1 /* 2 * Copyright 2018 Advanced Micro Devices, Inc. 3 * 4 * Permission is hereby granted, free of charge, to any person obtaining a 5 * copy of this software and associated documentation files (the "Software"), 6 * to deal in the Software without restriction, including without limitation 7 * the rights to use, copy, modify, merge, publish, distribute, sublicense, 8 * and/or sell copies of the Software, and to permit persons to whom the 9 * Software is furnished to do so, subject to the following conditions: 10 * 11 * The above copyright notice and this permission notice shall be included in 12 * all copies or substantial portions of the Software. 13 * 14 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 15 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 16 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL 17 * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR 18 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, 19 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR 20 * OTHER DEALINGS IN THE SOFTWARE. 21 * 22 * 23 */ 24 #ifndef _AMDGPU_RAS_H 25 #define _AMDGPU_RAS_H 26 27 #include <linux/debugfs.h> 28 #include <linux/list.h> 29 #include "ta_ras_if.h" 30 #include "amdgpu_ras_eeprom.h" 31 #include "amdgpu_smuio.h" 32 #include "amdgpu_aca.h" 33 34 struct amdgpu_iv_entry; 35 36 #define AMDGPU_RAS_GPU_ERR_MEM_TRAINING(x) AMDGPU_GET_REG_FIELD(x, 0, 0) 37 #define AMDGPU_RAS_GPU_ERR_FW_LOAD(x) AMDGPU_GET_REG_FIELD(x, 1, 1) 38 #define AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(x) AMDGPU_GET_REG_FIELD(x, 2, 2) 39 #define AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(x) AMDGPU_GET_REG_FIELD(x, 3, 3) 40 #define AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(x) AMDGPU_GET_REG_FIELD(x, 4, 4) 41 #define AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(x) AMDGPU_GET_REG_FIELD(x, 5, 5) 42 #define AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(x) AMDGPU_GET_REG_FIELD(x, 6, 6) 43 #define AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(x) AMDGPU_GET_REG_FIELD(x, 7, 7) 44 #define AMDGPU_RAS_GPU_ERR_SOCKET_ID(x) AMDGPU_GET_REG_FIELD(x, 10, 8) 45 #define AMDGPU_RAS_GPU_ERR_AID_ID(x) AMDGPU_GET_REG_FIELD(x, 12, 11) 46 #define AMDGPU_RAS_GPU_ERR_HBM_ID(x) AMDGPU_GET_REG_FIELD(x, 13, 13) 47 #define AMDGPU_RAS_GPU_ERR_BOOT_STATUS(x) AMDGPU_GET_REG_FIELD(x, 31, 31) 48 49 #define AMDGPU_RAS_BOOT_STATUS_POLLING_LIMIT 1000 50 #define AMDGPU_RAS_BOOT_STEADY_STATUS 0xBA 51 #define AMDGPU_RAS_BOOT_STATUS_MASK 0xFF 52 #define AMDGPU_RAS_BOOT_SUCEESS 0x80000000 53 54 #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS (0x1 << 0) 55 /* position of instance value in sub_block_index of 56 * ta_ras_trigger_error_input, the sub block uses lower 12 bits 57 */ 58 #define AMDGPU_RAS_INST_MASK 0xfffff000 59 #define AMDGPU_RAS_INST_SHIFT 0xc 60 61 #define AMDGPU_RAS_FEATURES_SOCKETID_SHIFT 29 62 #define AMDGPU_RAS_FEATURES_SOCKETID_MASK 0xe0000000 63 64 /* The high three bits indicates socketid */ 65 #define AMDGPU_RAS_GET_FEATURES(val) ((val) & ~AMDGPU_RAS_FEATURES_SOCKETID_MASK) 66 67 enum amdgpu_ras_block { 68 AMDGPU_RAS_BLOCK__UMC = 0, 69 AMDGPU_RAS_BLOCK__SDMA, 70 AMDGPU_RAS_BLOCK__GFX, 71 AMDGPU_RAS_BLOCK__MMHUB, 72 AMDGPU_RAS_BLOCK__ATHUB, 73 AMDGPU_RAS_BLOCK__PCIE_BIF, 74 AMDGPU_RAS_BLOCK__HDP, 75 AMDGPU_RAS_BLOCK__XGMI_WAFL, 76 AMDGPU_RAS_BLOCK__DF, 77 AMDGPU_RAS_BLOCK__SMN, 78 AMDGPU_RAS_BLOCK__SEM, 79 AMDGPU_RAS_BLOCK__MP0, 80 AMDGPU_RAS_BLOCK__MP1, 81 AMDGPU_RAS_BLOCK__FUSE, 82 AMDGPU_RAS_BLOCK__MCA, 83 AMDGPU_RAS_BLOCK__VCN, 84 AMDGPU_RAS_BLOCK__JPEG, 85 AMDGPU_RAS_BLOCK__IH, 86 AMDGPU_RAS_BLOCK__MPIO, 87 88 AMDGPU_RAS_BLOCK__LAST 89 }; 90 91 enum amdgpu_ras_mca_block { 92 AMDGPU_RAS_MCA_BLOCK__MP0 = 0, 93 AMDGPU_RAS_MCA_BLOCK__MP1, 94 AMDGPU_RAS_MCA_BLOCK__MPIO, 95 AMDGPU_RAS_MCA_BLOCK__IOHC, 96 97 AMDGPU_RAS_MCA_BLOCK__LAST 98 }; 99 100 #define AMDGPU_RAS_BLOCK_COUNT AMDGPU_RAS_BLOCK__LAST 101 #define AMDGPU_RAS_MCA_BLOCK_COUNT AMDGPU_RAS_MCA_BLOCK__LAST 102 #define AMDGPU_RAS_BLOCK_MASK ((1ULL << AMDGPU_RAS_BLOCK_COUNT) - 1) 103 104 enum amdgpu_ras_gfx_subblock { 105 /* CPC */ 106 AMDGPU_RAS_BLOCK__GFX_CPC_INDEX_START = 0, 107 AMDGPU_RAS_BLOCK__GFX_CPC_SCRATCH = 108 AMDGPU_RAS_BLOCK__GFX_CPC_INDEX_START, 109 AMDGPU_RAS_BLOCK__GFX_CPC_UCODE, 110 AMDGPU_RAS_BLOCK__GFX_DC_STATE_ME1, 111 AMDGPU_RAS_BLOCK__GFX_DC_CSINVOC_ME1, 112 AMDGPU_RAS_BLOCK__GFX_DC_RESTORE_ME1, 113 AMDGPU_RAS_BLOCK__GFX_DC_STATE_ME2, 114 AMDGPU_RAS_BLOCK__GFX_DC_CSINVOC_ME2, 115 AMDGPU_RAS_BLOCK__GFX_DC_RESTORE_ME2, 116 AMDGPU_RAS_BLOCK__GFX_CPC_INDEX_END = 117 AMDGPU_RAS_BLOCK__GFX_DC_RESTORE_ME2, 118 /* CPF */ 119 AMDGPU_RAS_BLOCK__GFX_CPF_INDEX_START, 120 AMDGPU_RAS_BLOCK__GFX_CPF_ROQ_ME2 = 121 AMDGPU_RAS_BLOCK__GFX_CPF_INDEX_START, 122 AMDGPU_RAS_BLOCK__GFX_CPF_ROQ_ME1, 123 AMDGPU_RAS_BLOCK__GFX_CPF_TAG, 124 AMDGPU_RAS_BLOCK__GFX_CPF_INDEX_END = AMDGPU_RAS_BLOCK__GFX_CPF_TAG, 125 /* CPG */ 126 AMDGPU_RAS_BLOCK__GFX_CPG_INDEX_START, 127 AMDGPU_RAS_BLOCK__GFX_CPG_DMA_ROQ = 128 AMDGPU_RAS_BLOCK__GFX_CPG_INDEX_START, 129 AMDGPU_RAS_BLOCK__GFX_CPG_DMA_TAG, 130 AMDGPU_RAS_BLOCK__GFX_CPG_TAG, 131 AMDGPU_RAS_BLOCK__GFX_CPG_INDEX_END = AMDGPU_RAS_BLOCK__GFX_CPG_TAG, 132 /* GDS */ 133 AMDGPU_RAS_BLOCK__GFX_GDS_INDEX_START, 134 AMDGPU_RAS_BLOCK__GFX_GDS_MEM = AMDGPU_RAS_BLOCK__GFX_GDS_INDEX_START, 135 AMDGPU_RAS_BLOCK__GFX_GDS_INPUT_QUEUE, 136 AMDGPU_RAS_BLOCK__GFX_GDS_OA_PHY_CMD_RAM_MEM, 137 AMDGPU_RAS_BLOCK__GFX_GDS_OA_PHY_DATA_RAM_MEM, 138 AMDGPU_RAS_BLOCK__GFX_GDS_OA_PIPE_MEM, 139 AMDGPU_RAS_BLOCK__GFX_GDS_INDEX_END = 140 AMDGPU_RAS_BLOCK__GFX_GDS_OA_PIPE_MEM, 141 /* SPI */ 142 AMDGPU_RAS_BLOCK__GFX_SPI_SR_MEM, 143 /* SQ */ 144 AMDGPU_RAS_BLOCK__GFX_SQ_INDEX_START, 145 AMDGPU_RAS_BLOCK__GFX_SQ_SGPR = AMDGPU_RAS_BLOCK__GFX_SQ_INDEX_START, 146 AMDGPU_RAS_BLOCK__GFX_SQ_LDS_D, 147 AMDGPU_RAS_BLOCK__GFX_SQ_LDS_I, 148 AMDGPU_RAS_BLOCK__GFX_SQ_VGPR, 149 AMDGPU_RAS_BLOCK__GFX_SQ_INDEX_END = AMDGPU_RAS_BLOCK__GFX_SQ_VGPR, 150 /* SQC (3 ranges) */ 151 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX_START, 152 /* SQC range 0 */ 153 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX0_START = 154 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX_START, 155 AMDGPU_RAS_BLOCK__GFX_SQC_INST_UTCL1_LFIFO = 156 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX0_START, 157 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_CU0_WRITE_DATA_BUF, 158 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_CU0_UTCL1_LFIFO, 159 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_CU1_WRITE_DATA_BUF, 160 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_CU1_UTCL1_LFIFO, 161 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_CU2_WRITE_DATA_BUF, 162 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_CU2_UTCL1_LFIFO, 163 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX0_END = 164 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_CU2_UTCL1_LFIFO, 165 /* SQC range 1 */ 166 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX1_START, 167 AMDGPU_RAS_BLOCK__GFX_SQC_INST_BANKA_TAG_RAM = 168 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX1_START, 169 AMDGPU_RAS_BLOCK__GFX_SQC_INST_BANKA_UTCL1_MISS_FIFO, 170 AMDGPU_RAS_BLOCK__GFX_SQC_INST_BANKA_MISS_FIFO, 171 AMDGPU_RAS_BLOCK__GFX_SQC_INST_BANKA_BANK_RAM, 172 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKA_TAG_RAM, 173 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKA_HIT_FIFO, 174 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKA_MISS_FIFO, 175 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKA_DIRTY_BIT_RAM, 176 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKA_BANK_RAM, 177 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX1_END = 178 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKA_BANK_RAM, 179 /* SQC range 2 */ 180 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX2_START, 181 AMDGPU_RAS_BLOCK__GFX_SQC_INST_BANKB_TAG_RAM = 182 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX2_START, 183 AMDGPU_RAS_BLOCK__GFX_SQC_INST_BANKB_UTCL1_MISS_FIFO, 184 AMDGPU_RAS_BLOCK__GFX_SQC_INST_BANKB_MISS_FIFO, 185 AMDGPU_RAS_BLOCK__GFX_SQC_INST_BANKB_BANK_RAM, 186 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKB_TAG_RAM, 187 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKB_HIT_FIFO, 188 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKB_MISS_FIFO, 189 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKB_DIRTY_BIT_RAM, 190 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKB_BANK_RAM, 191 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX2_END = 192 AMDGPU_RAS_BLOCK__GFX_SQC_DATA_BANKB_BANK_RAM, 193 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX_END = 194 AMDGPU_RAS_BLOCK__GFX_SQC_INDEX2_END, 195 /* TA */ 196 AMDGPU_RAS_BLOCK__GFX_TA_INDEX_START, 197 AMDGPU_RAS_BLOCK__GFX_TA_FS_DFIFO = 198 AMDGPU_RAS_BLOCK__GFX_TA_INDEX_START, 199 AMDGPU_RAS_BLOCK__GFX_TA_FS_AFIFO, 200 AMDGPU_RAS_BLOCK__GFX_TA_FL_LFIFO, 201 AMDGPU_RAS_BLOCK__GFX_TA_FX_LFIFO, 202 AMDGPU_RAS_BLOCK__GFX_TA_FS_CFIFO, 203 AMDGPU_RAS_BLOCK__GFX_TA_INDEX_END = AMDGPU_RAS_BLOCK__GFX_TA_FS_CFIFO, 204 /* TCA */ 205 AMDGPU_RAS_BLOCK__GFX_TCA_INDEX_START, 206 AMDGPU_RAS_BLOCK__GFX_TCA_HOLE_FIFO = 207 AMDGPU_RAS_BLOCK__GFX_TCA_INDEX_START, 208 AMDGPU_RAS_BLOCK__GFX_TCA_REQ_FIFO, 209 AMDGPU_RAS_BLOCK__GFX_TCA_INDEX_END = 210 AMDGPU_RAS_BLOCK__GFX_TCA_REQ_FIFO, 211 /* TCC (5 sub-ranges) */ 212 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX_START, 213 /* TCC range 0 */ 214 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX0_START = 215 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX_START, 216 AMDGPU_RAS_BLOCK__GFX_TCC_CACHE_DATA = 217 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX0_START, 218 AMDGPU_RAS_BLOCK__GFX_TCC_CACHE_DATA_BANK_0_1, 219 AMDGPU_RAS_BLOCK__GFX_TCC_CACHE_DATA_BANK_1_0, 220 AMDGPU_RAS_BLOCK__GFX_TCC_CACHE_DATA_BANK_1_1, 221 AMDGPU_RAS_BLOCK__GFX_TCC_CACHE_DIRTY_BANK_0, 222 AMDGPU_RAS_BLOCK__GFX_TCC_CACHE_DIRTY_BANK_1, 223 AMDGPU_RAS_BLOCK__GFX_TCC_HIGH_RATE_TAG, 224 AMDGPU_RAS_BLOCK__GFX_TCC_LOW_RATE_TAG, 225 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX0_END = 226 AMDGPU_RAS_BLOCK__GFX_TCC_LOW_RATE_TAG, 227 /* TCC range 1 */ 228 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX1_START, 229 AMDGPU_RAS_BLOCK__GFX_TCC_IN_USE_DEC = 230 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX1_START, 231 AMDGPU_RAS_BLOCK__GFX_TCC_IN_USE_TRANSFER, 232 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX1_END = 233 AMDGPU_RAS_BLOCK__GFX_TCC_IN_USE_TRANSFER, 234 /* TCC range 2 */ 235 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX2_START, 236 AMDGPU_RAS_BLOCK__GFX_TCC_RETURN_DATA = 237 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX2_START, 238 AMDGPU_RAS_BLOCK__GFX_TCC_RETURN_CONTROL, 239 AMDGPU_RAS_BLOCK__GFX_TCC_UC_ATOMIC_FIFO, 240 AMDGPU_RAS_BLOCK__GFX_TCC_WRITE_RETURN, 241 AMDGPU_RAS_BLOCK__GFX_TCC_WRITE_CACHE_READ, 242 AMDGPU_RAS_BLOCK__GFX_TCC_SRC_FIFO, 243 AMDGPU_RAS_BLOCK__GFX_TCC_SRC_FIFO_NEXT_RAM, 244 AMDGPU_RAS_BLOCK__GFX_TCC_CACHE_TAG_PROBE_FIFO, 245 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX2_END = 246 AMDGPU_RAS_BLOCK__GFX_TCC_CACHE_TAG_PROBE_FIFO, 247 /* TCC range 3 */ 248 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX3_START, 249 AMDGPU_RAS_BLOCK__GFX_TCC_LATENCY_FIFO = 250 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX3_START, 251 AMDGPU_RAS_BLOCK__GFX_TCC_LATENCY_FIFO_NEXT_RAM, 252 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX3_END = 253 AMDGPU_RAS_BLOCK__GFX_TCC_LATENCY_FIFO_NEXT_RAM, 254 /* TCC range 4 */ 255 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX4_START, 256 AMDGPU_RAS_BLOCK__GFX_TCC_WRRET_TAG_WRITE_RETURN = 257 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX4_START, 258 AMDGPU_RAS_BLOCK__GFX_TCC_ATOMIC_RETURN_BUFFER, 259 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX4_END = 260 AMDGPU_RAS_BLOCK__GFX_TCC_ATOMIC_RETURN_BUFFER, 261 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX_END = 262 AMDGPU_RAS_BLOCK__GFX_TCC_INDEX4_END, 263 /* TCI */ 264 AMDGPU_RAS_BLOCK__GFX_TCI_WRITE_RAM, 265 /* TCP */ 266 AMDGPU_RAS_BLOCK__GFX_TCP_INDEX_START, 267 AMDGPU_RAS_BLOCK__GFX_TCP_CACHE_RAM = 268 AMDGPU_RAS_BLOCK__GFX_TCP_INDEX_START, 269 AMDGPU_RAS_BLOCK__GFX_TCP_LFIFO_RAM, 270 AMDGPU_RAS_BLOCK__GFX_TCP_CMD_FIFO, 271 AMDGPU_RAS_BLOCK__GFX_TCP_VM_FIFO, 272 AMDGPU_RAS_BLOCK__GFX_TCP_DB_RAM, 273 AMDGPU_RAS_BLOCK__GFX_TCP_UTCL1_LFIFO0, 274 AMDGPU_RAS_BLOCK__GFX_TCP_UTCL1_LFIFO1, 275 AMDGPU_RAS_BLOCK__GFX_TCP_INDEX_END = 276 AMDGPU_RAS_BLOCK__GFX_TCP_UTCL1_LFIFO1, 277 /* TD */ 278 AMDGPU_RAS_BLOCK__GFX_TD_INDEX_START, 279 AMDGPU_RAS_BLOCK__GFX_TD_SS_FIFO_LO = 280 AMDGPU_RAS_BLOCK__GFX_TD_INDEX_START, 281 AMDGPU_RAS_BLOCK__GFX_TD_SS_FIFO_HI, 282 AMDGPU_RAS_BLOCK__GFX_TD_CS_FIFO, 283 AMDGPU_RAS_BLOCK__GFX_TD_INDEX_END = AMDGPU_RAS_BLOCK__GFX_TD_CS_FIFO, 284 /* EA (3 sub-ranges) */ 285 AMDGPU_RAS_BLOCK__GFX_EA_INDEX_START, 286 /* EA range 0 */ 287 AMDGPU_RAS_BLOCK__GFX_EA_INDEX0_START = 288 AMDGPU_RAS_BLOCK__GFX_EA_INDEX_START, 289 AMDGPU_RAS_BLOCK__GFX_EA_DRAMRD_CMDMEM = 290 AMDGPU_RAS_BLOCK__GFX_EA_INDEX0_START, 291 AMDGPU_RAS_BLOCK__GFX_EA_DRAMWR_CMDMEM, 292 AMDGPU_RAS_BLOCK__GFX_EA_DRAMWR_DATAMEM, 293 AMDGPU_RAS_BLOCK__GFX_EA_RRET_TAGMEM, 294 AMDGPU_RAS_BLOCK__GFX_EA_WRET_TAGMEM, 295 AMDGPU_RAS_BLOCK__GFX_EA_GMIRD_CMDMEM, 296 AMDGPU_RAS_BLOCK__GFX_EA_GMIWR_CMDMEM, 297 AMDGPU_RAS_BLOCK__GFX_EA_GMIWR_DATAMEM, 298 AMDGPU_RAS_BLOCK__GFX_EA_INDEX0_END = 299 AMDGPU_RAS_BLOCK__GFX_EA_GMIWR_DATAMEM, 300 /* EA range 1 */ 301 AMDGPU_RAS_BLOCK__GFX_EA_INDEX1_START, 302 AMDGPU_RAS_BLOCK__GFX_EA_DRAMRD_PAGEMEM = 303 AMDGPU_RAS_BLOCK__GFX_EA_INDEX1_START, 304 AMDGPU_RAS_BLOCK__GFX_EA_DRAMWR_PAGEMEM, 305 AMDGPU_RAS_BLOCK__GFX_EA_IORD_CMDMEM, 306 AMDGPU_RAS_BLOCK__GFX_EA_IOWR_CMDMEM, 307 AMDGPU_RAS_BLOCK__GFX_EA_IOWR_DATAMEM, 308 AMDGPU_RAS_BLOCK__GFX_EA_GMIRD_PAGEMEM, 309 AMDGPU_RAS_BLOCK__GFX_EA_GMIWR_PAGEMEM, 310 AMDGPU_RAS_BLOCK__GFX_EA_INDEX1_END = 311 AMDGPU_RAS_BLOCK__GFX_EA_GMIWR_PAGEMEM, 312 /* EA range 2 */ 313 AMDGPU_RAS_BLOCK__GFX_EA_INDEX2_START, 314 AMDGPU_RAS_BLOCK__GFX_EA_MAM_D0MEM = 315 AMDGPU_RAS_BLOCK__GFX_EA_INDEX2_START, 316 AMDGPU_RAS_BLOCK__GFX_EA_MAM_D1MEM, 317 AMDGPU_RAS_BLOCK__GFX_EA_MAM_D2MEM, 318 AMDGPU_RAS_BLOCK__GFX_EA_MAM_D3MEM, 319 AMDGPU_RAS_BLOCK__GFX_EA_INDEX2_END = 320 AMDGPU_RAS_BLOCK__GFX_EA_MAM_D3MEM, 321 AMDGPU_RAS_BLOCK__GFX_EA_INDEX_END = 322 AMDGPU_RAS_BLOCK__GFX_EA_INDEX2_END, 323 /* UTC VM L2 bank */ 324 AMDGPU_RAS_BLOCK__UTC_VML2_BANK_CACHE, 325 /* UTC VM walker */ 326 AMDGPU_RAS_BLOCK__UTC_VML2_WALKER, 327 /* UTC ATC L2 2MB cache */ 328 AMDGPU_RAS_BLOCK__UTC_ATCL2_CACHE_2M_BANK, 329 /* UTC ATC L2 4KB cache */ 330 AMDGPU_RAS_BLOCK__UTC_ATCL2_CACHE_4K_BANK, 331 AMDGPU_RAS_BLOCK__GFX_MAX 332 }; 333 334 enum amdgpu_ras_error_type { 335 AMDGPU_RAS_ERROR__NONE = 0, 336 AMDGPU_RAS_ERROR__PARITY = 1, 337 AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE = 2, 338 AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE = 4, 339 AMDGPU_RAS_ERROR__POISON = 8, 340 }; 341 342 enum amdgpu_ras_ret { 343 AMDGPU_RAS_SUCCESS = 0, 344 AMDGPU_RAS_FAIL, 345 AMDGPU_RAS_UE, 346 AMDGPU_RAS_CE, 347 AMDGPU_RAS_PT, 348 }; 349 350 enum amdgpu_ras_error_query_mode { 351 AMDGPU_RAS_INVALID_ERROR_QUERY = 0, 352 AMDGPU_RAS_DIRECT_ERROR_QUERY = 1, 353 AMDGPU_RAS_FIRMWARE_ERROR_QUERY = 2, 354 }; 355 356 /* ras error status reisger fields */ 357 #define ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 358 #define ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L 359 #define ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 360 #define ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L 361 #define ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 362 #define ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L 363 #define ERR_STATUS__ERR_CNT__SHIFT 0x17 364 #define ERR_STATUS__ERR_CNT_MASK 0x03800000L 365 366 #define AMDGPU_RAS_REG_ENTRY(ip, inst, reg_lo, reg_hi) \ 367 ip##_HWIP, inst, reg_lo##_BASE_IDX, reg_lo, reg_hi##_BASE_IDX, reg_hi 368 369 #define AMDGPU_RAS_REG_ENTRY_OFFSET(hwip, ip_inst, segment, reg) \ 370 (adev->reg_offset[hwip][ip_inst][segment] + (reg)) 371 372 #define AMDGPU_RAS_ERR_INFO_VALID (1 << 0) 373 #define AMDGPU_RAS_ERR_STATUS_VALID (1 << 1) 374 #define AMDGPU_RAS_ERR_ADDRESS_VALID (1 << 2) 375 376 #define AMDGPU_RAS_GPU_RESET_MODE2_RESET (0x1 << 0) 377 #define AMDGPU_RAS_GPU_RESET_MODE1_RESET (0x1 << 1) 378 379 struct amdgpu_ras_err_status_reg_entry { 380 uint32_t hwip; 381 uint32_t ip_inst; 382 uint32_t seg_lo; 383 uint32_t reg_lo; 384 uint32_t seg_hi; 385 uint32_t reg_hi; 386 uint32_t reg_inst; 387 uint32_t flags; 388 const char *block_name; 389 }; 390 391 struct amdgpu_ras_memory_id_entry { 392 uint32_t memory_id; 393 const char *name; 394 }; 395 396 struct ras_common_if { 397 enum amdgpu_ras_block block; 398 enum amdgpu_ras_error_type type; 399 uint32_t sub_block_index; 400 char name[32]; 401 }; 402 403 #define MAX_UMC_CHANNEL_NUM 32 404 405 struct ecc_info_per_ch { 406 uint16_t ce_count_lo_chip; 407 uint16_t ce_count_hi_chip; 408 uint64_t mca_umc_status; 409 uint64_t mca_umc_addr; 410 uint64_t mca_ceumc_addr; 411 }; 412 413 struct umc_ecc_info { 414 struct ecc_info_per_ch ecc[MAX_UMC_CHANNEL_NUM]; 415 416 /* Determine smu ecctable whether support 417 * record correctable error address 418 */ 419 int record_ce_addr_supported; 420 }; 421 422 struct amdgpu_ras { 423 /* ras infrastructure */ 424 /* for ras itself. */ 425 uint32_t features; 426 uint32_t schema; 427 struct list_head head; 428 /* sysfs */ 429 struct device_attribute features_attr; 430 struct device_attribute version_attr; 431 struct device_attribute schema_attr; 432 struct bin_attribute badpages_attr; 433 struct dentry *de_ras_eeprom_table; 434 /* block array */ 435 struct ras_manager *objs; 436 437 /* gpu recovery */ 438 struct work_struct recovery_work; 439 atomic_t in_recovery; 440 struct amdgpu_device *adev; 441 /* error handler data */ 442 struct ras_err_handler_data *eh_data; 443 struct mutex recovery_lock; 444 445 uint32_t flags; 446 bool reboot; 447 struct amdgpu_ras_eeprom_control eeprom_control; 448 449 bool error_query_ready; 450 451 /* bad page count threshold */ 452 uint32_t bad_page_cnt_threshold; 453 454 /* disable ras error count harvest in recovery */ 455 bool disable_ras_err_cnt_harvest; 456 457 /* is poison mode supported */ 458 bool poison_supported; 459 460 /* RAS count errors delayed work */ 461 struct delayed_work ras_counte_delay_work; 462 atomic_t ras_ue_count; 463 atomic_t ras_ce_count; 464 465 /* record umc error info queried from smu */ 466 struct umc_ecc_info umc_ecc; 467 468 /* Indicates smu whether need update bad channel info */ 469 bool update_channel_flag; 470 /* Record status of smu mca debug mode */ 471 bool is_aca_debug_mode; 472 473 /* Record special requirements of gpu reset caller */ 474 uint32_t gpu_reset_flags; 475 476 struct task_struct *page_retirement_thread; 477 wait_queue_head_t page_retirement_wq; 478 struct mutex page_retirement_lock; 479 atomic_t page_retirement_req_cnt; 480 /* Fatal error detected flag */ 481 atomic_t fed; 482 }; 483 484 struct ras_fs_data { 485 char sysfs_name[48]; 486 char debugfs_name[32]; 487 }; 488 489 struct ras_err_addr { 490 struct list_head node; 491 uint64_t err_status; 492 uint64_t err_ipid; 493 uint64_t err_addr; 494 }; 495 496 struct ras_err_info { 497 struct amdgpu_smuio_mcm_config_info mcm_info; 498 u64 ce_count; 499 u64 ue_count; 500 u64 de_count; 501 struct list_head err_addr_list; 502 }; 503 504 struct ras_err_node { 505 struct list_head node; 506 struct ras_err_info err_info; 507 }; 508 509 struct ras_err_data { 510 unsigned long ue_count; 511 unsigned long ce_count; 512 unsigned long de_count; 513 unsigned long err_addr_cnt; 514 struct eeprom_table_record *err_addr; 515 u32 err_list_count; 516 struct list_head err_node_list; 517 }; 518 519 #define for_each_ras_error(err_node, err_data) \ 520 list_for_each_entry(err_node, &(err_data)->err_node_list, node) 521 522 struct ras_err_handler_data { 523 /* point to bad page records array */ 524 struct eeprom_table_record *bps; 525 /* the count of entries */ 526 int count; 527 /* the space can place new entries */ 528 int space_left; 529 }; 530 531 typedef int (*ras_ih_cb)(struct amdgpu_device *adev, 532 void *err_data, 533 struct amdgpu_iv_entry *entry); 534 535 struct ras_ih_data { 536 /* interrupt bottom half */ 537 struct work_struct ih_work; 538 int inuse; 539 /* IP callback */ 540 ras_ih_cb cb; 541 /* full of entries */ 542 unsigned char *ring; 543 unsigned int ring_size; 544 unsigned int element_size; 545 unsigned int aligned_element_size; 546 unsigned int rptr; 547 unsigned int wptr; 548 }; 549 550 struct ras_manager { 551 struct ras_common_if head; 552 /* reference count */ 553 int use; 554 /* ras block link */ 555 struct list_head node; 556 /* the device */ 557 struct amdgpu_device *adev; 558 /* sysfs */ 559 struct device_attribute sysfs_attr; 560 int attr_inuse; 561 562 /* fs node name */ 563 struct ras_fs_data fs_data; 564 565 /* IH data */ 566 struct ras_ih_data ih_data; 567 568 struct ras_err_data err_data; 569 570 struct aca_handle aca_handle; 571 }; 572 573 struct ras_badpage { 574 unsigned int bp; 575 unsigned int size; 576 unsigned int flags; 577 }; 578 579 /* interfaces for IP */ 580 struct ras_fs_if { 581 struct ras_common_if head; 582 const char* sysfs_name; 583 char debugfs_name[32]; 584 }; 585 586 struct ras_query_if { 587 struct ras_common_if head; 588 unsigned long ue_count; 589 unsigned long ce_count; 590 unsigned long de_count; 591 }; 592 593 struct ras_inject_if { 594 struct ras_common_if head; 595 uint64_t address; 596 uint64_t value; 597 uint32_t instance_mask; 598 }; 599 600 struct ras_cure_if { 601 struct ras_common_if head; 602 uint64_t address; 603 }; 604 605 struct ras_ih_if { 606 struct ras_common_if head; 607 ras_ih_cb cb; 608 }; 609 610 struct ras_dispatch_if { 611 struct ras_common_if head; 612 struct amdgpu_iv_entry *entry; 613 }; 614 615 struct ras_debug_if { 616 union { 617 struct ras_common_if head; 618 struct ras_inject_if inject; 619 }; 620 int op; 621 }; 622 623 struct amdgpu_ras_block_object { 624 struct ras_common_if ras_comm; 625 626 int (*ras_block_match)(struct amdgpu_ras_block_object *block_obj, 627 enum amdgpu_ras_block block, uint32_t sub_block_index); 628 int (*ras_late_init)(struct amdgpu_device *adev, struct ras_common_if *ras_block); 629 void (*ras_fini)(struct amdgpu_device *adev, struct ras_common_if *ras_block); 630 ras_ih_cb ras_cb; 631 const struct amdgpu_ras_block_hw_ops *hw_ops; 632 }; 633 634 struct amdgpu_ras_block_hw_ops { 635 int (*ras_error_inject)(struct amdgpu_device *adev, 636 void *inject_if, uint32_t instance_mask); 637 void (*query_ras_error_count)(struct amdgpu_device *adev, void *ras_error_status); 638 void (*query_ras_error_status)(struct amdgpu_device *adev); 639 void (*query_ras_error_address)(struct amdgpu_device *adev, void *ras_error_status); 640 void (*reset_ras_error_count)(struct amdgpu_device *adev); 641 void (*reset_ras_error_status)(struct amdgpu_device *adev); 642 bool (*query_poison_status)(struct amdgpu_device *adev); 643 bool (*handle_poison_consumption)(struct amdgpu_device *adev); 644 }; 645 646 /* work flow 647 * vbios 648 * 1: ras feature enable (enabled by default) 649 * psp 650 * 2: ras framework init (in ip_init) 651 * IP 652 * 3: IH add 653 * 4: debugfs/sysfs create 654 * 5: query/inject 655 * 6: debugfs/sysfs remove 656 * 7: IH remove 657 * 8: feature disable 658 */ 659 660 661 int amdgpu_ras_recovery_init(struct amdgpu_device *adev); 662 663 void amdgpu_ras_resume(struct amdgpu_device *adev); 664 void amdgpu_ras_suspend(struct amdgpu_device *adev); 665 666 int amdgpu_ras_query_error_count(struct amdgpu_device *adev, 667 unsigned long *ce_count, 668 unsigned long *ue_count, 669 struct ras_query_if *query_info); 670 671 /* error handling functions */ 672 int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev, 673 struct eeprom_table_record *bps, int pages); 674 675 int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev, 676 unsigned long *new_cnt); 677 678 static inline enum ta_ras_block 679 amdgpu_ras_block_to_ta(enum amdgpu_ras_block block) { 680 switch (block) { 681 case AMDGPU_RAS_BLOCK__UMC: 682 return TA_RAS_BLOCK__UMC; 683 case AMDGPU_RAS_BLOCK__SDMA: 684 return TA_RAS_BLOCK__SDMA; 685 case AMDGPU_RAS_BLOCK__GFX: 686 return TA_RAS_BLOCK__GFX; 687 case AMDGPU_RAS_BLOCK__MMHUB: 688 return TA_RAS_BLOCK__MMHUB; 689 case AMDGPU_RAS_BLOCK__ATHUB: 690 return TA_RAS_BLOCK__ATHUB; 691 case AMDGPU_RAS_BLOCK__PCIE_BIF: 692 return TA_RAS_BLOCK__PCIE_BIF; 693 case AMDGPU_RAS_BLOCK__HDP: 694 return TA_RAS_BLOCK__HDP; 695 case AMDGPU_RAS_BLOCK__XGMI_WAFL: 696 return TA_RAS_BLOCK__XGMI_WAFL; 697 case AMDGPU_RAS_BLOCK__DF: 698 return TA_RAS_BLOCK__DF; 699 case AMDGPU_RAS_BLOCK__SMN: 700 return TA_RAS_BLOCK__SMN; 701 case AMDGPU_RAS_BLOCK__SEM: 702 return TA_RAS_BLOCK__SEM; 703 case AMDGPU_RAS_BLOCK__MP0: 704 return TA_RAS_BLOCK__MP0; 705 case AMDGPU_RAS_BLOCK__MP1: 706 return TA_RAS_BLOCK__MP1; 707 case AMDGPU_RAS_BLOCK__FUSE: 708 return TA_RAS_BLOCK__FUSE; 709 case AMDGPU_RAS_BLOCK__MCA: 710 return TA_RAS_BLOCK__MCA; 711 case AMDGPU_RAS_BLOCK__VCN: 712 return TA_RAS_BLOCK__VCN; 713 case AMDGPU_RAS_BLOCK__JPEG: 714 return TA_RAS_BLOCK__JPEG; 715 default: 716 WARN_ONCE(1, "RAS ERROR: unexpected block id %d\n", block); 717 return TA_RAS_BLOCK__UMC; 718 } 719 } 720 721 static inline enum ta_ras_error_type 722 amdgpu_ras_error_to_ta(enum amdgpu_ras_error_type error) { 723 switch (error) { 724 case AMDGPU_RAS_ERROR__NONE: 725 return TA_RAS_ERROR__NONE; 726 case AMDGPU_RAS_ERROR__PARITY: 727 return TA_RAS_ERROR__PARITY; 728 case AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE: 729 return TA_RAS_ERROR__SINGLE_CORRECTABLE; 730 case AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE: 731 return TA_RAS_ERROR__MULTI_UNCORRECTABLE; 732 case AMDGPU_RAS_ERROR__POISON: 733 return TA_RAS_ERROR__POISON; 734 default: 735 WARN_ONCE(1, "RAS ERROR: unexpected error type %d\n", error); 736 return TA_RAS_ERROR__NONE; 737 } 738 } 739 740 /* called in ip_init and ip_fini */ 741 int amdgpu_ras_init(struct amdgpu_device *adev); 742 int amdgpu_ras_late_init(struct amdgpu_device *adev); 743 int amdgpu_ras_fini(struct amdgpu_device *adev); 744 int amdgpu_ras_pre_fini(struct amdgpu_device *adev); 745 746 int amdgpu_ras_block_late_init(struct amdgpu_device *adev, 747 struct ras_common_if *ras_block); 748 749 void amdgpu_ras_block_late_fini(struct amdgpu_device *adev, 750 struct ras_common_if *ras_block); 751 752 int amdgpu_ras_feature_enable(struct amdgpu_device *adev, 753 struct ras_common_if *head, bool enable); 754 755 int amdgpu_ras_feature_enable_on_boot(struct amdgpu_device *adev, 756 struct ras_common_if *head, bool enable); 757 758 int amdgpu_ras_sysfs_create(struct amdgpu_device *adev, 759 struct ras_common_if *head); 760 761 int amdgpu_ras_sysfs_remove(struct amdgpu_device *adev, 762 struct ras_common_if *head); 763 764 void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev); 765 766 int amdgpu_ras_query_error_status(struct amdgpu_device *adev, 767 struct ras_query_if *info); 768 769 int amdgpu_ras_reset_error_count(struct amdgpu_device *adev, 770 enum amdgpu_ras_block block); 771 int amdgpu_ras_reset_error_status(struct amdgpu_device *adev, 772 enum amdgpu_ras_block block); 773 774 int amdgpu_ras_error_inject(struct amdgpu_device *adev, 775 struct ras_inject_if *info); 776 777 int amdgpu_ras_interrupt_add_handler(struct amdgpu_device *adev, 778 struct ras_common_if *head); 779 780 int amdgpu_ras_interrupt_remove_handler(struct amdgpu_device *adev, 781 struct ras_common_if *head); 782 783 int amdgpu_ras_interrupt_dispatch(struct amdgpu_device *adev, 784 struct ras_dispatch_if *info); 785 786 struct ras_manager *amdgpu_ras_find_obj(struct amdgpu_device *adev, 787 struct ras_common_if *head); 788 789 extern atomic_t amdgpu_ras_in_intr; 790 791 static inline bool amdgpu_ras_intr_triggered(void) 792 { 793 return !!atomic_read(&amdgpu_ras_in_intr); 794 } 795 796 static inline void amdgpu_ras_intr_cleared(void) 797 { 798 atomic_set(&amdgpu_ras_in_intr, 0); 799 } 800 801 void amdgpu_ras_global_ras_isr(struct amdgpu_device *adev); 802 803 void amdgpu_ras_set_error_query_ready(struct amdgpu_device *adev, bool ready); 804 805 bool amdgpu_ras_need_emergency_restart(struct amdgpu_device *adev); 806 807 void amdgpu_release_ras_context(struct amdgpu_device *adev); 808 809 int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev); 810 811 const char *get_ras_block_str(struct ras_common_if *ras_block); 812 813 bool amdgpu_ras_is_poison_mode_supported(struct amdgpu_device *adev); 814 815 int amdgpu_ras_is_supported(struct amdgpu_device *adev, unsigned int block); 816 817 int amdgpu_ras_reset_gpu(struct amdgpu_device *adev); 818 819 struct amdgpu_ras* amdgpu_ras_get_context(struct amdgpu_device *adev); 820 821 int amdgpu_ras_set_context(struct amdgpu_device *adev, struct amdgpu_ras *ras_con); 822 823 int amdgpu_ras_set_mca_debug_mode(struct amdgpu_device *adev, bool enable); 824 int amdgpu_ras_set_aca_debug_mode(struct amdgpu_device *adev, bool enable); 825 bool amdgpu_ras_get_aca_debug_mode(struct amdgpu_device *adev); 826 bool amdgpu_ras_get_error_query_mode(struct amdgpu_device *adev, 827 unsigned int *mode); 828 829 int amdgpu_ras_register_ras_block(struct amdgpu_device *adev, 830 struct amdgpu_ras_block_object *ras_block_obj); 831 void amdgpu_ras_interrupt_fatal_error_handler(struct amdgpu_device *adev); 832 void amdgpu_ras_get_error_type_name(uint32_t err_type, char *err_type_name); 833 bool amdgpu_ras_inst_get_memory_id_field(struct amdgpu_device *adev, 834 const struct amdgpu_ras_err_status_reg_entry *reg_entry, 835 uint32_t instance, 836 uint32_t *memory_id); 837 bool amdgpu_ras_inst_get_err_cnt_field(struct amdgpu_device *adev, 838 const struct amdgpu_ras_err_status_reg_entry *reg_entry, 839 uint32_t instance, 840 unsigned long *err_cnt); 841 void amdgpu_ras_inst_query_ras_error_count(struct amdgpu_device *adev, 842 const struct amdgpu_ras_err_status_reg_entry *reg_list, 843 uint32_t reg_list_size, 844 const struct amdgpu_ras_memory_id_entry *mem_list, 845 uint32_t mem_list_size, 846 uint32_t instance, 847 uint32_t err_type, 848 unsigned long *err_count); 849 void amdgpu_ras_inst_reset_ras_error_count(struct amdgpu_device *adev, 850 const struct amdgpu_ras_err_status_reg_entry *reg_list, 851 uint32_t reg_list_size, 852 uint32_t instance); 853 854 int amdgpu_ras_error_data_init(struct ras_err_data *err_data); 855 void amdgpu_ras_error_data_fini(struct ras_err_data *err_data); 856 int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data, 857 struct amdgpu_smuio_mcm_config_info *mcm_info, 858 struct ras_err_addr *err_addr, u64 count); 859 int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data, 860 struct amdgpu_smuio_mcm_config_info *mcm_info, 861 struct ras_err_addr *err_addr, u64 count); 862 int amdgpu_ras_error_statistic_de_count(struct ras_err_data *err_data, 863 struct amdgpu_smuio_mcm_config_info *mcm_info, 864 struct ras_err_addr *err_addr, u64 count); 865 void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32 num_instances); 866 int amdgpu_ras_bind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk, 867 const struct aca_info *aca_info, void *data); 868 int amdgpu_ras_unbind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk); 869 870 ssize_t amdgpu_ras_aca_sysfs_read(struct device *dev, struct device_attribute *attr, 871 struct aca_handle *handle, char *buf, void *data); 872 873 void amdgpu_ras_add_mca_err_addr(struct ras_err_info *err_info, 874 struct ras_err_addr *err_addr); 875 876 void amdgpu_ras_del_mca_err_addr(struct ras_err_info *err_info, 877 struct ras_err_addr *mca_err_addr); 878 879 void amdgpu_ras_set_fed(struct amdgpu_device *adev, bool status); 880 bool amdgpu_ras_get_fed_status(struct amdgpu_device *adev); 881 882 #endif 883