/amd/amdgpu/nv.h @@ -26,18 +26,10 @@ #include "nbio_v2_3.h" +extern const struct amdgpu_ip_block_version nv_common_ip_block; + void nv_grbm_select(struct amdgpu_device *adev, u32 me, u32 pipe, u32 queue, u32 vmid); void nv_set_virt_ops(struct amdgpu_device *adev); -int nv_set_ip_blocks(struct amdgpu_device *adev); -int navi10_reg_base_init(struct amdgpu_device *adev); -int navi14_reg_base_init(struct amdgpu_device *adev); -int navi12_reg_base_init(struct amdgpu_device *adev); -int sienna_cichlid_reg_base_init(struct amdgpu_device *adev); -void vangogh_reg_base_init(struct amdgpu_device *adev); -int dimgrey_cavefish_reg_base_init(struct amdgpu_device *adev); -int beige_goby_reg_base_init(struct amdgpu_device *adev); -int yellow_carp_reg_base_init(struct amdgpu_device *adev); -int cyan_skillfish_reg_base_init(struct amdgpu_device *adev); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c index 5872d68ed13d..ed2293686f0d 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c @@ -84,28 +84,28 @@ static int psp_v10_0_init_microcode(struct psp_context *psp) ta_hdr = (const struct ta_firmware_header_v1_0 *) adev->psp.ta_fw->data; - adev->psp.hdcp.feature_version = + adev->psp.hdcp_context.context.bin_desc.fw_version = le32_to_cpu(ta_hdr->hdcp.fw_version); - adev->psp.hdcp.size_bytes = + adev->psp.hdcp_context.context.bin_desc.size_bytes = le32_to_cpu(ta_hdr->hdcp.size_bytes); - adev->psp.hdcp.start_addr = + adev->psp.hdcp_context.context.bin_desc.start_addr = (uint8_t *)ta_hdr + le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes); - adev->psp.dtm.feature_version = + adev->psp.dtm_context.context.bin_desc.fw_version = le32_to_cpu(ta_hdr->dtm.fw_version); - adev->psp.dtm.size_bytes = + adev->psp.dtm_context.context.bin_desc.size_bytes = le32_to_cpu(ta_hdr->dtm.size_bytes); - adev->psp.dtm.start_addr = - (uint8_t *)adev->psp.hdcp.start_addr + + adev->psp.dtm_context.context.bin_desc.start_addr = + (uint8_t *)adev->psp.hdcp_context.context.bin_desc.start_addr + le32_to_cpu(ta_hdr->dtm.offset_bytes); - adev->psp.securedisplay.feature_version = + adev->psp.securedisplay_context.context.bin_desc.fw_version = le32_to_cpu(ta_hdr->securedisplay.fw_version); - adev->psp.securedisplay.size_bytes = + adev->psp.securedisplay_context.context.bin_desc.size_bytes = le32_to_cpu(ta_hdr->securedisplay.size_bytes); - adev->psp.securedisplay.start_addr = - (uint8_t *)adev->psp.hdcp.start_addr + + adev->psp.securedisplay_context.context.bin_desc.start_addr = + (uint8_t *)adev->psp.hdcp_context.context.bin_desc.start_addr + le32_to_cpu(ta_hdr->securedisplay.offset_bytes); adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version); diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c index 29bf9f09944b..d0e76b36d4ab 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c @@ -93,35 +93,35 @@ static int psp_v11_0_init_microcode(struct psp_context *psp) DRM_DEBUG("\n"); - switch (adev->asic_type) { - case CHIP_VEGA20: + switch (adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(11, 0, 2): chip_name = "vega20"; break; - case CHIP_NAVI10: + case IP_VERSION(11, 0, 0): chip_name = "navi10"; break; - case CHIP_NAVI14: + case IP_VERSION(11, 0, 5): chip_name = "navi14"; break; - case CHIP_NAVI12: + case IP_VERSION(11, 0, 9): chip_name = "navi12"; break; - case CHIP_ARCTURUS: + case IP_VERSION(11, 0, 4): chip_name = "arcturus"; break; - case CHIP_SIENNA_CICHLID: + case IP_VERSION(11, 0, 7): chip_name = "sienna_cichlid"; break; - case CHIP_NAVY_FLOUNDER: + case IP_VERSION(11, 0, 11): chip_name = "navy_flounder"; break; - case CHIP_VANGOGH: + case IP_VERSION(11, 5, 0): chip_name = "vangogh"; break; - case CHIP_DIMGREY_CAVEFISH: + case IP_VERSION(11, 0, 12): chip_name = "dimgrey_cavefish"; break; - case CHIP_BEIGE_GOBY: + case IP_VERSION(11, 0, 13): chip_name = "beige_goby"; break; default: @@ -129,9 +129,9 @@ static int psp_v11_0_init_microcode(struct psp_context *psp) } - switch (adev->asic_type) { - case CHIP_VEGA20: - case CHIP_ARCTURUS: + switch (adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 4): err = psp_init_sos_microcode(psp, chip_name); if (err) return err; @@ -151,20 +151,26 @@ static int psp_v11_0_init_microcode(struct psp_context *psp) goto out2; ta_hdr = (const struct ta_firmware_header_v1_0 *)adev->psp.ta_fw->data; - adev->psp.xgmi.feature_version = le32_to_cpu(ta_hdr->xgmi.fw_version); - adev->psp.xgmi.size_bytes = le32_to_cpu(ta_hdr->xgmi.size_bytes); - adev->psp.xgmi.start_addr = (uint8_t *)ta_hdr + + adev->psp.xgmi_context.context.bin_desc.fw_version = + le32_to_cpu(ta_hdr->xgmi.fw_version); + adev->psp.xgmi_context.context.bin_desc.size_bytes = + le32_to_cpu(ta_hdr->xgmi.size_bytes); + adev->psp.xgmi_context.context.bin_desc.start_addr = + (uint8_t *)ta_hdr + le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes); adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version); - adev->psp.ras.feature_version = le32_to_cpu(ta_hdr->ras.fw_version); - adev->psp.ras.size_bytes = le32_to_cpu(ta_hdr->ras.size_bytes); - adev->psp.ras.start_addr = (uint8_t *)adev->psp.xgmi.start_addr + + adev->psp.ras_context.context.bin_desc.fw_version = + le32_to_cpu(ta_hdr->ras.fw_version); + adev->psp.ras_context.context.bin_desc.size_bytes = + le32_to_cpu(ta_hdr->ras.size_bytes); + adev->psp.ras_context.context.bin_desc.start_addr = + (uint8_t *)adev->psp.xgmi_context.context.bin_desc.start_addr + le32_to_cpu(ta_hdr->ras.offset_bytes); } break; - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 5): + case IP_VERSION(11, 0, 9): err = psp_init_sos_microcode(psp, chip_name); if (err) return err; @@ -186,22 +192,31 @@ static int psp_v11_0_init_microcode(struct psp_context *psp) goto out2; ta_hdr = (const struct ta_firmware_header_v1_0 *)adev->psp.ta_fw->data; - adev->psp.hdcp.feature_version = le32_to_cpu(ta_hdr->hdcp.fw_version); - adev->psp.hdcp.size_bytes = le32_to_cpu(ta_hdr->hdcp.size_bytes); - adev->psp.hdcp.start_addr = (uint8_t *)ta_hdr + - le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes); + adev->psp.hdcp_context.context.bin_desc.fw_version = + le32_to_cpu(ta_hdr->hdcp.fw_version); + adev->psp.hdcp_context.context.bin_desc.size_bytes = + le32_to_cpu(ta_hdr->hdcp.size_bytes); + adev->psp.hdcp_context.context.bin_desc.start_addr = + (uint8_t *)ta_hdr + + le32_to_cpu( + ta_hdr->header.ucode_array_offset_bytes); adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version); - adev->psp.dtm.feature_version = le32_to_cpu(ta_hdr->dtm.fw_version); - adev->psp.dtm.size_bytes = le32_to_cpu(ta_hdr->dtm.size_bytes); - adev->psp.dtm.start_addr = (uint8_t *)adev->psp.hdcp.start_addr + + adev->psp.dtm_context.context.bin_desc.fw_version = + le32_to_cpu(ta_hdr->dtm.fw_version); + adev->psp.dtm_context.context.bin_desc.size_bytes = + le32_to_cpu(ta_hdr->dtm.size_bytes); + adev->psp.dtm_context.context.bin_desc.start_addr = + (uint8_t *)adev->psp.hdcp_context.context + .bin_desc.start_addr + le32_to_cpu(ta_hdr->dtm.offset_bytes); } break; - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: + case IP_VERSION(11, 0, 7): + case IP_VERSION(11, 0, 11): + case IP_VERSION(11, 0, 12): + case IP_VERSION(11, 0, 13): err = psp_init_sos_microcode(psp, chip_name); if (err) return err; @@ -209,15 +224,7 @@ static int psp_v11_0_init_microcode(struct psp_context *psp) if (err) return err; break; - case CHIP_BEIGE_GOBY: - err = psp_init_sos_microcode(psp, chip_name); - if (err) - return err; - err = psp_init_ta_microcode(psp, chip_name); - if (err) - return err; - break; - case CHIP_VANGOGH: + case IP_VERSION(11, 5, 0): err = psp_init_asd_microcode(psp, chip_name); if (err) return err; @@ -270,13 +277,15 @@ static bool psp_v11_0_is_sos_alive(struct psp_context *psp) return sol_reg != 0x0; } -static int psp_v11_0_bootloader_load_kdb(struct psp_context *psp) +static int psp_v11_0_bootloader_load_component(struct psp_context *psp, + struct psp_bin_desc *bin_desc, + enum psp_bootloader_cmd bl_cmd) { int ret; uint32_t psp_gfxdrv_command_reg = 0; struct amdgpu_device *adev = psp->adev; - /* Check tOS sign of life register to confirm sys driver and sOS + /* Check sOS sign of life register to confirm sys driver and sOS * are already been loaded. */ if (psp_v11_0_is_sos_alive(psp)) @@ -286,13 +295,13 @@ static int psp_v11_0_bootloader_load_kdb(struct psp_context *psp) if (ret) return ret; - /* Copy PSP KDB binary to memory */ - psp_copy_fw(psp, psp->kdb.start_addr, psp->kdb.size_bytes); + /* Copy PSP System Driver binary to memory */ + psp_copy_fw(psp, bin_desc->start_addr, bin_desc->size_bytes); - /* Provide the PSP KDB to bootloader */ + /* Provide the sys driver to bootloader */ WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_36, (uint32_t)(psp->fw_pri_mc_addr >> 20)); - psp_gfxdrv_command_reg = PSP_BL__LOAD_KEY_DATABASE; + psp_gfxdrv_command_reg = bl_cmd; WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35, psp_gfxdrv_command_reg); @@ -301,69 +310,19 @@ static int psp_v11_0_bootloader_load_kdb(struct psp_context *psp) return ret; } -static int psp_v11_0_bootloader_load_spl(struct psp_context *psp) +static int psp_v11_0_bootloader_load_kdb(struct psp_context *psp) { - int ret; - uint32_t psp_gfxdrv_command_reg = 0; - struct amdgpu_device *adev = psp->adev; - - /* Check tOS sign of life register to confirm sys driver and sOS - * are already been loaded. - */ - if (psp_v11_0_is_sos_alive(psp)) - return 0; - - ret = psp_v11_0_wait_for_bootloader(psp); - if (ret) - return ret; - - /* Copy PSP SPL binary to memory */ - psp_copy_fw(psp, psp->spl.start_addr, psp->spl.size_bytes); - - /* Provide the PSP SPL to bootloader */ - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_36, - (uint32_t)(psp->fw_pri_mc_addr >> 20)); - psp_gfxdrv_command_reg = PSP_BL__LOAD_TOS_SPL_TABLE; - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35, - psp_gfxdrv_command_reg); - - ret = psp_v11_0_wait_for_bootloader(psp); + return psp_v11_0_bootloader_load_component(psp, &psp->kdb, PSP_BL__LOAD_KEY_DATABASE); +} - return ret; +static int psp_v11_0_bootloader_load_spl(struct psp_context *psp) +{ + return psp_v11_0_bootloader_load_component(psp, &psp->spl, PSP_BL__LOAD_TOS_SPL_TABLE); } static int psp_v11_0_bootloader_load_sysdrv(struct psp_context *psp) { - int ret; - uint32_t psp_gfxdrv_command_reg = 0; - struct amdgpu_device *adev = psp->adev; - - /* Check sOS sign of life register to confirm sys driver and sOS - * are already been loaded. - */ - if (psp_v11_0_is_sos_alive(psp)) - return 0; - - ret = psp_v11_0_wait_for_bootloader(psp); - if (ret) - return ret; - - /* Copy PSP System Driver binary to memory */ - psp_copy_fw(psp, psp->sys.start_addr, psp->sys.size_bytes); - - /* Provide the sys driver to bootloader */ - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_36, - (uint32_t)(psp->fw_pri_mc_addr >> 20)); - psp_gfxdrv_command_reg = PSP_BL__LOAD_SYSDRV; - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35, - psp_gfxdrv_command_reg); - - /* there might be handshake issue with hardware which needs delay */ - mdelay(20); - - ret = psp_v11_0_wait_for_bootloader(psp); - - return ret; + return psp_v11_0_bootloader_load_component(psp, &psp->sys, PSP_BL__LOAD_SYSDRV); } static int psp_v11_0_bootloader_load_sos(struct psp_context *psp) @@ -691,7 +650,7 @@ static int psp_v11_0_memory_training(struct psp_context *psp, uint32_t ops) return -ENOMEM; } - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { memcpy_fromio(buf, adev->mman.aper_base_kaddr, sz); ret = psp_v11_0_memory_training_send_msg(psp, PSP_BL__DRAM_LONG_TRAIN); if (ret) { diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c index cc649406234b..a2588200ea58 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c @@ -84,22 +84,22 @@ static int psp_v12_0_init_microcode(struct psp_context *psp) ta_hdr = (const struct ta_firmware_header_v1_0 *) adev->psp.ta_fw->data; - adev->psp.hdcp.feature_version = + adev->psp.hdcp_context.context.bin_desc.fw_version = le32_to_cpu(ta_hdr->hdcp.fw_version); - adev->psp.hdcp.size_bytes = + adev->psp.hdcp_context.context.bin_desc.size_bytes = le32_to_cpu(ta_hdr->hdcp.size_bytes); - adev->psp.hdcp.start_addr = + adev->psp.hdcp_context.context.bin_desc.start_addr = (uint8_t *)ta_hdr + le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes); adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version); - adev->psp.dtm.feature_version = + adev->psp.dtm_context.context.bin_desc.fw_version = le32_to_cpu(ta_hdr->dtm.fw_version); - adev->psp.dtm.size_bytes = + adev->psp.dtm_context.context.bin_desc.size_bytes = le32_to_cpu(ta_hdr->dtm.size_bytes); - adev->psp.dtm.start_addr = - (uint8_t *)adev->psp.hdcp.start_addr + + adev->psp.dtm_context.context.bin_desc.start_addr = + (uint8_t *)adev->psp.hdcp_context.context.bin_desc.start_addr + le32_to_cpu(ta_hdr->dtm.offset_bytes); } diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c index 47a500f64db2..17655bc6d2f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c @@ -47,18 +47,19 @@ static int psp_v13_0_init_microcode(struct psp_context *psp) const char *chip_name; int err = 0; - switch (adev->asic_type) { - case CHIP_ALDEBARAN: + switch (adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(13, 0, 2): chip_name = "aldebaran"; break; - case CHIP_YELLOW_CARP: + case IP_VERSION(13, 0, 1): + case IP_VERSION(13, 0, 3): chip_name = "yellow_carp"; break; default: BUG(); } - switch (adev->asic_type) { - case CHIP_ALDEBARAN: + switch (adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(13, 0, 2): err = psp_init_sos_microcode(psp, chip_name); if (err) return err; @@ -66,7 +67,8 @@ static int psp_v13_0_init_microcode(struct psp_context *psp) if (err) return err; break; - case CHIP_YELLOW_CARP: + case IP_VERSION(13, 0, 1): + case IP_VERSION(13, 0, 3): err = psp_init_asd_microcode(psp, chip_name); if (err) return err; diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c index 8931000dcd41..e8e4749e9c79 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c @@ -469,8 +469,8 @@ static int sdma_v4_0_irq_id_to_seq(unsigned client_id) static void sdma_v4_0_init_golden_registers(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(4, 0, 0): soc15_program_register_sequence(adev, golden_settings_sdma_4, ARRAY_SIZE(golden_settings_sdma_4)); @@ -478,7 +478,7 @@ static void sdma_v4_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_sdma_vg10, ARRAY_SIZE(golden_settings_sdma_vg10)); break; - case CHIP_VEGA12: + case IP_VERSION(4, 0, 1): soc15_program_register_sequence(adev, golden_settings_sdma_4, ARRAY_SIZE(golden_settings_sdma_4)); @@ -486,7 +486,7 @@ static void sdma_v4_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_sdma_vg12, ARRAY_SIZE(golden_settings_sdma_vg12)); break; - case CHIP_VEGA20: + case IP_VERSION(4, 2, 0): soc15_program_register_sequence(adev, golden_settings_sdma0_4_2_init, ARRAY_SIZE(golden_settings_sdma0_4_2_init)); @@ -497,17 +497,18 @@ static void sdma_v4_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_sdma1_4_2, ARRAY_SIZE(golden_settings_sdma1_4_2)); break; - case CHIP_ARCTURUS: + case IP_VERSION(4, 2, 2): soc15_program_register_sequence(adev, golden_settings_sdma_arct, ARRAY_SIZE(golden_settings_sdma_arct)); break; - case CHIP_ALDEBARAN: + case IP_VERSION(4, 4, 0): soc15_program_register_sequence(adev, golden_settings_sdma_aldebaran, ARRAY_SIZE(golden_settings_sdma_aldebaran)); break; - case CHIP_RAVEN: + case IP_VERSION(4, 1, 0): + case IP_VERSION(4, 1, 1): soc15_program_register_sequence(adev, golden_settings_sdma_4_1, ARRAY_SIZE(golden_settings_sdma_4_1)); @@ -520,7 +521,7 @@ static void sdma_v4_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_sdma_rv1, ARRAY_SIZE(golden_settings_sdma_rv1)); break; - case CHIP_RENOIR: + case IP_VERSION(4, 1, 2): soc15_program_register_sequence(adev, golden_settings_sdma_4_3, ARRAY_SIZE(golden_settings_sdma_4_3)); @@ -538,12 +539,12 @@ static void sdma_v4_0_setup_ulv(struct amdgpu_device *adev) * The only chips with SDMAv4 and ULV are VG10 and VG20. * Server SKUs take a different hysteresis setting from other SKUs. */ - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(4, 0, 0): if (adev->pdev->device == 0x6860) break; return; - case CHIP_VEGA20: + case IP_VERSION(4, 2, 0): if (adev->pdev->device == 0x66a1) break; return; @@ -589,8 +590,8 @@ static void sdma_v4_0_destroy_inst_ctx(struct amdgpu_device *adev) /* arcturus shares the same FW memory across all SDMA isntances */ - if (adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_ALDEBARAN) + if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) || + adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 4, 0)) break; } @@ -620,17 +621,18 @@ static int sdma_v4_0_init_microcode(struct amdgpu_device *adev) DRM_DEBUG("\n"); - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(4, 0, 0): chip_name = "vega10"; break; - case CHIP_VEGA12: + case IP_VERSION(4, 0, 1): chip_name = "vega12"; break; - case CHIP_VEGA20: + case IP_VERSION(4, 2, 0): chip_name = "vega20"; break; - case CHIP_RAVEN: + case IP_VERSION(4, 1, 0): + case IP_VERSION(4, 1, 1): if (adev->apu_flags & AMD_APU_IS_RAVEN2) chip_name = "raven2"; else if (adev->apu_flags & AMD_APU_IS_PICASSO) @@ -638,16 +640,16 @@ static int sdma_v4_0_init_microcode(struct amdgpu_device *adev) else chip_name = "raven"; break; - case CHIP_ARCTURUS: + case IP_VERSION(4, 2, 2): chip_name = "arcturus"; break; - case CHIP_RENOIR: + case IP_VERSION(4, 1, 2): if (adev->apu_flags & AMD_APU_IS_RENOIR) chip_name = "renoir"; else chip_name = "green_sardine"; break; - case CHIP_ALDEBARAN: + case IP_VERSION(4, 4, 0): chip_name = "aldebaran"; break; default: @@ -665,8 +667,8 @@ static int sdma_v4_0_init_microcode(struct amdgpu_device *adev) goto out; for (i = 1; i < adev->sdma.num_instances; i++) { - if (adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_ALDEBARAN) { + if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) || + adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 4, 0)) { /* Acturus & Aldebaran will leverage the same FW memory for every SDMA instance */ memcpy((void *)&adev->sdma.instance[i], @@ -1106,7 +1108,7 @@ static void sdma_v4_0_ctx_switch_enable(struct amdgpu_device *adev, bool enable) * Arcturus for the moment and firmware version 14 * and above. */ - if (adev->asic_type == CHIP_ARCTURUS && + if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) && adev->sdma.instance[i].fw_version >= 14) WREG32_SDMA(i, mmSDMA0_PUB_DUMMY_REG2, enable); /* Extend page fault timeout to avoid interrupt storm */ @@ -1393,9 +1395,10 @@ static void sdma_v4_0_init_pg(struct amdgpu_device *adev) if (!(adev->pg_flags & AMD_PG_SUPPORT_SDMA)) return; - switch (adev->asic_type) { - case CHIP_RAVEN: - case CHIP_RENOIR: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(4, 1, 0): + case IP_VERSION(4, 1, 1): + case IP_VERSION(4, 1, 2): sdma_v4_1_init_power_gating(adev); sdma_v4_1_update_power_gating(adev, true); break; @@ -1835,13 +1838,13 @@ static bool sdma_v4_0_fw_support_paging_queue(struct amdgpu_device *adev) { uint fw_version = adev->sdma.instance[0].fw_version; - switch (adev->asic_type) { - case CHIP_VEGA10: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(4, 0, 0): return fw_version >= 430; - case CHIP_VEGA12: + case IP_VERSION(4, 0, 1): /*return fw_version >= 31;*/ return false; - case CHIP_VEGA20: + case IP_VERSION(4, 2, 0): return fw_version >= 123; default: return false; @@ -1853,15 +1856,6 @@ static int sdma_v4_0_early_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; int r; - if (adev->flags & AMD_IS_APU) - adev->sdma.num_instances = 1; - else if (adev->asic_type == CHIP_ARCTURUS) - adev->sdma.num_instances = 8; - else if (adev->asic_type == CHIP_ALDEBARAN) - adev->sdma.num_instances = 5; - else - adev->sdma.num_instances = 2; - r = sdma_v4_0_init_microcode(adev); if (r) { DRM_ERROR("Failed to load sdma firmware!\n"); @@ -1869,7 +1863,8 @@ static int sdma_v4_0_early_init(void *handle) } /* TODO: Page queue breaks driver reload under SRIOV */ - if ((adev->asic_type == CHIP_VEGA10) && amdgpu_sriov_vf((adev))) + if ((adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 0, 0)) && + amdgpu_sriov_vf((adev))) adev->sdma.has_page_queue = false; else if (sdma_v4_0_fw_support_paging_queue(adev)) adev->sdma.has_page_queue = true; @@ -2141,14 +2136,14 @@ static int sdma_v4_0_process_trap_irq(struct amdgpu_device *adev, amdgpu_fence_process(&adev->sdma.instance[instance].ring); break; case 1: - if (adev->asic_type == CHIP_VEGA20) + if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 0)) amdgpu_fence_process(&adev->sdma.instance[instance].page); break; case 2: /* XXX compute */ break; case 3: - if (adev->asic_type != CHIP_VEGA20) + if (adev->ip_versions[SDMA0_HWIP][0] != IP_VERSION(4, 2, 0)) amdgpu_fence_process(&adev->sdma.instance[instance].page); break; } @@ -2364,9 +2359,10 @@ static int sdma_v4_0_set_powergating_state(void *handle, { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - switch (adev->asic_type) { - case CHIP_RAVEN: - case CHIP_RENOIR: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(4, 1, 0): + case IP_VERSION(4, 1, 1): + case IP_VERSION(4, 1, 2): sdma_v4_1_update_power_gating(adev, state == AMD_PG_STATE_GATE); break; @@ -2551,7 +2547,7 @@ static void sdma_v4_0_set_ring_funcs(struct amdgpu_device *adev) int i; for (i = 0; i < adev->sdma.num_instances; i++) { - if (adev->asic_type == CHIP_ARCTURUS && i >= 5) + if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) && i >= 5) adev->sdma.instance[i].ring.funcs = &sdma_v4_0_ring_funcs_2nd_mmhub; else @@ -2559,7 +2555,7 @@ static void sdma_v4_0_set_ring_funcs(struct amdgpu_device *adev) &sdma_v4_0_ring_funcs; adev->sdma.instance[i].ring.me = i; if (adev->sdma.has_page_queue) { - if (adev->asic_type == CHIP_ARCTURUS && i >= 5) + if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) && i >= 5) adev->sdma.instance[i].page.funcs = &sdma_v4_0_page_ring_funcs_2nd_mmhub; else @@ -2786,12 +2782,12 @@ static const struct amdgpu_sdma_ras_funcs sdma_v4_0_ras_funcs = { static void sdma_v4_0_set_ras_funcs(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_VEGA20: - case CHIP_ARCTURUS: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(4, 2, 0): + case IP_VERSION(4, 2, 2): adev->sdma.funcs = &sdma_v4_0_ras_funcs; break; - case CHIP_ALDEBARAN: + case IP_VERSION(4, 4, 0): adev->sdma.funcs = &sdma_v4_4_ras_funcs; break; default: diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c index 50bf3b71bc93..81e033549dda 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c @@ -187,8 +187,8 @@ static u32 sdma_v5_0_get_reg_offset(struct amdgpu_device *adev, u32 instance, u3 static void sdma_v5_0_init_golden_registers(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_NAVI10: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(5, 0, 0): soc15_program_register_sequence(adev, golden_settings_sdma_5, (const u32)ARRAY_SIZE(golden_settings_sdma_5)); @@ -196,7 +196,7 @@ static void sdma_v5_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_sdma_nv10, (const u32)ARRAY_SIZE(golden_settings_sdma_nv10)); break; - case CHIP_NAVI14: + case IP_VERSION(5, 0, 2): soc15_program_register_sequence(adev, golden_settings_sdma_5, (const u32)ARRAY_SIZE(golden_settings_sdma_5)); @@ -204,7 +204,7 @@ static void sdma_v5_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_sdma_nv14, (const u32)ARRAY_SIZE(golden_settings_sdma_nv14)); break; - case CHIP_NAVI12: + case IP_VERSION(5, 0, 5): if (amdgpu_sriov_vf(adev)) soc15_program_register_sequence(adev, golden_settings_sdma_5_sriov, @@ -217,7 +217,7 @@ static void sdma_v5_0_init_golden_registers(struct amdgpu_device *adev) golden_settings_sdma_nv12, (const u32)ARRAY_SIZE(golden_settings_sdma_nv12)); break; - case CHIP_CYAN_SKILLFISH: + case IP_VERSION(5, 0, 1): soc15_program_register_sequence(adev, golden_settings_sdma_cyan_skillfish, (const u32)ARRAY_SIZE(golden_settings_sdma_cyan_skillfish)); @@ -248,22 +248,22 @@ static int sdma_v5_0_init_microcode(struct amdgpu_device *adev) const struct common_firmware_header *header = NULL; const struct sdma_firmware_header_v1_0 *hdr; - if (amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_NAVI12)) + if (amdgpu_sriov_vf(adev) && (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(5, 0, 5))) return 0; DRM_DEBUG("\n"); - switch (adev->asic_type) { - case CHIP_NAVI10: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(5, 0, 0): chip_name = "navi10"; break; - case CHIP_NAVI14: + case IP_VERSION(5, 0, 2): chip_name = "navi14"; break; - case CHIP_NAVI12: + case IP_VERSION(5, 0, 5): chip_name = "navi12"; break; - case CHIP_CYAN_SKILLFISH: + case IP_VERSION(5, 0, 1): if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2) chip_name = "cyan_skillfish2"; else @@ -481,8 +481,6 @@ static void sdma_v5_0_ring_emit_ib(struct amdgpu_ring *ring, * sdma_v5_0_ring_emit_mem_sync - flush the IB by graphics cache rinse * * @ring: amdgpu ring pointer - * @job: job to retrieve vmid from - * @ib: IB object to schedule * * flush the IB by graphics cache rinse. */ @@ -1295,8 +1293,6 @@ static int sdma_v5_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - adev->sdma.num_instances = 2; - sdma_v5_0_set_ring_funcs(adev); sdma_v5_0_set_buffer_funcs(adev); sdma_v5_0_set_vm_pte_funcs(adev); @@ -1636,10 +1632,10 @@ static int sdma_v5_0_set_clockgating_state(void *handle, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_NAVI10: - case CHIP_NAVI14: - case CHIP_NAVI12: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(5, 0, 0): + case IP_VERSION(5, 0, 2): + case IP_VERSION(5, 0, 5): sdma_v5_0_update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); sdma_v5_0_update_medium_grain_light_sleep(adev, diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c index e32efcfb0c8b..d3d6d5b045b8 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c @@ -136,23 +136,23 @@ static int sdma_v5_2_init_microcode(struct amdgpu_device *adev) DRM_DEBUG("\n"); - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(5, 2, 0): chip_name = "sienna_cichlid"; break; - case CHIP_NAVY_FLOUNDER: + case IP_VERSION(5, 2, 2): chip_name = "navy_flounder"; break; - case CHIP_VANGOGH: + case IP_VERSION(5, 2, 1): chip_name = "vangogh"; break; - case CHIP_DIMGREY_CAVEFISH: + case IP_VERSION(5, 2, 4): chip_name = "dimgrey_cavefish"; break; - case CHIP_BEIGE_GOBY: + case IP_VERSION(5, 2, 5): chip_name = "beige_goby"; break; - case CHIP_YELLOW_CARP: + case IP_VERSION(5, 2, 3): chip_name = "yellow_carp"; break; default: @@ -174,7 +174,7 @@ static int sdma_v5_2_init_microcode(struct amdgpu_device *adev) (void *)&adev->sdma.instance[0], sizeof(struct amdgpu_sdma_instance)); - if (amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_SIENNA_CICHLID)) + if (amdgpu_sriov_vf(adev) && (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(5, 2, 0))) return 0; DRM_DEBUG("psp_load == '%s'\n", @@ -368,17 +368,15 @@ static void sdma_v5_2_ring_emit_ib(struct amdgpu_ring *ring, * sdma_v5_2_ring_emit_mem_sync - flush the IB by graphics cache rinse * * @ring: amdgpu ring pointer - * @job: job to retrieve vmid from - * @ib: IB object to schedule * * flush the IB by graphics cache rinse. */ static void sdma_v5_2_ring_emit_mem_sync(struct amdgpu_ring *ring) { - uint32_t gcr_cntl = - SDMA_GCR_GL2_INV | SDMA_GCR_GL2_WB | SDMA_GCR_GLM_INV | - SDMA_GCR_GL1_INV | SDMA_GCR_GLV_INV | SDMA_GCR_GLK_INV | - SDMA_GCR_GLI_INV(1); + uint32_t gcr_cntl = SDMA_GCR_GL2_INV | SDMA_GCR_GL2_WB | + SDMA_GCR_GLM_INV | SDMA_GCR_GL1_INV | + SDMA_GCR_GLV_INV | SDMA_GCR_GLK_INV | + SDMA_GCR_GLI_INV(1); /* flush entire cache L0/L1/L2, this can be optimized by performance requirement */ amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_GCR_REQ)); @@ -544,9 +542,6 @@ static void sdma_v5_2_ctx_switch_enable(struct amdgpu_device *adev, bool enable) } for (i = 0; i < adev->sdma.num_instances; i++) { - f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL)); - f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL, - AUTO_CTXSW_ENABLE, enable ? 1 : 0); if (enable && amdgpu_sdma_phase_quantum) { WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_PHASE0_QUANTUM), phase_quantum); @@ -555,7 +550,13 @@ static void sdma_v5_2_ctx_switch_enable(struct amdgpu_device *adev, bool enable) WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_PHASE2_QUANTUM), phase_quantum); } - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), f32_cntl); + + if (!amdgpu_sriov_vf(adev)) { + f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL)); + f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL, + AUTO_CTXSW_ENABLE, enable ? 1 : 0); + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), f32_cntl); + } } } @@ -578,10 +579,12 @@ static void sdma_v5_2_enable(struct amdgpu_device *adev, bool enable) sdma_v5_2_rlc_stop(adev); } - for (i = 0; i < adev->sdma.num_instances; i++) { - f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL)); - f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, enable ? 0 : 1); - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), f32_cntl); + if (!amdgpu_sriov_vf(adev)) { + for (i = 0; i < adev->sdma.num_instances; i++) { + f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL)); + f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, enable ? 0 : 1); + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), f32_cntl); + } } } @@ -610,7 +613,8 @@ static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev) ring = &adev->sdma.instance[i].ring; wb_offset = (ring->rptr_offs * 4); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0); + if (!amdgpu_sriov_vf(adev)) + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0); /* Set ring buffer size in dwords */ rb_bufsz = order_base_2(ring->ring_size / 4); @@ -685,32 +689,34 @@ static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev) sdma_v5_2_ring_set_wptr(ring); /* set minor_ptr_update to 0 after wptr programed */ - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0); - /* set utc l1 enable flag always to 1 */ - temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL)); - temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1); - - /* enable MCBP */ - temp = REG_SET_FIELD(temp, SDMA0_CNTL, MIDCMD_PREEMPT_ENABLE, 1); - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), temp); - - /* Set up RESP_MODE to non-copy addresses */ - temp = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL)); - temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3); - temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL), temp); - - /* program default cache read and write policy */ - temp = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE)); - /* clean read policy and write policy bits */ - temp &= 0xFF0FFF; - temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) | - (CACHE_WRITE_POLICY_L2__DEFAULT << 14) | - SDMA0_UTCL1_PAGE__LLC_NOALLOC_MASK); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE), temp); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0); + /* SRIOV VF has no control of any of registers below */ if (!amdgpu_sriov_vf(adev)) { + /* set utc l1 enable flag always to 1 */ + temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL)); + temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1); + + /* enable MCBP */ + temp = REG_SET_FIELD(temp, SDMA0_CNTL, MIDCMD_PREEMPT_ENABLE, 1); + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), temp); + + /* Set up RESP_MODE to non-copy addresses */ + temp = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL)); + temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3); + temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL), temp); + + /* program default cache read and write policy */ + temp = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE)); + /* clean read policy and write policy bits */ + temp &= 0xFF0FFF; + temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) | + (CACHE_WRITE_POLICY_L2__DEFAULT << 14) | + SDMA0_UTCL1_PAGE__LLC_NOALLOC_MASK); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE), temp); + /* unhalt engine */ temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL)); temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0); @@ -1217,23 +1223,6 @@ static int sdma_v5_2_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - adev->sdma.num_instances = 4; - break; - case CHIP_NAVY_FLOUNDER: - case CHIP_DIMGREY_CAVEFISH: - adev->sdma.num_instances = 2; - break; - case CHIP_VANGOGH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: - adev->sdma.num_instances = 1; - break; - default: - break; - } - sdma_v5_2_set_ring_funcs(adev); sdma_v5_2_set_buffer_funcs(adev); sdma_v5_2_set_vm_pte_funcs(adev); @@ -1455,13 +1444,14 @@ static int sdma_v5_2_set_trap_irq_state(struct amdgpu_device *adev, enum amdgpu_interrupt_state state) { u32 sdma_cntl; - u32 reg_offset = sdma_v5_2_get_reg_offset(adev, type, mmSDMA0_CNTL); - sdma_cntl = RREG32(reg_offset); - sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA0_CNTL, TRAP_ENABLE, - state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0); - WREG32(reg_offset, sdma_cntl); + if (!amdgpu_sriov_vf(adev)) { + sdma_cntl = RREG32(reg_offset); + sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA0_CNTL, TRAP_ENABLE, + state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0); + WREG32(reg_offset, sdma_cntl); + } return 0; } @@ -1555,7 +1545,7 @@ static void sdma_v5_2_update_medium_grain_clock_gating(struct amdgpu_device *ade for (i = 0; i < adev->sdma.num_instances; i++) { - if (adev->sdma.instance[i].fw_version < 70 && adev->asic_type == CHIP_VANGOGH) + if (adev->sdma.instance[i].fw_version < 70 && adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(5, 2, 1)) adev->cg_flags &= ~AMD_CG_SUPPORT_SDMA_MGCG; if (enable && (adev->cg_flags & AMD_CG_SUPPORT_SDMA_MGCG)) { @@ -1592,7 +1582,7 @@ static void sdma_v5_2_update_medium_grain_light_sleep(struct amdgpu_device *adev for (i = 0; i < adev->sdma.num_instances; i++) { - if (adev->sdma.instance[i].fw_version < 70 && adev->asic_type == CHIP_VANGOGH) + if (adev->sdma.instance[i].fw_version < 70 && adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(5, 2, 1)) adev->cg_flags &= ~AMD_CG_SUPPORT_SDMA_LS; if (enable && (adev->cg_flags & AMD_CG_SUPPORT_SDMA_LS)) { @@ -1621,13 +1611,13 @@ static int sdma_v5_2_set_clockgating_state(void *handle, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_SIENNA_CICHLID: - case CHIP_NAVY_FLOUNDER: - case CHIP_VANGOGH: - case CHIP_DIMGREY_CAVEFISH: - case CHIP_BEIGE_GOBY: - case CHIP_YELLOW_CARP: + switch (adev->ip_versions[SDMA0_HWIP][0]) { + case IP_VERSION(5, 2, 0): + case IP_VERSION(5, 2, 2): + case IP_VERSION(5, 2, 1): + case IP_VERSION(5, 2, 4): + case IP_VERSION(5, 2, 5): + case IP_VERSION(5, 2, 3): sdma_v5_2_update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); sdma_v5_2_update_medium_grain_light_sleep(adev, diff --git a/drivers/gpu/drm/amd/amdgpu/sienna_cichlid_reg_init.c b/drivers/gpu/drm/amd/amdgpu/sienna_cichlid_reg_init.c deleted file mode 100644 index 5ee69f70c49b..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/sienna_cichlid_reg_init.c +++ /dev/null @@ -1,54 +0,0 @@ -/* - * Copyright 2019 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ -#include "amdgpu.h" -#include "nv.h" - -#include "soc15_common.h" -#include "soc15_hw_ip.h" -#include "sienna_cichlid_ip_offset.h" - -int sienna_cichlid_reg_base_init(struct amdgpu_device *adev) -{ - /* HW has more IP blocks, only initialized the blocke needed by driver */ - uint32_t i; - for (i = 0 ; i < MAX_INSTANCE ; ++i) { - adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i])); - adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i])); - adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i])); - adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIO_BASE.instance[i])); - adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i])); - adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i])); - adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(VCN_BASE.instance[i])); - adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i])); - adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DCN_BASE.instance[i])); - adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i])); - adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA1_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA2_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SDMA3_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i])); - adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i])); - } - return 0; -} diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 0fc97c364fd7..0fc1747e4a70 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -85,6 +85,8 @@ #define mmMP0_MISC_LIGHT_SLEEP_CTRL 0x01ba #define mmMP0_MISC_LIGHT_SLEEP_CTRL_BASE_IDX 0 +static const struct amd_ip_funcs soc15_common_ip_funcs; + /* Vega, Raven, Arcturus */ static const struct amdgpu_video_codec_info vega_video_codecs_encode_array[] = { @@ -154,31 +156,38 @@ static const struct amdgpu_video_codecs rn_video_codecs_decode = static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode, const struct amdgpu_video_codecs **codecs) { - switch (adev->asic_type) { - case CHIP_VEGA20: - case CHIP_VEGA10: - case CHIP_VEGA12: - if (encode) - *codecs = &vega_video_codecs_encode; - else - *codecs = &vega_video_codecs_decode; - return 0; - case CHIP_RAVEN: - if (encode) - *codecs = &vega_video_codecs_encode; - else - *codecs = &rv_video_codecs_decode; - return 0; - case CHIP_ARCTURUS: - case CHIP_ALDEBARAN: - case CHIP_RENOIR: - if (encode) - *codecs = &vega_video_codecs_encode; - else - *codecs = &rn_video_codecs_decode; - return 0; - default: - return -EINVAL; + if (adev->ip_versions[VCE_HWIP][0]) { + switch (adev->ip_versions[VCE_HWIP][0]) { + case IP_VERSION(4, 0, 0): + case IP_VERSION(4, 1, 0): + if (encode) + *codecs = &vega_video_codecs_encode; + else + *codecs = &vega_video_codecs_decode; + return 0; + default: + return -EINVAL; + } + } else { + switch (adev->ip_versions[UVD_HWIP][0]) { + case IP_VERSION(1, 0, 0): + case IP_VERSION(1, 0, 1): + if (encode) + *codecs = &vega_video_codecs_encode; + else + *codecs = &rv_video_codecs_decode; + return 0; + case IP_VERSION(2, 5, 0): + case IP_VERSION(2, 6, 0): + case IP_VERSION(2, 2, 0): + if (encode) + *codecs = &vega_video_codecs_encode; + else + *codecs = &rn_video_codecs_decode; + return 0; + default: + return -EINVAL; + } } } @@ -332,9 +341,11 @@ static u32 soc15_get_xclk(struct amdgpu_device *adev) { u32 reference_clock = adev->clock.spll.reference_freq; - if (adev->asic_type == CHIP_RENOIR) + if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 0) || + adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 1)) return 10000; - if (adev->asic_type == CHIP_RAVEN) + if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 0) || + adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 1)) return reference_clock / 4; return reference_clock; @@ -565,28 +576,29 @@ soc15_asic_reset_method(struct amdgpu_device *adev) dev_warn(adev->dev, "Specified reset method:%d isn't supported, using AUTO instead.\n", amdgpu_reset_method); - switch (adev->asic_type) { - case CHIP_RAVEN: - case CHIP_RENOIR: + switch (adev->ip_versions[MP1_HWIP][0]) { + case IP_VERSION(10, 0, 0): + case IP_VERSION(10, 0, 1): + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): return AMD_RESET_METHOD_MODE2; - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_ARCTURUS: - baco_reset = amdgpu_dpm_is_baco_supported(adev); - break; - case CHIP_VEGA20: - if (adev->psp.sos.fw_version >= 0x80067) + case IP_VERSION(9, 0, 0): + case IP_VERSION(11, 0, 2): + if (adev->asic_type == CHIP_VEGA20) { + if (adev->psp.sos.fw_version >= 0x80067) + baco_reset = amdgpu_dpm_is_baco_supported(adev); + /* + * 1. PMFW version > 0x284300: all cases use baco + * 2. PMFW version <= 0x284300: only sGPU w/o RAS use baco + */ + if (ras && adev->ras_enabled && + adev->pm.fw_version <= 0x283400) + baco_reset = false; + } else { baco_reset = amdgpu_dpm_is_baco_supported(adev); - - /* - * 1. PMFW version > 0x284300: all cases use baco - * 2. PMFW version <= 0x284300: only sGPU w/o RAS use baco - */ - if (ras && adev->ras_enabled && - adev->pm.fw_version <= 0x283400) - baco_reset = false; + } break; - case CHIP_ALDEBARAN: + case IP_VERSION(13, 0, 2): /* * 1.connected to cpu: driver issue mode2 reset * 2.discret gpu: driver issue mode1 reset @@ -629,15 +641,17 @@ static int soc15_asic_reset(struct amdgpu_device *adev) static bool soc15_supports_baco(struct amdgpu_device *adev) { - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_ARCTURUS: - return amdgpu_dpm_is_baco_supported(adev); - case CHIP_VEGA20: - if (adev->psp.sos.fw_version >= 0x80067) + switch (adev->ip_versions[MP1_HWIP][0]) { + case IP_VERSION(9, 0, 0): + case IP_VERSION(11, 0, 2): + if (adev->asic_type == CHIP_VEGA20) { + if (adev->psp.sos.fw_version >= 0x80067) + return amdgpu_dpm_is_baco_supported(adev); + return false; + } else { return amdgpu_dpm_is_baco_supported(adev); - return false; + } + break; default: return false; } @@ -704,7 +718,7 @@ static void soc15_enable_doorbell_aperture(struct amdgpu_device *adev, adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, enable); } -static const struct amdgpu_ip_block_version vega10_common_ip_block = +const struct amdgpu_ip_block_version vega10_common_ip_block = { .type = AMD_IP_BLOCK_TYPE_COMMON, .major = 2, @@ -730,7 +744,7 @@ static void soc15_reg_base_init(struct amdgpu_device *adev) vega10_reg_base_init(adev); break; case CHIP_RENOIR: - /* It's safe to do ip discovery here for Renior, + /* It's safe to do ip discovery here for Renoir, * it doesn't support SRIOV. */ if (amdgpu_discovery) { r = amdgpu_discovery_reg_base_init(adev); @@ -766,185 +780,6 @@ void soc15_set_virt_ops(struct amdgpu_device *adev) soc15_reg_base_init(adev); } -int soc15_set_ip_blocks(struct amdgpu_device *adev) -{ - /* for bare metal case */ - if (!amdgpu_sriov_vf(adev)) - soc15_reg_base_init(adev); - - if (adev->flags & AMD_IS_APU) { - adev->nbio.funcs = &nbio_v7_0_funcs; - adev->nbio.hdp_flush_reg = &nbio_v7_0_hdp_flush_reg; - } else if (adev->asic_type == CHIP_VEGA20 || - adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_ALDEBARAN) { - adev->nbio.funcs = &nbio_v7_4_funcs; - adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg; - } else { - adev->nbio.funcs = &nbio_v6_1_funcs; - adev->nbio.hdp_flush_reg = &nbio_v6_1_hdp_flush_reg; - } - adev->hdp.funcs = &hdp_v4_0_funcs; - - if (adev->asic_type == CHIP_VEGA20 || - adev->asic_type == CHIP_ARCTURUS || - adev->asic_type == CHIP_ALDEBARAN) - adev->df.funcs = &df_v3_6_funcs; - else - adev->df.funcs = &df_v1_7_funcs; - - if (adev->asic_type == CHIP_VEGA20 || - adev->asic_type == CHIP_ARCTURUS) - adev->smuio.funcs = &smuio_v11_0_funcs; - else if (adev->asic_type == CHIP_ALDEBARAN) - adev->smuio.funcs = &smuio_v13_0_funcs; - else - adev->smuio.funcs = &smuio_v9_0_funcs; - - adev->rev_id = soc15_get_rev_id(adev); - - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: - amdgpu_device_ip_block_add(adev, &vega10_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); - - /* For Vega10 SR-IOV, PSP need to be initialized before IH */ - if (amdgpu_sriov_vf(adev)) { - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) { - if (adev->asic_type == CHIP_VEGA20) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - else - amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block); - } - if (adev->asic_type == CHIP_VEGA20) - amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block); - else - amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block); - } else { - if (adev->asic_type == CHIP_VEGA20) - amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block); - else - amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) { - if (adev->asic_type == CHIP_VEGA20) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - else - amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block); - } - } - amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); - if (is_support_sw_smu(adev)) { - if (!amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - } else { - amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); - } - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - if (!(adev->asic_type == CHIP_VEGA20 && amdgpu_sriov_vf(adev))) { - amdgpu_device_ip_block_add(adev, &uvd_v7_0_ip_block); - amdgpu_device_ip_block_add(adev, &vce_v4_0_ip_block); - } - break; - case CHIP_RAVEN: - amdgpu_device_ip_block_add(adev, &vega10_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); - amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); - amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - amdgpu_device_ip_block_add(adev, &vcn_v1_0_ip_block); - break; - case CHIP_ARCTURUS: - amdgpu_device_ip_block_add(adev, &vega10_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); - - if (amdgpu_sriov_vf(adev)) { - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block); - } else { - amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - } - - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); - amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); - - if (amdgpu_sriov_vf(adev)) { - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &vcn_v2_5_ip_block); - } else { - amdgpu_device_ip_block_add(adev, &vcn_v2_5_ip_block); - } - if (!amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &jpeg_v2_5_ip_block); - break; - case CHIP_RENOIR: - amdgpu_device_ip_block_add(adev, &vega10_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); - amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v12_0_ip_block); - amdgpu_device_ip_block_add(adev, &smu_v12_0_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); - if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) - amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); -#if defined(CONFIG_DRM_AMD_DC) - else if (amdgpu_device_has_dc_support(adev)) - amdgpu_device_ip_block_add(adev, &dm_ip_block); -#endif - amdgpu_device_ip_block_add(adev, &vcn_v2_0_ip_block); - amdgpu_device_ip_block_add(adev, &jpeg_v2_0_ip_block); - break; - case CHIP_ALDEBARAN: - amdgpu_device_ip_block_add(adev, &vega10_common_ip_block); - amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); - - if (amdgpu_sriov_vf(adev)) { - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v13_0_ip_block); - amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block); - } else { - amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block); - if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) - amdgpu_device_ip_block_add(adev, &psp_v13_0_ip_block); - } - - amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); - - amdgpu_device_ip_block_add(adev, &smu_v13_0_ip_block); - amdgpu_device_ip_block_add(adev, &vcn_v2_6_ip_block); - amdgpu_device_ip_block_add(adev, &jpeg_v2_6_ip_block); - break; - default: - return -EINVAL; - } - - return 0; -} - static bool soc15_need_full_reset(struct amdgpu_device *adev) { /* change this when we implement soft reset */ @@ -1136,8 +971,10 @@ static int soc15_common_early_init(void *handle) #define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE) struct amdgpu_device *adev = (struct amdgpu_device *)handle; - adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET; - adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET; + if (!amdgpu_sriov_vf(adev)) { + adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET; + adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET; + } adev->smc_rreg = NULL; adev->smc_wreg = NULL; adev->pcie_rreg = &soc15_pcie_rreg; @@ -1153,10 +990,13 @@ static int soc15_common_early_init(void *handle) adev->se_cac_rreg = &soc15_se_cac_rreg; adev->se_cac_wreg = &soc15_se_cac_wreg; - + adev->rev_id = soc15_get_rev_id(adev); adev->external_rev_id = 0xFF; - switch (adev->asic_type) { - case CHIP_VEGA10: + /* TODO: split the GC and PG flags based on the relevant IP version for which + * they are relevant. + */ + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 0, 1): adev->asic_funcs = &soc15_asic_funcs; adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | @@ -1180,7 +1020,7 @@ static int soc15_common_early_init(void *handle) adev->pg_flags = 0; adev->external_rev_id = 0x1; break; - case CHIP_VEGA12: + case IP_VERSION(9, 2, 1): adev->asic_funcs = &soc15_asic_funcs; adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | @@ -1203,7 +1043,7 @@ static int soc15_common_early_init(void *handle) adev->pg_flags = 0; adev->external_rev_id = adev->rev_id + 0x14; break; - case CHIP_VEGA20: + case IP_VERSION(9, 4, 0): adev->asic_funcs = &vega20_asic_funcs; adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | @@ -1226,7 +1066,8 @@ static int soc15_common_early_init(void *handle) adev->pg_flags = 0; adev->external_rev_id = adev->rev_id + 0x28; break; - case CHIP_RAVEN: + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 2): adev->asic_funcs = &soc15_asic_funcs; if (adev->rev_id >= 0x8) @@ -1299,7 +1140,7 @@ static int soc15_common_early_init(void *handle) adev->pg_flags = AMD_PG_SUPPORT_SDMA | AMD_PG_SUPPORT_VCN; } break; - case CHIP_ARCTURUS: + case IP_VERSION(9, 4, 1): adev->asic_funcs = &vega20_asic_funcs; adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | @@ -1318,7 +1159,7 @@ static int soc15_common_early_init(void *handle) adev->pg_flags = AMD_PG_SUPPORT_VCN | AMD_PG_SUPPORT_VCN_DPG; adev->external_rev_id = adev->rev_id + 0x32; break; - case CHIP_RENOIR: + case IP_VERSION(9, 3, 0): adev->asic_funcs = &soc15_asic_funcs; if (adev->apu_flags & AMD_APU_IS_RENOIR) @@ -1349,7 +1190,7 @@ static int soc15_common_early_init(void *handle) AMD_PG_SUPPORT_JPEG | AMD_PG_SUPPORT_VCN_DPG; break; - case CHIP_ALDEBARAN: + case IP_VERSION(9, 4, 2): adev->asic_funcs = &vega20_asic_funcs; adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | @@ -1397,7 +1238,9 @@ static int soc15_common_sw_init(void *handle) if (amdgpu_sriov_vf(adev)) xgpu_ai_mailbox_add_irq_id(adev); - adev->df.funcs->sw_init(adev); + if (adev->df.funcs && + adev->df.funcs->sw_init) + adev->df.funcs->sw_init(adev); return 0; } @@ -1409,7 +1252,10 @@ static int soc15_common_sw_fini(void *handle) if (adev->nbio.ras_funcs && adev->nbio.ras_funcs->ras_fini) adev->nbio.ras_funcs->ras_fini(adev); - adev->df.funcs->sw_fini(adev); + + if (adev->df.funcs && + adev->df.funcs->sw_fini) + adev->df.funcs->sw_fini(adev); return 0; } @@ -1446,7 +1292,7 @@ static int soc15_common_hw_init(void *handle) * for the purpose of expose those registers * to process space */ - if (adev->nbio.funcs->remap_hdp_registers) + if (adev->nbio.funcs->remap_hdp_registers && !amdgpu_sriov_vf(adev)) adev->nbio.funcs->remap_hdp_registers(adev); /* enable the doorbell aperture */ @@ -1564,10 +1410,10 @@ static int soc15_common_set_clockgating_state(void *handle, if (amdgpu_sriov_vf(adev)) return 0; - switch (adev->asic_type) { - case CHIP_VEGA10: - case CHIP_VEGA12: - case CHIP_VEGA20: + switch (adev->ip_versions[NBIO_HWIP][0]) { + case IP_VERSION(6, 1, 0): + case IP_VERSION(6, 2, 0): + case IP_VERSION(7, 4, 0): adev->nbio.funcs->update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); adev->nbio.funcs->update_medium_grain_light_sleep(adev, @@ -1583,8 +1429,9 @@ static int soc15_common_set_clockgating_state(void *handle, adev->df.funcs->update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); break; - case CHIP_RAVEN: - case CHIP_RENOIR: + case IP_VERSION(7, 0, 0): + case IP_VERSION(7, 0, 1): + case IP_VERSION(2, 5, 0): adev->nbio.funcs->update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE); adev->nbio.funcs->update_medium_grain_light_sleep(adev, @@ -1596,8 +1443,8 @@ static int soc15_common_set_clockgating_state(void *handle, soc15_update_drm_light_sleep(adev, state == AMD_CG_STATE_GATE); break; - case CHIP_ARCTURUS: - case CHIP_ALDEBARAN: + case IP_VERSION(7, 4, 1): + case IP_VERSION(7, 4, 4): adev->hdp.funcs->update_clock_gating(adev, state == AMD_CG_STATE_GATE); break; @@ -1619,7 +1466,7 @@ static void soc15_common_get_clockgating_state(void *handle, u32 *flags) adev->hdp.funcs->get_clock_gating_state(adev, flags); - if (adev->asic_type != CHIP_ALDEBARAN) { + if (adev->ip_versions[MP0_HWIP][0] != IP_VERSION(13, 0, 2)) { /* AMD_CG_SUPPORT_DRM_MGCG */ data = RREG32(SOC15_REG_OFFSET(MP0, 0, mmMP0_MISC_CGTT_CTRL0)); @@ -1645,7 +1492,7 @@ static int soc15_common_set_powergating_state(void *handle, return 0; } -const struct amd_ip_funcs soc15_common_ip_funcs = { +static const struct amd_ip_funcs soc15_common_ip_funcs = { .name = "soc15_common", .early_init = soc15_common_early_init, .late_init = soc15_common_late_init, diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.h b/drivers/gpu/drm/amd/amdgpu/soc15.h index 034cfdfc4dbe..efc2a253e8db 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.h +++ b/drivers/gpu/drm/amd/amdgpu/soc15.h @@ -28,11 +28,11 @@ #include "nbio_v7_0.h" #include "nbio_v7_4.h" +extern const struct amdgpu_ip_block_version vega10_common_ip_block; + #define SOC15_FLUSH_GPU_TLB_NUM_WREG 6 #define SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT 3 -extern const struct amd_ip_funcs soc15_common_ip_funcs; - struct soc15_reg_golden { u32 hwip; u32 instance; @@ -102,7 +102,6 @@ struct soc15_ras_field_entry { void soc15_grbm_select(struct amdgpu_device *adev, u32 me, u32 pipe, u32 queue, u32 vmid); void soc15_set_virt_ops(struct amdgpu_device *adev); -int soc15_set_ip_blocks(struct amdgpu_device *adev); void soc15_program_register_sequence(struct amdgpu_device *adev, const struct soc15_reg_golden *registers, diff --git a/drivers/gpu/drm/amd/amdgpu/soc15_common.h b/drivers/gpu/drm/amd/amdgpu/soc15_common.h index 8a9ca87d8663..473767e03676 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15_common.h +++ b/drivers/gpu/drm/amd/amdgpu/soc15_common.h @@ -51,6 +51,8 @@ #define RREG32_SOC15_IP(ip, reg) __RREG32_SOC15_RLC__(reg, 0, ip##_HWIP) +#define RREG32_SOC15_IP_NO_KIQ(ip, reg) __RREG32_SOC15_RLC__(reg, AMDGPU_REGS_NO_KIQ, ip##_HWIP) + #define RREG32_SOC15_NO_KIQ(ip, inst, reg) \ __RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, \ AMDGPU_REGS_NO_KIQ, ip##_HWIP) @@ -65,6 +67,9 @@ #define WREG32_SOC15_IP(ip, reg, value) \ __WREG32_SOC15_RLC__(reg, value, 0, ip##_HWIP) +#define WREG32_SOC15_IP_NO_KIQ(ip, reg, value) \ + __WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_NO_KIQ, ip##_HWIP) + #define WREG32_SOC15_NO_KIQ(ip, inst, reg, value) \ __WREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, \ value, AMDGPU_REGS_NO_KIQ, ip##_HWIP) diff --git a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h index 0f214a398dd8..5093826a43d1 100644 --- a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h +++ b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h @@ -38,9 +38,8 @@ enum ras_command { TA_RAS_COMMAND__TRIGGER_ERROR, }; -enum ta_ras_status -{ - TA_RAS_STATUS__SUCCESS = 0x00, +enum ta_ras_status { + TA_RAS_STATUS__SUCCESS = 0x0000, TA_RAS_STATUS__RESET_NEEDED = 0xA001, TA_RAS_STATUS__ERROR_INVALID_PARAMETER = 0xA002, TA_RAS_STATUS__ERROR_RAS_NOT_AVAILABLE = 0xA003, @@ -55,7 +54,17 @@ enum ta_ras_status TA_RAS_STATUS__ERROR_GET_DEV_INFO = 0xA00C, TA_RAS_STATUS__ERROR_UNSUPPORTED_DEV = 0xA00D, TA_RAS_STATUS__ERROR_NOT_INITIALIZED = 0xA00E, - TA_RAS_STATUS__ERROR_TEE_INTERNAL = 0xA00F + TA_RAS_STATUS__ERROR_TEE_INTERNAL = 0xA00F, + TA_RAS_STATUS__ERROR_UNSUPPORTED_FUNCTION = 0xA010, + TA_RAS_STATUS__ERROR_SYS_DRV_REG_ACCESS = 0xA011, + TA_RAS_STATUS__ERROR_RAS_READ_WRITE = 0xA012, + TA_RAS_STATUS__ERROR_NULL_PTR = 0xA013, + TA_RAS_STATUS__ERROR_UNSUPPORTED_IP = 0xA014, + TA_RAS_STATUS__ERROR_PCS_STATE_QUIET = 0xA015, + TA_RAS_STATUS__ERROR_PCS_STATE_ERROR = 0xA016, + TA_RAS_STATUS__ERROR_PCS_STATE_HANG = 0xA017, + TA_RAS_STATUS__ERROR_PCS_STATE_UNKNOWN = 0xA018, + TA_RAS_STATUS__ERROR_UNSUPPORTED_ERROR_INJ = 0xA019 }; enum ta_ras_block { @@ -73,9 +82,18 @@ enum ta_ras_block { TA_RAS_BLOCK__MP0, TA_RAS_BLOCK__MP1, TA_RAS_BLOCK__FUSE, + TA_RAS_BLOCK__MCA, TA_NUM_BLOCK_MAX }; +enum ta_ras_mca_block { + TA_RAS_MCA_BLOCK__MP0 = 0, + TA_RAS_MCA_BLOCK__MP1 = 1, + TA_RAS_MCA_BLOCK__MPIO = 2, + TA_RAS_MCA_BLOCK__IOHC = 3, + TA_MCA_NUM_BLOCK_MAX +}; + enum ta_ras_error_type { TA_RAS_ERROR__NONE = 0, TA_RAS_ERROR__PARITY = 1, @@ -105,17 +123,15 @@ struct ta_ras_trigger_error_input { uint64_t value; // method if error injection. i.e persistent, coherent etc. }; -struct ta_ras_init_flags -{ - uint8_t poison_mode_en; - uint8_t dgpu_mode; +struct ta_ras_init_flags { + uint8_t poison_mode_en; + uint8_t dgpu_mode; }; -struct ta_ras_output_flags -{ - uint8_t ras_init_success_flag; - uint8_t err_inject_switch_disable_flag; - uint8_t reg_access_failure_flag; +struct ta_ras_output_flags { + uint8_t ras_init_success_flag; + uint8_t err_inject_switch_disable_flag; + uint8_t reg_access_failure_flag; }; /* Common input structure for RAS callbacks */ @@ -126,14 +142,13 @@ union ta_ras_cmd_input { struct ta_ras_disable_features_input disable_features; struct ta_ras_trigger_error_input trigger_error; - uint32_t reserve_pad[256]; + uint32_t reserve_pad[256]; }; -union ta_ras_cmd_output -{ - struct ta_ras_output_flags flags; +union ta_ras_cmd_output { + struct ta_ras_output_flags flags; - uint32_t reserve_pad[256]; + uint32_t reserve_pad[256]; }; /* Shared Memory structures */ diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c index bb30336b1e8d..6dd1e19e8d43 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c +++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c @@ -50,6 +50,165 @@ static inline uint32_t get_umc_v6_7_reg_offset(struct amdgpu_device *adev, return adev->umc.channel_offs * ch_inst + UMC_V6_7_INST_DIST * umc_inst; } +static inline uint32_t get_umc_v6_7_channel_index(struct amdgpu_device *adev, + uint32_t umc_inst, + uint32_t ch_inst) +{ + return adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst]; +} + +static void umc_v6_7_ecc_info_query_correctable_error_count(struct amdgpu_device *adev, + uint32_t channel_index, + unsigned long *error_count) +{ + uint32_t ecc_err_cnt; + uint64_t mc_umc_status; + struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); + + /* + * select the lower chip and check the error count + * skip add error count, calc error counter only from mca_umc_status + */ + ecc_err_cnt = ras->umc_ecc.ecc[channel_index].ce_count_lo_chip; + + /* + * select the higher chip and check the err counter + * skip add error count, calc error counter only from mca_umc_status + */ + ecc_err_cnt = ras->umc_ecc.ecc[channel_index].ce_count_hi_chip; + + /* check for SRAM correctable error + MCUMC_STATUS is a 64 bit register */ + mc_umc_status = ras->umc_ecc.ecc[channel_index].mca_umc_status; + if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 && + REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1) + *error_count += 1; +} + +static void umc_v6_7_ecc_info_querry_uncorrectable_error_count(struct amdgpu_device *adev, + uint32_t channel_index, + unsigned long *error_count) +{ + uint64_t mc_umc_status; + struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); + + /* check the MCUMC_STATUS */ + mc_umc_status = ras->umc_ecc.ecc[channel_index].mca_umc_status; + if ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) && + (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1 || + REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 || + REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 || + REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 || + REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1)) + *error_count += 1; +} + +static void umc_v6_7_ecc_info_query_ras_error_count(struct amdgpu_device *adev, + void *ras_error_status) +{ + struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; + + uint32_t umc_inst = 0; + uint32_t ch_inst = 0; + uint32_t umc_reg_offset = 0; + uint32_t channel_index = 0; + + /*TODO: driver needs to toggle DF Cstate to ensure + * safe access of UMC registers. Will add the protection */ + LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) { + umc_reg_offset = get_umc_v6_7_reg_offset(adev, + umc_inst, + ch_inst); + channel_index = get_umc_v6_7_channel_index(adev, + umc_inst, + ch_inst); + umc_v6_7_ecc_info_query_correctable_error_count(adev, + channel_index, + &(err_data->ce_count)); + umc_v6_7_ecc_info_querry_uncorrectable_error_count(adev, + channel_index, + &(err_data->ue_count)); + } +} + +static void umc_v6_7_ecc_info_query_error_address(struct amdgpu_device *adev, + struct ras_err_data *err_data, + uint32_t umc_reg_offset, + uint32_t ch_inst, + uint32_t umc_inst) +{ + uint64_t mc_umc_status, err_addr, retired_page; + struct eeprom_table_record *err_rec; + uint32_t channel_index; + struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); + + channel_index = + adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst]; + + mc_umc_status = ras->umc_ecc.ecc[channel_index].mca_umc_status; + + if (mc_umc_status == 0) + return; + + if (!err_data->err_addr) + return; + + err_rec = &err_data->err_addr[err_data->err_addr_cnt]; + + /* calculate error address if ue/ce error is detected */ + if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 && + (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 || + REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)) { + + err_addr = ras->umc_ecc.ecc[channel_index].mca_umc_addr; + err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr); + + /* translate umc channel address to soc pa, 3 parts are included */ + retired_page = ADDR_OF_8KB_BLOCK(err_addr) | + ADDR_OF_256B_BLOCK(channel_index) | + OFFSET_IN_256B_BLOCK(err_addr); + + /* we only save ue error information currently, ce is skipped */ + if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) + == 1) { + err_rec->address = err_addr; + /* page frame address is saved */ + err_rec->retired_page = retired_page >> AMDGPU_GPU_PAGE_SHIFT; + err_rec->ts = (uint64_t)ktime_get_real_seconds(); + err_rec->err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE; + err_rec->cu = 0; + err_rec->mem_channel = channel_index; + err_rec->mcumc_id = umc_inst; + + err_data->err_addr_cnt++; + } + } +} + +static void umc_v6_7_ecc_info_query_ras_error_address(struct amdgpu_device *adev, + void *ras_error_status) +{ + struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; + + uint32_t umc_inst = 0; + uint32_t ch_inst = 0; + uint32_t umc_reg_offset = 0; + + /*TODO: driver needs to toggle DF Cstate to ensure + * safe access of UMC resgisters. Will add the protection + * when firmware interface is ready */ + LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) { + umc_reg_offset = get_umc_v6_7_reg_offset(adev, + umc_inst, + ch_inst); + umc_v6_7_ecc_info_query_error_address(adev, + err_data, + umc_reg_offset, + ch_inst, + umc_inst); + } +} + static void umc_v6_7_query_correctable_error_count(struct amdgpu_device *adev, uint32_t umc_reg_offset, unsigned long *error_count) @@ -288,9 +447,45 @@ static void umc_v6_7_query_ras_error_address(struct amdgpu_device *adev, } } +static uint32_t umc_v6_7_query_ras_poison_mode_per_channel( + struct amdgpu_device *adev, + uint32_t umc_reg_offset) +{ + uint32_t ecc_ctrl_addr, ecc_ctrl; + + ecc_ctrl_addr = + SOC15_REG_OFFSET(UMC, 0, regUMCCH0_0_EccCtrl); + ecc_ctrl = RREG32_PCIE((ecc_ctrl_addr + + umc_reg_offset) * 4); + + return REG_GET_FIELD(ecc_ctrl, UMCCH0_0_EccCtrl, UCFatalEn); +} + +static bool umc_v6_7_query_ras_poison_mode(struct amdgpu_device *adev) +{ + uint32_t umc_inst = 0; + uint32_t ch_inst = 0; + uint32_t umc_reg_offset = 0; + + LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) { + umc_reg_offset = get_umc_v6_7_reg_offset(adev, + umc_inst, + ch_inst); + /* Enabling fatal error in one channel will be considered + as fatal error mode */ + if (umc_v6_7_query_ras_poison_mode_per_channel(adev, umc_reg_offset)) + return false; + } + + return true; +} + const struct amdgpu_umc_ras_funcs umc_v6_7_ras_funcs = { .ras_late_init = amdgpu_umc_ras_late_init, .ras_fini = amdgpu_umc_ras_fini, .query_ras_error_count = umc_v6_7_query_ras_error_count, .query_ras_error_address = umc_v6_7_query_ras_error_address, + .query_ras_poison_mode = umc_v6_7_query_ras_poison_mode, + .ecc_info_query_ras_error_count = umc_v6_7_ecc_info_query_ras_error_count, + .ecc_info_query_ras_error_address = umc_v6_7_ecc_info_query_ras_error_address, }; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c index 7232241e3bfb..0fef925b6602 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c @@ -698,6 +698,19 @@ static int uvd_v3_1_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + cancel_delayed_work_sync(&adev->uvd.idle_work); + + if (RREG32(mmUVD_STATUS) != 0) + uvd_v3_1_stop(adev); + + return 0; +} + +static int uvd_v3_1_suspend(void *handle) +{ + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + /* * Proper cleanups before halting the HW engine: * - cancel the delayed idle work @@ -722,17 +735,6 @@ static int uvd_v3_1_hw_fini(void *handle) AMD_CG_STATE_GATE); } - if (RREG32(mmUVD_STATUS) != 0) - uvd_v3_1_stop(adev); - - return 0; -} - -static int uvd_v3_1_suspend(void *handle) -{ - int r; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = uvd_v3_1_hw_fini(adev); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c index 52d6de969f46..c108b8381795 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c @@ -212,6 +212,19 @@ static int uvd_v4_2_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + cancel_delayed_work_sync(&adev->uvd.idle_work); + + if (RREG32(mmUVD_STATUS) != 0) + uvd_v4_2_stop(adev); + + return 0; +} + +static int uvd_v4_2_suspend(void *handle) +{ + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + /* * Proper cleanups before halting the HW engine: * - cancel the delayed idle work @@ -236,17 +249,6 @@ static int uvd_v4_2_hw_fini(void *handle) AMD_CG_STATE_GATE); } - if (RREG32(mmUVD_STATUS) != 0) - uvd_v4_2_stop(adev); - - return 0; -} - -static int uvd_v4_2_suspend(void *handle) -{ - int r; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = uvd_v4_2_hw_fini(adev); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c index db6d06758e4d..563493d1f830 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c @@ -210,6 +210,19 @@ static int uvd_v5_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + cancel_delayed_work_sync(&adev->uvd.idle_work); + + if (RREG32(mmUVD_STATUS) != 0) + uvd_v5_0_stop(adev); + + return 0; +} + +static int uvd_v5_0_suspend(void *handle) +{ + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + /* * Proper cleanups before halting the HW engine: * - cancel the delayed idle work @@ -234,17 +247,6 @@ static int uvd_v5_0_hw_fini(void *handle) AMD_CG_STATE_GATE); } - if (RREG32(mmUVD_STATUS) != 0) - uvd_v5_0_stop(adev); - - return 0; -} - -static int uvd_v5_0_suspend(void *handle) -{ - int r; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = uvd_v5_0_hw_fini(adev); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c index bc571833632e..2d558c2f417d 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c @@ -332,15 +332,9 @@ err: static int uvd_v6_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct dma_fence *fence = NULL; - struct amdgpu_bo *bo = NULL; + struct amdgpu_bo *bo = ring->adev->uvd.ib_bo; long r; - r = amdgpu_bo_create_reserved(ring->adev, 128 * 1024, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, - &bo, NULL, NULL); - if (r) - return r; - r = uvd_v6_0_enc_get_create_msg(ring, 1, bo, NULL); if (r) goto error; @@ -357,9 +351,6 @@ static int uvd_v6_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) error: dma_fence_put(fence); - amdgpu_bo_unpin(bo); - amdgpu_bo_unreserve(bo); - amdgpu_bo_unref(&bo); return r; } @@ -543,6 +534,19 @@ static int uvd_v6_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + cancel_delayed_work_sync(&adev->uvd.idle_work); + + if (RREG32(mmUVD_STATUS) != 0) + uvd_v6_0_stop(adev); + + return 0; +} + +static int uvd_v6_0_suspend(void *handle) +{ + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + /* * Proper cleanups before halting the HW engine: * - cancel the delayed idle work @@ -567,17 +571,6 @@ static int uvd_v6_0_hw_fini(void *handle) AMD_CG_STATE_GATE); } - if (RREG32(mmUVD_STATUS) != 0) - uvd_v6_0_stop(adev); - - return 0; -} - -static int uvd_v6_0_suspend(void *handle) -{ - int r; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = uvd_v6_0_hw_fini(adev); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c index b6e82d75561f..b483f03b4591 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c @@ -338,15 +338,9 @@ err: static int uvd_v7_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct dma_fence *fence = NULL; - struct amdgpu_bo *bo = NULL; + struct amdgpu_bo *bo = ring->adev->uvd.ib_bo; long r; - r = amdgpu_bo_create_reserved(ring->adev, 128 * 1024, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, - &bo, NULL, NULL); - if (r) - return r; - r = uvd_v7_0_enc_get_create_msg(ring, 1, bo, NULL); if (r) goto error; @@ -363,9 +357,6 @@ static int uvd_v7_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) error: dma_fence_put(fence); - amdgpu_bo_unpin(bo); - amdgpu_bo_unreserve(bo); - amdgpu_bo_unref(&bo); return r; } @@ -606,6 +597,23 @@ static int uvd_v7_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + cancel_delayed_work_sync(&adev->uvd.idle_work); + + if (!amdgpu_sriov_vf(adev)) + uvd_v7_0_stop(adev); + else { + /* full access mode, so don't touch any UVD register */ + DRM_DEBUG("For SRIOV client, shouldn't do anything.\n"); + } + + return 0; +} + +static int uvd_v7_0_suspend(void *handle) +{ + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + /* * Proper cleanups before halting the HW engine: * - cancel the delayed idle work @@ -630,21 +638,6 @@ static int uvd_v7_0_hw_fini(void *handle) AMD_CG_STATE_GATE); } - if (!amdgpu_sriov_vf(adev)) - uvd_v7_0_stop(adev); - else { - /* full access mode, so don't touch any UVD register */ - DRM_DEBUG("For SRIOV client, shouldn't do anything.\n"); - } - - return 0; -} - -static int uvd_v7_0_suspend(void *handle) -{ - int r; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = uvd_v7_0_hw_fini(adev); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/vangogh_reg_init.c b/drivers/gpu/drm/amd/amdgpu/vangogh_reg_init.c deleted file mode 100644 index d64d681a05dc..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/vangogh_reg_init.c +++ /dev/null @@ -1,50 +0,0 @@ -/* - * Copyright 2019 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ -#include "amdgpu.h" -#include "nv.h" - -#include "soc15_common.h" -#include "soc15_hw_ip.h" -#include "vangogh_ip_offset.h" - -void vangogh_reg_base_init(struct amdgpu_device *adev) -{ - /* HW has more IP blocks, only initialized the blocke needed by driver */ - uint32_t i; - for (i = 0 ; i < MAX_INSTANCE ; ++i) { - adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i])); - adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i])); - adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i])); - adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIO_BASE.instance[i])); - adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i])); - adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i])); - adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(VCN_BASE.instance[i])); - adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i])); - adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DCN_BASE.instance[i])); - adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i])); - adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i])); - adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i])); - } -} diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c index b70c17f0c52e..67eb01fef789 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c @@ -431,10 +431,12 @@ static int vce_v2_0_sw_init(void *handle) return r; for (i = 0; i < adev->vce.num_rings; i++) { + enum amdgpu_ring_priority_level hw_prio = amdgpu_vce_get_ring_prio(i); + ring = &adev->vce.ring[i]; sprintf(ring->name, "vce%d", i); r = amdgpu_ring_init(adev, ring, 512, &adev->vce.irq, 0, - AMDGPU_RING_PRIO_DEFAULT, NULL); + hw_prio, NULL); if (r) return r; } @@ -479,6 +481,17 @@ static int vce_v2_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + cancel_delayed_work_sync(&adev->vce.idle_work); + + return 0; +} + +static int vce_v2_0_suspend(void *handle) +{ + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + /* * Proper cleanups before halting the HW engine: * - cancel the delayed idle work @@ -502,14 +515,6 @@ static int vce_v2_0_hw_fini(void *handle) AMD_CG_STATE_GATE); } - return 0; -} - -static int vce_v2_0_suspend(void *handle) -{ - int r; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = vce_v2_0_hw_fini(adev); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index 9de66893ccd6..142e291983b4 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -440,10 +440,12 @@ static int vce_v3_0_sw_init(void *handle) return r; for (i = 0; i < adev->vce.num_rings; i++) { + enum amdgpu_ring_priority_level hw_prio = amdgpu_vce_get_ring_prio(i); + ring = &adev->vce.ring[i]; sprintf(ring->name, "vce%d", i); r = amdgpu_ring_init(adev, ring, 512, &adev->vce.irq, 0, - AMDGPU_RING_PRIO_DEFAULT, NULL); + hw_prio, NULL); if (r) return r; } @@ -490,6 +492,21 @@ static int vce_v3_0_hw_fini(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + cancel_delayed_work_sync(&adev->vce.idle_work); + + r = vce_v3_0_wait_for_idle(handle); + if (r) + return r; + + vce_v3_0_stop(adev); + return vce_v3_0_set_clockgating_state(adev, AMD_CG_STATE_GATE); +} + +static int vce_v3_0_suspend(void *handle) +{ + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + /* * Proper cleanups before halting the HW engine: * - cancel the delayed idle work @@ -513,19 +530,6 @@ static int vce_v3_0_hw_fini(void *handle) AMD_CG_STATE_GATE); } - r = vce_v3_0_wait_for_idle(handle); - if (r) - return r; - - vce_v3_0_stop(adev); - return vce_v3_0_set_clockgating_state(adev, AMD_CG_STATE_GATE); -} - -static int vce_v3_0_suspend(void *handle) -{ - int r; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = vce_v3_0_hw_fini(adev); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c index fec902b800c2..d1fc4e0b8265 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c @@ -463,6 +463,8 @@ static int vce_v4_0_sw_init(void *handle) } for (i = 0; i < adev->vce.num_rings; i++) { + enum amdgpu_ring_priority_level hw_prio = amdgpu_vce_get_ring_prio(i); + ring = &adev->vce.ring[i]; sprintf(ring->name, "vce%d", i); if (amdgpu_sriov_vf(adev)) { @@ -478,7 +480,7 @@ static int vce_v4_0_sw_init(void *handle) ring->doorbell_index = adev->doorbell_index.uvd_vce.vce_ring2_3 * 2 + 1; } r = amdgpu_ring_init(adev, ring, 512, &adev->vce.irq, 0, - AMDGPU_RING_PRIO_DEFAULT, NULL); + hw_prio, NULL); if (r) return r; } @@ -542,29 +544,8 @@ static int vce_v4_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - /* - * Proper cleanups before halting the HW engine: - * - cancel the delayed idle work - * - enable powergating - * - enable clockgating - * - disable dpm - * - * TODO: to align with the VCN implementation, move the - * jobs for clockgating/powergating/dpm setting to - * ->set_powergating_state(). - */ cancel_delayed_work_sync(&adev->vce.idle_work); - if (adev->pm.dpm_enabled) { - amdgpu_dpm_enable_vce(adev, false); - } else { - amdgpu_asic_set_vce_clocks(adev, 0, 0); - amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE, - AMD_PG_STATE_GATE); - amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE, - AMD_CG_STATE_GATE); - } - if (!amdgpu_sriov_vf(adev)) { /* vce_v4_0_wait_for_idle(handle); */ vce_v4_0_stop(adev); @@ -584,7 +565,7 @@ static int vce_v4_0_suspend(void *handle) if (adev->vce.vcpu_bo == NULL) return 0; - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { unsigned size = amdgpu_bo_size(adev->vce.vcpu_bo); void *ptr = adev->vce.cpu_addr; @@ -594,6 +575,29 @@ static int vce_v4_0_suspend(void *handle) drm_dev_exit(idx); } + /* + * Proper cleanups before halting the HW engine: + * - cancel the delayed idle work + * - enable powergating + * - enable clockgating + * - disable dpm + * + * TODO: to align with the VCN implementation, move the + * jobs for clockgating/powergating/dpm setting to + * ->set_powergating_state(). + */ + cancel_delayed_work_sync(&adev->vce.idle_work); + + if (adev->pm.dpm_enabled) { + amdgpu_dpm_enable_vce(adev, false); + } else { + amdgpu_asic_set_vce_clocks(adev, 0, 0); + amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE, + AMD_PG_STATE_GATE); + amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE, + AMD_CG_STATE_GATE); + } + r = vce_v4_0_hw_fini(adev); if (r) return r; @@ -611,7 +615,7 @@ static int vce_v4_0_resume(void *handle) if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { unsigned size = amdgpu_bo_size(adev->vce.vcpu_bo); void *ptr = adev->vce.cpu_addr; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index 121ee9f2b8d1..3799226defc0 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -66,7 +66,6 @@ static int vcn_v1_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - adev->vcn.num_vcn_inst = 1; adev->vcn.num_enc_rings = 2; vcn_v1_0_set_dec_ring_funcs(adev); @@ -112,15 +111,7 @@ static int vcn_v1_0_sw_init(void *handle) /* Override the work func */ adev->vcn.idle_work.work.func = vcn_v1_0_idle_work_handler; - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { - const struct common_firmware_header *hdr; - hdr = (const struct common_firmware_header *)adev->vcn.fw->data; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].ucode_id = AMDGPU_UCODE_ID_VCN; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].fw = adev->vcn.fw; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); - dev_info(adev->dev, "Will use PSP to load VCN firmware\n"); - } + amdgpu_vcn_setup_ucode(adev); r = amdgpu_vcn_resume(adev); if (r) @@ -145,10 +136,12 @@ static int vcn_v1_0_sw_init(void *handle) SOC15_REG_OFFSET(UVD, 0, mmUVD_NO_OP); for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i); + ring = &adev->vcn.inst->ring_enc[i]; sprintf(ring->name, "vcn_enc%d", i); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst->irq, 0, - AMDGPU_RING_PRIO_DEFAULT, NULL); + hw_prio, NULL); if (r) return r; } @@ -253,6 +246,13 @@ static int vcn_v1_0_suspend(void *handle) { int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + bool idle_work_unexecuted; + + idle_work_unexecuted = cancel_delayed_work_sync(&adev->vcn.idle_work); + if (idle_work_unexecuted) { + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_uvd(adev, false); + } r = vcn_v1_0_hw_fini(adev); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c index f4686e918e0d..313fc1b53999 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c @@ -22,6 +22,7 @@ */ #include <linux/firmware.h> +#include <drm/drm_drv.h> #include "amdgpu.h" #include "amdgpu_vcn.h" @@ -68,7 +69,6 @@ static int vcn_v2_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - adev->vcn.num_vcn_inst = 1; if (amdgpu_sriov_vf(adev)) adev->vcn.num_enc_rings = 1; else @@ -115,15 +115,7 @@ static int vcn_v2_0_sw_init(void *handle) if (r) return r; - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { - const struct common_firmware_header *hdr; - hdr = (const struct common_firmware_header *)adev->vcn.fw->data; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].ucode_id = AMDGPU_UCODE_ID_VCN; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].fw = adev->vcn.fw; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); - dev_info(adev->dev, "Will use PSP to load VCN firmware\n"); - } + amdgpu_vcn_setup_ucode(adev); r = amdgpu_vcn_resume(adev); if (r) @@ -159,6 +151,8 @@ static int vcn_v2_0_sw_init(void *handle) adev->vcn.inst->external.nop = SOC15_REG_OFFSET(UVD, 0, mmUVD_NO_OP); for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i); + ring = &adev->vcn.inst->ring_enc[i]; ring->use_doorbell = true; if (!amdgpu_sriov_vf(adev)) @@ -167,7 +161,7 @@ static int vcn_v2_0_sw_init(void *handle) ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1 + i; sprintf(ring->name, "vcn_enc%d", i); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst->irq, 0, - AMDGPU_RING_PRIO_DEFAULT, NULL); + hw_prio, NULL); if (r) return r; } @@ -192,11 +186,14 @@ static int vcn_v2_0_sw_init(void *handle) */ static int vcn_v2_0_sw_fini(void *handle) { - int r; + int r, idx; struct amdgpu_device *adev = (struct amdgpu_device *)handle; volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared_cpu_addr; - fw_shared->present_flag_0 = 0; + if (drm_dev_enter(adev_to_drm(adev), &idx)) { + fw_shared->present_flag_0 = 0; + drm_dev_exit(idx); + } amdgpu_virt_free_mm_table(adev); @@ -1879,15 +1876,14 @@ static int vcn_v2_0_start_sriov(struct amdgpu_device *adev) /* mc resume*/ if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { - tmp = AMDGPU_UCODE_ID_VCN; MMSCH_V2_0_INSERT_DIRECT_WT( SOC15_REG_OFFSET(UVD, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), - adev->firmware.ucode[tmp].tmr_mc_addr_lo); + adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo); MMSCH_V2_0_INSERT_DIRECT_WT( SOC15_REG_OFFSET(UVD, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), - adev->firmware.ucode[tmp].tmr_mc_addr_hi); + adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_hi); offset = 0; } else { MMSCH_V2_0_INSERT_DIRECT_WT( diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index e0c0c3734432..44fc4c218433 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -22,6 +22,7 @@ */ #include <linux/firmware.h> +#include <drm/drm_drv.h> #include "amdgpu.h" #include "amdgpu_vcn.h" @@ -82,7 +83,7 @@ static int vcn_v2_5_early_init(void *handle) } else { u32 harvest; int i; - adev->vcn.num_vcn_inst = VCN25_MAX_HW_INSTANCES_ARCTURUS; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { harvest = RREG32_SOC15(VCN, i, mmCC_UVD_HARVESTING); if (harvest & CC_UVD_HARVESTING__UVD_DISABLE_MASK) @@ -138,22 +139,7 @@ static int vcn_v2_5_sw_init(void *handle) if (r) return r; - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { - const struct common_firmware_header *hdr; - hdr = (const struct common_firmware_header *)adev->vcn.fw->data; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].ucode_id = AMDGPU_UCODE_ID_VCN; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].fw = adev->vcn.fw; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); - - if (adev->vcn.num_vcn_inst == VCN25_MAX_HW_INSTANCES_ARCTURUS) { - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN1].ucode_id = AMDGPU_UCODE_ID_VCN1; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN1].fw = adev->vcn.fw; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); - } - dev_info(adev->dev, "Will use PSP to load VCN firmware\n"); - } + amdgpu_vcn_setup_ucode(adev); r = amdgpu_vcn_resume(adev); if (r) @@ -194,6 +180,8 @@ static int vcn_v2_5_sw_init(void *handle) return r; for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i); + ring = &adev->vcn.inst[j].ring_enc[i]; ring->use_doorbell = true; @@ -203,7 +191,7 @@ static int vcn_v2_5_sw_init(void *handle) sprintf(ring->name, "vcn_enc_%d.%d", j, i); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[j].irq, 0, - AMDGPU_RING_PRIO_DEFAULT, NULL); + hw_prio, NULL); if (r) return r; } @@ -233,17 +221,21 @@ static int vcn_v2_5_sw_init(void *handle) */ static int vcn_v2_5_sw_fini(void *handle) { - int i, r; + int i, r, idx; struct amdgpu_device *adev = (struct amdgpu_device *)handle; volatile struct amdgpu_fw_shared *fw_shared; - for (i = 0; i < adev->vcn.num_vcn_inst; i++) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - fw_shared = adev->vcn.inst[i].fw_shared_cpu_addr; - fw_shared->present_flag_0 = 0; + if (drm_dev_enter(adev_to_drm(adev), &idx)) { + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + if (adev->vcn.harvest_config & (1 << i)) + continue; + fw_shared = adev->vcn.inst[i].fw_shared_cpu_addr; + fw_shared->present_flag_0 = 0; + } + drm_dev_exit(idx); } + if (amdgpu_sriov_vf(adev)) amdgpu_virt_free_mm_table(adev); @@ -1713,7 +1705,7 @@ static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev) for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { if (adev->vcn.harvest_config & (1 << i)) continue; - if (adev->asic_type == CHIP_ARCTURUS) + if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(2, 5, 0)) adev->vcn.inst[i].ring_dec.funcs = &vcn_v2_5_dec_ring_vm_funcs; else /* CHIP_ALDEBARAN */ adev->vcn.inst[i].ring_dec.funcs = &vcn_v2_6_dec_ring_vm_funcs; @@ -1730,7 +1722,7 @@ static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev) if (adev->vcn.harvest_config & (1 << j)) continue; for (i = 0; i < adev->vcn.num_enc_rings; ++i) { - if (adev->asic_type == CHIP_ARCTURUS) + if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(2, 5, 0)) adev->vcn.inst[j].ring_enc[i].funcs = &vcn_v2_5_enc_ring_vm_funcs; else /* CHIP_ALDEBARAN */ adev->vcn.inst[j].ring_enc[i].funcs = &vcn_v2_6_enc_ring_vm_funcs; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index 3d18aab88b4e..da11ceba0698 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -60,11 +60,6 @@ static int amdgpu_ih_clientid_vcns[] = { SOC15_IH_CLIENTID_VCN1 }; -static int amdgpu_ucode_id_vcns[] = { - AMDGPU_UCODE_ID_VCN, - AMDGPU_UCODE_ID_VCN1 -}; - static int vcn_v3_0_start_sriov(struct amdgpu_device *adev); static void vcn_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev); static void vcn_v3_0_set_enc_ring_funcs(struct amdgpu_device *adev); @@ -87,7 +82,6 @@ static void vcn_v3_0_enc_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v3_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - int i; if (amdgpu_sriov_vf(adev)) { adev->vcn.num_vcn_inst = VCN_INSTANCES_SIENNA_CICHLID; @@ -95,24 +89,12 @@ static int vcn_v3_0_early_init(void *handle) adev->vcn.num_enc_rings = 1; } else { - if (adev->asic_type == CHIP_SIENNA_CICHLID) { - u32 harvest; - - adev->vcn.num_vcn_inst = VCN_INSTANCES_SIENNA_CICHLID; - for (i = 0; i < adev->vcn.num_vcn_inst; i++) { - harvest = RREG32_SOC15(VCN, i, mmCC_UVD_HARVESTING); - if (harvest & CC_UVD_HARVESTING__UVD_DISABLE_MASK) - adev->vcn.harvest_config |= 1 << i; - } - - if (adev->vcn.harvest_config == (AMDGPU_VCN_HARVEST_VCN0 | - AMDGPU_VCN_HARVEST_VCN1)) - /* both instances are harvested, disable the block */ - return -ENOENT; - } else - adev->vcn.num_vcn_inst = 1; + if (adev->vcn.harvest_config == (AMDGPU_VCN_HARVEST_VCN0 | + AMDGPU_VCN_HARVEST_VCN1)) + /* both instances are harvested, disable the block */ + return -ENOENT; - if (adev->asic_type == CHIP_BEIGE_GOBY) + if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(3, 0, 33)) adev->vcn.num_enc_rings = 0; else adev->vcn.num_enc_rings = 2; @@ -143,22 +125,7 @@ static int vcn_v3_0_sw_init(void *handle) if (r) return r; - if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { - const struct common_firmware_header *hdr; - hdr = (const struct common_firmware_header *)adev->vcn.fw->data; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].ucode_id = AMDGPU_UCODE_ID_VCN; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].fw = adev->vcn.fw; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); - - if (adev->vcn.num_vcn_inst == VCN_INSTANCES_SIENNA_CICHLID) { - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN1].ucode_id = AMDGPU_UCODE_ID_VCN1; - adev->firmware.ucode[AMDGPU_UCODE_ID_VCN1].fw = adev->vcn.fw; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); - } - dev_info(adev->dev, "Will use PSP to load VCN firmware\n"); - } + amdgpu_vcn_setup_ucode(adev); r = amdgpu_vcn_resume(adev); if (r) @@ -224,6 +191,8 @@ static int vcn_v3_0_sw_init(void *handle) return r; for (j = 0; j < adev->vcn.num_enc_rings; ++j) { + enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(j); + /* VCN ENC TRAP */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i], j + VCN_2_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst[i].irq); @@ -239,8 +208,7 @@ static int vcn_v3_0_sw_init(void *handle) } sprintf(ring->name, "vcn_enc_%d.%d", i, j); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0, - AMDGPU_RING_PRIO_DEFAULT, - &adev->vcn.inst[i].sched_score); + hw_prio, &adev->vcn.inst[i].sched_score); if (r) return r; } @@ -275,7 +243,7 @@ static int vcn_v3_0_sw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; int i, r, idx; - if (drm_dev_enter(&adev->ddev, &idx)) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { for (i = 0; i < adev->vcn.num_vcn_inst; i++) { volatile struct amdgpu_fw_shared *fw_shared; @@ -1271,7 +1239,7 @@ static int vcn_v3_0_start(struct amdgpu_device *adev) fw_shared->rb.wptr = lower_32_bits(ring->wptr); fw_shared->multi_queue.decode_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET); - if (adev->asic_type != CHIP_BEIGE_GOBY) { + if (adev->ip_versions[UVD_HWIP][0] != IP_VERSION(3, 0, 33)) { fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET); ring = &adev->vcn.inst[i].ring_enc[0]; WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr)); @@ -1305,7 +1273,6 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev) uint32_t param, resp, expected; uint32_t offset, cache_size; uint32_t tmp, timeout; - uint32_t id; struct amdgpu_mm_table *table = &adev->virt.mm_table; uint32_t *table_loc; @@ -1349,13 +1316,12 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev) cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4); if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { - id = amdgpu_ucode_id_vcns[i]; MMSCH_V3_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), - adev->firmware.ucode[id].tmr_mc_addr_lo); + adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo); MMSCH_V3_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), - adev->firmware.ucode[id].tmr_mc_addr_hi); + adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi); offset = 0; MMSCH_V3_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CACHE_OFFSET0), @@ -1643,7 +1609,7 @@ static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev, UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK, ~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK); - if (adev->asic_type != CHIP_BEIGE_GOBY) { + if (adev->ip_versions[UVD_HWIP][0] != IP_VERSION(3, 0, 33)) { /* Restore */ fw_shared = adev->vcn.inst[inst_idx].fw_shared_cpu_addr; fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET); diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c index a9ca6988009e..3070466f54e1 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c @@ -640,6 +640,7 @@ const struct amd_ip_funcs vega10_ih_ip_funcs = { static const struct amdgpu_ih_funcs vega10_ih_funcs = { .get_wptr = vega10_ih_get_wptr, .decode_iv = amdgpu_ih_decode_iv_helper, + .decode_iv_ts = amdgpu_ih_decode_iv_ts_helper, .set_rptr = vega10_ih_set_rptr }; diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c index f51dfc38ac65..3b4eb8285943 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c @@ -688,6 +688,7 @@ const struct amd_ip_funcs vega20_ih_ip_funcs = { static const struct amdgpu_ih_funcs vega20_ih_funcs = { .get_wptr = vega20_ih_get_wptr, .decode_iv = amdgpu_ih_decode_iv_helper, + .decode_iv_ts = amdgpu_ih_decode_iv_ts_helper, .set_rptr = vega20_ih_set_rptr }; diff --git a/drivers/gpu/drm/amd/amdgpu/yellow_carp_reg_init.c b/drivers/gpu/drm/amd/amdgpu/yellow_carp_reg_init.c deleted file mode 100644 index 3d89421275ed..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/yellow_carp_reg_init.c +++ /dev/null @@ -1,51 +0,0 @@ -/* - * Copyright 2019 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ -#include "amdgpu.h" -#include "nv.h" - -#include "soc15_common.h" -#include "soc15_hw_ip.h" -#include "yellow_carp_offset.h" - -int yellow_carp_reg_base_init(struct amdgpu_device *adev) -{ - /* HW has more IP blocks, only initialized the block needed by driver */ - uint32_t i; - for (i = 0 ; i < MAX_INSTANCE ; ++i) { - adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i])); - adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i])); - adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i])); - adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i])); - adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIO_BASE.instance[i])); - adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i])); - adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i])); - adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(VCN_BASE.instance[i])); - adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i])); - adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DCN_BASE.instance[i])); - adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i])); - adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(SDMA0_BASE.instance[i])); - adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i])); - adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i])); - } - return 0; -} |