1 /* 2 * Copyright 2016 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_VCN_H__ 25 #define __AMDGPU_VCN_H__ 26 27 #define AMDGPU_VCN_STACK_SIZE (128*1024) 28 #define AMDGPU_VCN_CONTEXT_SIZE (512*1024) 29 30 #define AMDGPU_VCN_FIRMWARE_OFFSET 256 31 #define AMDGPU_VCN_MAX_ENC_RINGS 3 32 33 #define AMDGPU_MAX_VCN_INSTANCES 2 34 35 #define AMDGPU_VCN_HARVEST_VCN0 (1 << 0) 36 #define AMDGPU_VCN_HARVEST_VCN1 (1 << 1) 37 38 #define VCN_DEC_KMD_CMD 0x80000000 39 #define VCN_DEC_CMD_FENCE 0x00000000 40 #define VCN_DEC_CMD_TRAP 0x00000001 41 #define VCN_DEC_CMD_WRITE_REG 0x00000004 42 #define VCN_DEC_CMD_REG_READ_COND_WAIT 0x00000006 43 #define VCN_DEC_CMD_PACKET_START 0x0000000a 44 #define VCN_DEC_CMD_PACKET_END 0x0000000b 45 46 #define VCN_ENC_CMD_NO_OP 0x00000000 47 #define VCN_ENC_CMD_END 0x00000001 48 #define VCN_ENC_CMD_IB 0x00000002 49 #define VCN_ENC_CMD_FENCE 0x00000003 50 #define VCN_ENC_CMD_TRAP 0x00000004 51 #define VCN_ENC_CMD_REG_WRITE 0x0000000b 52 #define VCN_ENC_CMD_REG_WAIT 0x0000000c 53 54 #define VCN_VID_SOC_ADDRESS_2_0 0x1fa00 55 #define VCN_AON_SOC_ADDRESS_2_0 0x1f800 56 #define VCN_VID_IP_ADDRESS_2_0 0x0 57 #define VCN_AON_IP_ADDRESS_2_0 0x30000 58 59 #define RREG32_SOC15_DPG_MODE(ip, inst, reg, mask, sram_sel) \ 60 ({ WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_MASK, mask); \ 61 WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_CTL, \ 62 UVD_DPG_LMA_CTL__MASK_EN_MASK | \ 63 ((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) \ 64 << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT) | \ 65 (sram_sel << UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT)); \ 66 RREG32_SOC15(ip, inst, mmUVD_DPG_LMA_DATA); \ 67 }) 68 69 #define WREG32_SOC15_DPG_MODE(ip, inst, reg, value, mask, sram_sel) \ 70 do { \ 71 WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_DATA, value); \ 72 WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_MASK, mask); \ 73 WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_CTL, \ 74 UVD_DPG_LMA_CTL__READ_WRITE_MASK | \ 75 ((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) \ 76 << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT) | \ 77 (sram_sel << UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT)); \ 78 } while (0) 79 80 #define SOC15_DPG_MODE_OFFSET_2_0(ip, inst, reg) \ 81 ({ \ 82 uint32_t internal_reg_offset, addr; \ 83 bool video_range, aon_range; \ 84 \ 85 addr = (adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg); \ 86 addr <<= 2; \ 87 video_range = ((((0xFFFFF & addr) >= (VCN_VID_SOC_ADDRESS_2_0)) && \ 88 ((0xFFFFF & addr) < ((VCN_VID_SOC_ADDRESS_2_0 + 0x2600))))); \ 89 aon_range = ((((0xFFFFF & addr) >= (VCN_AON_SOC_ADDRESS_2_0)) && \ 90 ((0xFFFFF & addr) < ((VCN_AON_SOC_ADDRESS_2_0 + 0x600))))); \ 91 if (video_range) \ 92 internal_reg_offset = ((0xFFFFF & addr) - (VCN_VID_SOC_ADDRESS_2_0) + \ 93 (VCN_VID_IP_ADDRESS_2_0)); \ 94 else if (aon_range) \ 95 internal_reg_offset = ((0xFFFFF & addr) - (VCN_AON_SOC_ADDRESS_2_0) + \ 96 (VCN_AON_IP_ADDRESS_2_0)); \ 97 else \ 98 internal_reg_offset = (0xFFFFF & addr); \ 99 \ 100 internal_reg_offset >>= 2; \ 101 }) 102 103 #define RREG32_SOC15_DPG_MODE_2_0(offset, mask_en) \ 104 ({ \ 105 WREG32_SOC15(VCN, 0, mmUVD_DPG_LMA_CTL, \ 106 (0x0 << UVD_DPG_LMA_CTL__READ_WRITE__SHIFT | \ 107 mask_en << UVD_DPG_LMA_CTL__MASK_EN__SHIFT | \ 108 offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT)); \ 109 RREG32_SOC15(VCN, 0, mmUVD_DPG_LMA_DATA); \ 110 }) 111 112 #define WREG32_SOC15_DPG_MODE_2_0(offset, value, mask_en, indirect) \ 113 do { \ 114 if (!indirect) { \ 115 WREG32_SOC15(VCN, 0, mmUVD_DPG_LMA_DATA, value); \ 116 WREG32_SOC15(VCN, 0, mmUVD_DPG_LMA_CTL, \ 117 (0x1 << UVD_DPG_LMA_CTL__READ_WRITE__SHIFT | \ 118 mask_en << UVD_DPG_LMA_CTL__MASK_EN__SHIFT | \ 119 offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT)); \ 120 } else { \ 121 *adev->vcn.dpg_sram_curr_addr++ = offset; \ 122 *adev->vcn.dpg_sram_curr_addr++ = value; \ 123 } \ 124 } while (0) 125 126 enum engine_status_constants { 127 UVD_PGFSM_STATUS__UVDM_UVDU_PWR_ON = 0x2AAAA0, 128 UVD_PGFSM_STATUS__UVDM_UVDU_PWR_ON_2_0 = 0xAAAA0, 129 UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON = 0x00000002, 130 UVD_STATUS__UVD_BUSY = 0x00000004, 131 GB_ADDR_CONFIG_DEFAULT = 0x26010011, 132 UVD_STATUS__IDLE = 0x2, 133 UVD_STATUS__BUSY = 0x5, 134 UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF = 0x1, 135 UVD_STATUS__RBC_BUSY = 0x1, 136 UVD_PGFSM_STATUS_UVDJ_PWR_ON = 0, 137 }; 138 139 enum internal_dpg_state { 140 VCN_DPG_STATE__UNPAUSE = 0, 141 VCN_DPG_STATE__PAUSE, 142 }; 143 144 struct dpg_pause_state { 145 enum internal_dpg_state fw_based; 146 enum internal_dpg_state jpeg; 147 }; 148 149 struct amdgpu_vcn_reg{ 150 unsigned data0; 151 unsigned data1; 152 unsigned cmd; 153 unsigned nop; 154 unsigned context_id; 155 unsigned ib_vmid; 156 unsigned ib_bar_low; 157 unsigned ib_bar_high; 158 unsigned ib_size; 159 unsigned gp_scratch8; 160 unsigned scratch9; 161 }; 162 163 struct amdgpu_vcn_inst { 164 struct amdgpu_bo *vcpu_bo; 165 void *cpu_addr; 166 uint64_t gpu_addr; 167 void *saved_bo; 168 struct amdgpu_ring ring_dec; 169 struct amdgpu_ring ring_enc[AMDGPU_VCN_MAX_ENC_RINGS]; 170 struct amdgpu_irq_src irq; 171 struct amdgpu_vcn_reg external; 172 }; 173 174 struct amdgpu_vcn { 175 unsigned fw_version; 176 struct delayed_work idle_work; 177 const struct firmware *fw; /* VCN firmware */ 178 unsigned num_enc_rings; 179 enum amd_powergating_state cur_state; 180 struct dpg_pause_state pause_state; 181 182 bool indirect_sram; 183 struct amdgpu_bo *dpg_sram_bo; 184 void *dpg_sram_cpu_addr; 185 uint64_t dpg_sram_gpu_addr; 186 uint32_t *dpg_sram_curr_addr; 187 188 uint8_t num_vcn_inst; 189 struct amdgpu_vcn_inst inst[AMDGPU_MAX_VCN_INSTANCES]; 190 struct amdgpu_vcn_reg internal; 191 192 unsigned harvest_config; 193 int (*pause_dpg_mode)(struct amdgpu_device *adev, 194 struct dpg_pause_state *new_state); 195 }; 196 197 int amdgpu_vcn_sw_init(struct amdgpu_device *adev); 198 int amdgpu_vcn_sw_fini(struct amdgpu_device *adev); 199 int amdgpu_vcn_suspend(struct amdgpu_device *adev); 200 int amdgpu_vcn_resume(struct amdgpu_device *adev); 201 void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring); 202 void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring); 203 204 int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring); 205 int amdgpu_vcn_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout); 206 207 int amdgpu_vcn_enc_ring_test_ring(struct amdgpu_ring *ring); 208 int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout); 209 210 #endif 211