| 1 |
|
| 2 | |
| 3 | * Copyright 2023 Advanced Micro Devices, Inc.
|
| 4 | *
|
| 5 | * Permission is hereby granted, free of charge, to any person obtaining a
|
| 6 | * copy of this software and associated documentation files (the "Software"),
|
| 7 | * to deal in the Software without restriction, including without limitation
|
| 8 | * the rights to use, copy, modify, merge, publish, distribute, sublicense,
|
| 9 | * and/or sell copies of the Software, and to permit persons to whom the
|
| 10 | * Software is furnished to do so, subject to the following conditions:
|
| 11 | *
|
| 12 | * The above copyright notice and this permission notice shall be included in
|
| 13 | * all copies or substantial portions of the Software.
|
| 14 | *
|
| 15 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
| 16 | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
| 17 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
|
| 18 | * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
|
| 19 | * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
|
| 20 | * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
|
| 21 | * OTHER DEALINGS IN THE SOFTWARE.
|
| 22 | *
|
| 23 |
|
| 24 |
|
| 25 | #ifndef __AMDGPU_UMSCH_MM_H__
|
| 26 | #define __AMDGPU_UMSCH_MM_H__
|
| 27 |
|
| 28 | enum UMSCH_SWIP_ENGINE_TYPE {
|
| 29 | UMSCH_SWIP_ENGINE_TYPE_VCN0 = 0,
|
| 30 | UMSCH_SWIP_ENGINE_TYPE_VCN1 = 1,
|
| 31 | UMSCH_SWIP_ENGINE_TYPE_VCN = 2,
|
| 32 | UMSCH_SWIP_ENGINE_TYPE_VPE = 3,
|
| 33 | UMSCH_SWIP_ENGINE_TYPE_MAX
|
| 34 | };
|
| 35 |
|
| 36 | enum UMSCH_CONTEXT_PRIORITY_LEVEL {
|
| 37 | CONTEXT_PRIORITY_LEVEL_IDLE = 0,
|
| 38 | CONTEXT_PRIORITY_LEVEL_NORMAL = 1,
|
| 39 | CONTEXT_PRIORITY_LEVEL_FOCUS = 2,
|
| 40 | CONTEXT_PRIORITY_LEVEL_REALTIME = 3,
|
| 41 | CONTEXT_PRIORITY_NUM_LEVELS
|
| 42 | };
|
| 43 |
|
| 44 | struct umsch_mm_set_resource_input {
|
| 45 | uint32_t vmid_mask_mm_vcn;
|
| 46 | uint32_t vmid_mask_mm_vpe;
|
| 47 | uint32_t collaboration_mask_vpe;
|
| 48 | uint32_t logging_vmid;
|
| 49 | uint32_t engine_mask;
|
| 50 | union {
|
| 51 | struct {
|
| 52 | uint32_t disable_reset : 1;
|
| 53 | uint32_t disable_umsch_mm_log : 1;
|
| 54 | uint32_t use_rs64mem_for_proc_ctx_csa : 1;
|
| 55 | uint32_t reserved : 29;
|
| 56 | };
|
| 57 | uint32_t uint32_all;
|
| 58 | };
|
| 59 | };
|
| 60 |
|
| 61 | struct amdgpu_umsch_fwlog {
|
| 62 | uint32_t rptr;
|
| 63 | uint32_t wptr;
|
| 64 | uint32_t buffer_size;
|
| 65 | uint32_t header_size;
|
| 66 | uint32_t wrapped;
|
| 67 | };
|
| 68 |
|
| 69 | struct umsch_mm_add_queue_input {
|
| 70 | uint32_t process_id;
|
| 71 | uint64_t page_table_base_addr;
|
| 72 | uint64_t process_va_start;
|
| 73 | uint64_t process_va_end;
|
| 74 | uint64_t process_quantum;
|
| 75 | uint64_t process_csa_addr;
|
| 76 | uint64_t context_quantum;
|
| 77 | uint64_t context_csa_addr;
|
| 78 | uint32_t inprocess_context_priority;
|
| 79 | enum UMSCH_CONTEXT_PRIORITY_LEVEL context_global_priority_level;
|
| 80 | uint32_t doorbell_offset_0;
|
| 81 | uint32_t doorbell_offset_1;
|
| 82 | enum UMSCH_SWIP_ENGINE_TYPE engine_type;
|
| 83 | uint32_t affinity;
|
| 84 | uint64_t mqd_addr;
|
| 85 | uint64_t h_context;
|
| 86 | uint64_t h_queue;
|
| 87 | uint32_t vm_context_cntl;
|
| 88 |
|
| 89 | uint32_t process_csa_array_index;
|
| 90 | uint32_t context_csa_array_index;
|
| 91 |
|
| 92 | struct {
|
| 93 | uint32_t is_context_suspended : 1;
|
| 94 | uint32_t collaboration_mode : 1;
|
| 95 | uint32_t reserved : 30;
|
| 96 | };
|
| 97 | };
|
| 98 |
|
| 99 | struct umsch_mm_remove_queue_input {
|
| 100 | uint32_t doorbell_offset_0;
|
| 101 | uint32_t doorbell_offset_1;
|
| 102 | uint64_t context_csa_addr;
|
| 103 | uint32_t context_csa_array_index;
|
| 104 | };
|
| 105 |
|
| 106 | struct MQD_INFO {
|
| 107 | uint32_t rb_base_hi;
|
| 108 | uint32_t rb_base_lo;
|
| 109 | uint32_t rb_size;
|
| 110 | uint32_t wptr_val;
|
| 111 | uint32_t rptr_val;
|
| 112 | uint32_t unmapped;
|
| 113 | uint32_t vmid;
|
| 114 | };
|
| 115 |
|
| 116 | struct amdgpu_umsch_mm;
|
| 117 |
|
| 118 | struct umsch_mm_funcs {
|
| 119 | int (*set_hw_resources)(struct amdgpu_umsch_mm *umsch);
|
| 120 | int (*add_queue)(struct amdgpu_umsch_mm *umsch,
|
| 121 | struct umsch_mm_add_queue_input *input);
|
| 122 | int (*remove_queue)(struct amdgpu_umsch_mm *umsch,
|
| 123 | struct umsch_mm_remove_queue_input *input);
|
| 124 | int (*set_regs)(struct amdgpu_umsch_mm *umsch);
|
| 125 | int (*init_microcode)(struct amdgpu_umsch_mm *umsch);
|
| 126 | int (*load_microcode)(struct amdgpu_umsch_mm *umsch);
|
| 127 | int (*ring_init)(struct amdgpu_umsch_mm *umsch);
|
| 128 | int (*ring_start)(struct amdgpu_umsch_mm *umsch);
|
| 129 | int (*ring_stop)(struct amdgpu_umsch_mm *umsch);
|
| 130 | int (*ring_fini)(struct amdgpu_umsch_mm *umsch);
|
| 131 | };
|
| 132 |
|
| 133 | struct amdgpu_umsch_mm {
|
| 134 | struct amdgpu_ring ring;
|
| 135 |
|
| 136 | uint32_t rb_wptr;
|
| 137 | uint32_t rb_rptr;
|
| 138 |
|
| 139 | const struct umsch_mm_funcs *funcs;
|
| 140 |
|
| 141 | const struct firmware *fw;
|
| 142 | uint32_t fw_version;
|
| 143 | uint32_t feature_version;
|
| 144 |
|
| 145 | struct amdgpu_bo *ucode_fw_obj;
|
| 146 | uint64_t ucode_fw_gpu_addr;
|
| 147 | uint32_t *ucode_fw_ptr;
|
| 148 | uint64_t irq_start_addr;
|
| 149 | uint64_t uc_start_addr;
|
| 150 | uint32_t ucode_size;
|
| 151 |
|
| 152 | struct amdgpu_bo *data_fw_obj;
|
| 153 | uint64_t data_fw_gpu_addr;
|
| 154 | uint32_t *data_fw_ptr;
|
| 155 | uint64_t data_start_addr;
|
| 156 | uint32_t data_size;
|
| 157 |
|
| 158 | struct amdgpu_bo *cmd_buf_obj;
|
| 159 | uint64_t cmd_buf_gpu_addr;
|
| 160 | uint32_t *cmd_buf_ptr;
|
| 161 | uint32_t *cmd_buf_curr_ptr;
|
| 162 |
|
| 163 | uint32_t wb_index;
|
| 164 | uint64_t sch_ctx_gpu_addr;
|
| 165 | uint32_t *sch_ctx_cpu_addr;
|
| 166 |
|
| 167 | uint32_t vmid_mask_mm_vcn;
|
| 168 | uint32_t vmid_mask_mm_vpe;
|
| 169 | uint32_t engine_mask;
|
| 170 | uint32_t vcn0_hqd_mask;
|
| 171 | uint32_t vcn1_hqd_mask;
|
| 172 | uint32_t vcn_hqd_mask[2];
|
| 173 | uint32_t vpe_hqd_mask;
|
| 174 | uint32_t agdb_index[CONTEXT_PRIORITY_NUM_LEVELS];
|
| 175 |
|
| 176 | struct mutex mutex_hidden;
|
| 177 | struct amdgpu_bo *dbglog_bo;
|
| 178 | void *log_cpu_addr;
|
| 179 | uint64_t log_gpu_addr;
|
| 180 | uint32_t mem_size;
|
| 181 | uint32_t log_offset;
|
| 182 | };
|
| 183 |
|
| 184 | int amdgpu_umsch_mm_submit_pkt(struct amdgpu_umsch_mm *umsch, void *pkt, int ndws);
|
| 185 | int amdgpu_umsch_mm_query_fence(struct amdgpu_umsch_mm *umsch);
|
| 186 |
|
| 187 | int amdgpu_umsch_mm_init_microcode(struct amdgpu_umsch_mm *umsch);
|
| 188 | int amdgpu_umsch_mm_allocate_ucode_buffer(struct amdgpu_umsch_mm *umsch);
|
| 189 | int amdgpu_umsch_mm_allocate_ucode_data_buffer(struct amdgpu_umsch_mm *umsch);
|
| 190 |
|
| 191 | int amdgpu_umsch_mm_psp_execute_cmd_buf(struct amdgpu_umsch_mm *umsch);
|
| 192 |
|
| 193 | int amdgpu_umsch_mm_ring_init(struct amdgpu_umsch_mm *umsch);
|
| 194 |
|
| 195 | void amdgpu_debugfs_umsch_fwlog_init(struct amdgpu_device *adev,
|
| 196 | struct amdgpu_umsch_mm *umsch);
|
| 197 |
|
| 198 | void amdgpu_umsch_fwlog_init(struct amdgpu_umsch_mm *umsch_mm);
|
| 199 |
|
| 200 | #define WREG32_SOC15_UMSCH(reg, value) \
|
| 201 | do { \
|
| 202 | uint32_t reg_offset = adev->reg_offset[VCN_HWIP][0][reg##_BASE_IDX] + reg; \
|
| 203 | if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { \
|
| 204 | *adev->umsch_mm.cmd_buf_curr_ptr++ = (reg_offset << 2); \
|
| 205 | *adev->umsch_mm.cmd_buf_curr_ptr++ = value; \
|
| 206 | } else { \
|
| 207 | WREG32(reg_offset, value); \
|
| 208 | } \
|
| 209 | } while (0)
|
| 210 |
|
| 211 | #define umsch_mm_set_hw_resources(umsch) \
|
| 212 | ((umsch)->funcs->set_hw_resources ? (umsch)->funcs->set_hw_resources((umsch)) : 0)
|
| 213 | #define umsch_mm_add_queue(umsch, input) \
|
| 214 | ((umsch)->funcs->add_queue ? (umsch)->funcs->add_queue((umsch), (input)) : 0)
|
| 215 | #define umsch_mm_remove_queue(umsch, input) \
|
| 216 | ((umsch)->funcs->remove_queue ? (umsch)->funcs->remove_queue((umsch), (input)) : 0)
|
| 217 |
|
| 218 | #define umsch_mm_set_regs(umsch) \
|
| 219 | ((umsch)->funcs->set_regs ? (umsch)->funcs->set_regs((umsch)) : 0)
|
| 220 | #define umsch_mm_init_microcode(umsch) \
|
| 221 | ((umsch)->funcs->init_microcode ? (umsch)->funcs->init_microcode((umsch)) : 0)
|
| 222 | #define umsch_mm_load_microcode(umsch) \
|
| 223 | ((umsch)->funcs->load_microcode ? (umsch)->funcs->load_microcode((umsch)) : 0)
|
| 224 |
|
| 225 | #define umsch_mm_ring_init(umsch) \
|
| 226 | ((umsch)->funcs->ring_init ? (umsch)->funcs->ring_init((umsch)) : 0)
|
| 227 | #define umsch_mm_ring_start(umsch) \
|
| 228 | ((umsch)->funcs->ring_start ? (umsch)->funcs->ring_start((umsch)) : 0)
|
| 229 | #define umsch_mm_ring_stop(umsch) \
|
| 230 | ((umsch)->funcs->ring_stop ? (umsch)->funcs->ring_stop((umsch)) : 0)
|
| 231 | #define umsch_mm_ring_fini(umsch) \
|
| 232 | ((umsch)->funcs->ring_fini ? (umsch)->funcs->ring_fini((umsch)) : 0)
|
| 233 |
|
| 234 | static inline void amdgpu_umsch_mm_lock(struct amdgpu_umsch_mm *umsch)
|
| 235 | {
|
| 236 | mutex_lock(&umsch->mutex_hidden);
|
| 237 | }
|
| 238 |
|
| 239 | static inline void amdgpu_umsch_mm_unlock(struct amdgpu_umsch_mm *umsch)
|
| 240 | {
|
| 241 | mutex_unlock(&umsch->mutex_hidden);
|
| 242 | }
|
| 243 |
|
| 244 | extern const struct amdgpu_ip_block_version umsch_mm_v4_0_ip_block;
|
| 245 |
|
| 246 | #endif
|