diff options
author | Dave Airlie <airlied@redhat.com> | 2018-05-16 01:21:51 +0300 |
---|---|---|
committer | Dave Airlie <airlied@redhat.com> | 2018-05-16 01:31:29 +0300 |
commit | 95d2c3e15da613afd53b4b8f2cdb352dc7d12221 (patch) | |
tree | b407c31c3ad2a7e133e61ba36edb31274492fb84 /drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | |
parent | b8a71080ad288eb3fe42f101e64526cdd2823f93 (diff) | |
parent | 8344c53f57057b42a5da87e9557c40fcda18fb7a (diff) | |
download | linux-95d2c3e15da613afd53b4b8f2cdb352dc7d12221.tar.xz |
Merge branch 'drm-next-4.18' of git://people.freedesktop.org/~agd5f/linux into drm-next
Main changes for 4.18. I'd like to do a separate pull for vega20 later
this week or next. Highlights:
- Reserve pre-OS scanout buffer during init for seemless transition from
console to driver
- VEGAM support
- Improved GPU scheduler documentation
- Initial gfxoff support for raven
- SR-IOV fixes
- Default to non-AGP on PowerPC for radeon
- Fine grained clock voltage control for vega10
- Power profiles for vega10
- Further clean up of powerplay/driver interface
- Underlay fixes
- Display link bw updates
- Gamma fixes
- Scatter/Gather display support on CZ/ST
- Misc bug fixes and clean ups
[airlied: fixup v3d vs scheduler API change]
Link: https://patchwork.freedesktop.org/patch/msgid/20180515185450.1113-1-alexander.deucher@amd.com
Signed-off-by: Dave Airlie <airlied@redhat.com>
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c')
-rw-r--r-- | drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 267 |
1 files changed, 185 insertions, 82 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index e5962e61beb5..fc1911834ab5 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -41,7 +41,6 @@ #define GFX9_MEC_HPD_SIZE 2048 #define RLCG_UCODE_LOADING_START_ADDRESS 0x00002000L #define RLC_SAVE_RESTORE_ADDR_STARTING_OFFSET 0x00000000L -#define GFX9_RLC_FORMAT_DIRECT_REG_LIST_LENGTH 34 #define mmPWR_MISC_CNTL_STATUS 0x0183 #define mmPWR_MISC_CNTL_STATUS_BASE_IDX 0 @@ -185,6 +184,30 @@ static const struct soc15_reg_golden golden_settings_gc_9_2_1_vg12[] = SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x01bd9f33, 0x01000000) }; +static const u32 GFX_RLC_SRM_INDEX_CNTL_ADDR_OFFSETS[] = +{ + mmRLC_SRM_INDEX_CNTL_ADDR_0 - mmRLC_SRM_INDEX_CNTL_ADDR_0, + mmRLC_SRM_INDEX_CNTL_ADDR_1 - mmRLC_SRM_INDEX_CNTL_ADDR_0, + mmRLC_SRM_INDEX_CNTL_ADDR_2 - mmRLC_SRM_INDEX_CNTL_ADDR_0, + mmRLC_SRM_INDEX_CNTL_ADDR_3 - mmRLC_SRM_INDEX_CNTL_ADDR_0, + mmRLC_SRM_INDEX_CNTL_ADDR_4 - mmRLC_SRM_INDEX_CNTL_ADDR_0, + mmRLC_SRM_INDEX_CNTL_ADDR_5 - mmRLC_SRM_INDEX_CNTL_ADDR_0, + mmRLC_SRM_INDEX_CNTL_ADDR_6 - mmRLC_SRM_INDEX_CNTL_ADDR_0, + mmRLC_SRM_INDEX_CNTL_ADDR_7 - mmRLC_SRM_INDEX_CNTL_ADDR_0, +}; + +static const u32 GFX_RLC_SRM_INDEX_CNTL_DATA_OFFSETS[] = +{ + mmRLC_SRM_INDEX_CNTL_DATA_0 - mmRLC_SRM_INDEX_CNTL_DATA_0, + mmRLC_SRM_INDEX_CNTL_DATA_1 - mmRLC_SRM_INDEX_CNTL_DATA_0, + mmRLC_SRM_INDEX_CNTL_DATA_2 - mmRLC_SRM_INDEX_CNTL_DATA_0, + mmRLC_SRM_INDEX_CNTL_DATA_3 - mmRLC_SRM_INDEX_CNTL_DATA_0, + mmRLC_SRM_INDEX_CNTL_DATA_4 - mmRLC_SRM_INDEX_CNTL_DATA_0, + mmRLC_SRM_INDEX_CNTL_DATA_5 - mmRLC_SRM_INDEX_CNTL_DATA_0, + mmRLC_SRM_INDEX_CNTL_DATA_6 - mmRLC_SRM_INDEX_CNTL_DATA_0, + mmRLC_SRM_INDEX_CNTL_DATA_7 - mmRLC_SRM_INDEX_CNTL_DATA_0, +}; + #define VEGA10_GB_ADDR_CONFIG_GOLDEN 0x2a114042 #define VEGA12_GB_ADDR_CONFIG_GOLDEN 0x24104041 #define RAVEN_GB_ADDR_CONFIG_GOLDEN 0x24000042 @@ -401,6 +424,27 @@ static void gfx_v9_0_free_microcode(struct amdgpu_device *adev) kfree(adev->gfx.rlc.register_list_format); } +static void gfx_v9_0_init_rlc_ext_microcode(struct amdgpu_device *adev) +{ + const struct rlc_firmware_header_v2_1 *rlc_hdr; + + rlc_hdr = (const struct rlc_firmware_header_v2_1 *)adev->gfx.rlc_fw->data; + adev->gfx.rlc_srlc_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_ucode_ver); + adev->gfx.rlc_srlc_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_feature_ver); + adev->gfx.rlc.save_restore_list_cntl_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_cntl_size_bytes); + adev->gfx.rlc.save_restore_list_cntl = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_cntl_offset_bytes); + adev->gfx.rlc_srlg_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_ucode_ver); + adev->gfx.rlc_srlg_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_feature_ver); + adev->gfx.rlc.save_restore_list_gpm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_gpm_size_bytes); + adev->gfx.rlc.save_restore_list_gpm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_gpm_offset_bytes); + adev->gfx.rlc_srls_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_ucode_ver); + adev->gfx.rlc_srls_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_feature_ver); + adev->gfx.rlc.save_restore_list_srm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_srm_size_bytes); + adev->gfx.rlc.save_restore_list_srm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_srm_offset_bytes); + adev->gfx.rlc.reg_list_format_direct_reg_list_length = + le32_to_cpu(rlc_hdr->reg_list_format_direct_reg_list_length); +} + static int gfx_v9_0_init_microcode(struct amdgpu_device *adev) { const char *chip_name; @@ -412,6 +456,8 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev) const struct rlc_firmware_header_v2_0 *rlc_hdr; unsigned int *tmp = NULL; unsigned int i = 0; + uint16_t version_major; + uint16_t version_minor; DRM_DEBUG("\n"); @@ -468,6 +514,12 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev) goto out; err = amdgpu_ucode_validate(adev->gfx.rlc_fw); rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data; + + version_major = le16_to_cpu(rlc_hdr->header.header_version_major); + version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor); + if (version_major == 2 && version_minor == 1) + adev->gfx.rlc.is_rlc_v2_1 = true; + adev->gfx.rlc_fw_version = le32_to_cpu(rlc_hdr->header.ucode_version); adev->gfx.rlc_feature_version = le32_to_cpu(rlc_hdr->ucode_feature_version); adev->gfx.rlc.save_and_restore_offset = @@ -508,6 +560,9 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev) for (i = 0 ; i < (rlc_hdr->reg_list_size_bytes >> 2); i++) adev->gfx.rlc.register_restore[i] = le32_to_cpu(tmp[i]); + if (adev->gfx.rlc.is_rlc_v2_1) + gfx_v9_0_init_rlc_ext_microcode(adev); + snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name); err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev); if (err) @@ -566,6 +621,26 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev) adev->firmware.fw_size += ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); + if (adev->gfx.rlc.is_rlc_v2_1) { + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL]; + info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL; + info->fw = adev->gfx.rlc_fw; + adev->firmware.fw_size += + ALIGN(adev->gfx.rlc.save_restore_list_cntl_size_bytes, PAGE_SIZE); + + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM]; + info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM; + info->fw = adev->gfx.rlc_fw; + adev->firmware.fw_size += + ALIGN(adev->gfx.rlc.save_restore_list_gpm_size_bytes, PAGE_SIZE); + + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM]; + info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM; + info->fw = adev->gfx.rlc_fw; + adev->firmware.fw_size += + ALIGN(adev->gfx.rlc.save_restore_list_srm_size_bytes, PAGE_SIZE); + } + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1]; info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1; info->fw = adev->gfx.mec_fw; @@ -1600,6 +1675,7 @@ static void gfx_v9_0_gpu_init(struct amdgpu_device *adev) gfx_v9_0_setup_rb(adev); gfx_v9_0_get_cu_info(adev, &adev->gfx.cu_info); + adev->gfx.config.db_debug2 = RREG32_SOC15(GC, 0, mmDB_DEBUG2); /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ @@ -1616,7 +1692,10 @@ static void gfx_v9_0_gpu_init(struct amdgpu_device *adev) tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE, SH_MEM_ALIGNMENT_MODE_UNALIGNED); WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, tmp); - tmp = adev->gmc.shared_aperture_start >> 48; + tmp = REG_SET_FIELD(0, SH_MEM_BASES, PRIVATE_BASE, + (adev->gmc.private_aperture_start >> 48)); + tmp = REG_SET_FIELD(tmp, SH_MEM_BASES, SHARED_BASE, + (adev->gmc.shared_aperture_start >> 48)); WREG32_SOC15(GC, 0, mmSH_MEM_BASES, tmp); } } @@ -1708,55 +1787,42 @@ static void gfx_v9_0_init_csb(struct amdgpu_device *adev) adev->gfx.rlc.clear_state_size); } -static void gfx_v9_0_parse_ind_reg_list(int *register_list_format, +static void gfx_v9_1_parse_ind_reg_list(int *register_list_format, int indirect_offset, int list_size, int *unique_indirect_regs, int *unique_indirect_reg_count, - int max_indirect_reg_count, int *indirect_start_offsets, - int *indirect_start_offsets_count, - int max_indirect_start_offsets_count) + int *indirect_start_offsets_count) { int idx; - bool new_entry = true; for (; indirect_offset < list_size; indirect_offset++) { + indirect_start_offsets[*indirect_start_offsets_count] = indirect_offset; + *indirect_start_offsets_count = *indirect_start_offsets_count + 1; - if (new_entry) { - new_entry = false; - indirect_start_offsets[*indirect_start_offsets_count] = indirect_offset; - *indirect_start_offsets_count = *indirect_start_offsets_count + 1; - BUG_ON(*indirect_start_offsets_count >= max_indirect_start_offsets_count); - } + while (register_list_format[indirect_offset] != 0xFFFFFFFF) { + indirect_offset += 2; - if (register_list_format[indirect_offset] == 0xFFFFFFFF) { - new_entry = true; - continue; - } + /* look for the matching indice */ + for (idx = 0; idx < *unique_indirect_reg_count; idx++) { + if (unique_indirect_regs[idx] == + register_list_format[indirect_offset] || + !unique_indirect_regs[idx]) + break; + } - indirect_offset += 2; + BUG_ON(idx >= *unique_indirect_reg_count); - /* look for the matching indice */ - for (idx = 0; idx < *unique_indirect_reg_count; idx++) { - if (unique_indirect_regs[idx] == - register_list_format[indirect_offset]) - break; - } + if (!unique_indirect_regs[idx]) + unique_indirect_regs[idx] = register_list_format[indirect_offset]; - if (idx >= *unique_indirect_reg_count) { - unique_indirect_regs[*unique_indirect_reg_count] = - register_list_format[indirect_offset]; - idx = *unique_indirect_reg_count; - *unique_indirect_reg_count = *unique_indirect_reg_count + 1; - BUG_ON(*unique_indirect_reg_count >= max_indirect_reg_count); + indirect_offset++; } - - register_list_format[indirect_offset] = idx; } } -static int gfx_v9_0_init_rlc_save_restore_list(struct amdgpu_device *adev) +static int gfx_v9_1_init_rlc_save_restore_list(struct amdgpu_device *adev) { int unique_indirect_regs[] = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}; int unique_indirect_reg_count = 0; @@ -1765,7 +1831,7 @@ static int gfx_v9_0_init_rlc_save_restore_list(struct amdgpu_device *adev) int indirect_start_offsets_count = 0; int list_size = 0; - int i = 0; + int i = 0, j = 0; u32 tmp = 0; u32 *register_list_format = @@ -1776,15 +1842,14 @@ static int gfx_v9_0_init_rlc_save_restore_list(struct amdgpu_device *adev) adev->gfx.rlc.reg_list_format_size_bytes); /* setup unique_indirect_regs array and indirect_start_offsets array */ - gfx_v9_0_parse_ind_reg_list(register_list_format, - GFX9_RLC_FORMAT_DIRECT_REG_LIST_LENGTH, - adev->gfx.rlc.reg_list_format_size_bytes >> 2, - unique_indirect_regs, - &unique_indirect_reg_count, - ARRAY_SIZE(unique_indirect_regs), - indirect_start_offsets, - &indirect_start_offsets_count, - ARRAY_SIZE(indirect_start_offsets)); + unique_indirect_reg_count = ARRAY_SIZE(unique_indirect_regs); + gfx_v9_1_parse_ind_reg_list(register_list_format, + adev->gfx.rlc.reg_list_format_direct_reg_list_length, + adev->gfx.rlc.reg_list_format_size_bytes >> 2, + unique_indirect_regs, + &unique_indirect_reg_count, + indirect_start_offsets, + &indirect_start_offsets_count); /* enable auto inc in case it is disabled */ tmp = RREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_SRM_CNTL)); @@ -1798,19 +1863,37 @@ static int gfx_v9_0_init_rlc_save_restore_list(struct amdgpu_device *adev) WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_SRM_ARAM_DATA), adev->gfx.rlc.register_restore[i]); - /* load direct register */ - WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_SRM_ARAM_ADDR), 0); - for (i = 0; i < adev->gfx.rlc.reg_list_size_bytes >> 2; i++) - WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_SRM_ARAM_DATA), - adev->gfx.rlc.register_restore[i]); - /* load indirect register */ WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_GPM_SCRATCH_ADDR), adev->gfx.rlc.reg_list_format_start); - for (i = 0; i < adev->gfx.rlc.reg_list_format_size_bytes >> 2; i++) + + /* direct register portion */ + for (i = 0; i < adev->gfx.rlc.reg_list_format_direct_reg_list_length; i++) WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_GPM_SCRATCH_DATA), register_list_format[i]); + /* indirect register portion */ + while (i < (adev->gfx.rlc.reg_list_format_size_bytes >> 2)) { + if (register_list_format[i] == 0xFFFFFFFF) { + WREG32_SOC15(GC, 0, mmRLC_GPM_SCRATCH_DATA, register_list_format[i++]); + continue; + } + + WREG32_SOC15(GC, 0, mmRLC_GPM_SCRATCH_DATA, register_list_format[i++]); + WREG32_SOC15(GC, 0, mmRLC_GPM_SCRATCH_DATA, register_list_format[i++]); + + for (j = 0; j < unique_indirect_reg_count; j++) { + if (register_list_format[i] == unique_indirect_regs[j]) { + WREG32_SOC15(GC, 0, mmRLC_GPM_SCRATCH_DATA, j); + break; + } + } + + BUG_ON(j >= unique_indirect_reg_count); + + i++; + } + /* set save/restore list size */ list_size = adev->gfx.rlc.reg_list_size_bytes >> 2; list_size = list_size >> 1; @@ -1823,14 +1906,19 @@ static int gfx_v9_0_init_rlc_save_restore_list(struct amdgpu_device *adev) adev->gfx.rlc.starting_offsets_start); for (i = 0; i < ARRAY_SIZE(indirect_start_offsets); i++) WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_GPM_SCRATCH_DATA), - indirect_start_offsets[i]); + indirect_start_offsets[i]); /* load unique indirect regs*/ for (i = 0; i < ARRAY_SIZE(unique_indirect_regs); i++) { - WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_SRM_INDEX_CNTL_ADDR_0) + i, - unique_indirect_regs[i] & 0x3FFFF); - WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_SRM_INDEX_CNTL_DATA_0) + i, - unique_indirect_regs[i] >> 20); + if (unique_indirect_regs[i] != 0) { + WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_SRM_INDEX_CNTL_ADDR_0) + + GFX_RLC_SRM_INDEX_CNTL_ADDR_OFFSETS[i], + unique_indirect_regs[i] & 0x3FFFF); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_SRM_INDEX_CNTL_DATA_0) + + GFX_RLC_SRM_INDEX_CNTL_DATA_OFFSETS[i], + unique_indirect_regs[i] >> 20); + } } kfree(register_list_format); @@ -2010,6 +2098,9 @@ static void gfx_v9_0_enable_gfx_dynamic_mg_power_gating(struct amdgpu_device *ad static void gfx_v9_0_init_pg(struct amdgpu_device *adev) { + if (!adev->gfx.rlc.is_rlc_v2_1) + return; + if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG | AMD_PG_SUPPORT_GFX_SMG | AMD_PG_SUPPORT_GFX_DMG | @@ -2017,27 +2108,12 @@ static void gfx_v9_0_init_pg(struct amdgpu_device *adev) AMD_PG_SUPPORT_GDS | AMD_PG_SUPPORT_RLC_SMU_HS)) { gfx_v9_0_init_csb(adev); - gfx_v9_0_init_rlc_save_restore_list(adev); + gfx_v9_1_init_rlc_save_restore_list(adev); gfx_v9_0_enable_save_restore_machine(adev); - if (adev->asic_type == CHIP_RAVEN) { - WREG32(mmRLC_JUMP_TABLE_RESTORE, - adev->gfx.rlc.cp_table_gpu_addr >> 8); - gfx_v9_0_init_gfx_power_gating(adev); - - if (adev->pg_flags & AMD_PG_SUPPORT_RLC_SMU_HS) { - gfx_v9_0_enable_sck_slow_down_on_power_up(adev, true); - gfx_v9_0_enable_sck_slow_down_on_power_down(adev, true); - } else { - gfx_v9_0_enable_sck_slow_down_on_power_up(adev, false); - gfx_v9_0_enable_sck_slow_down_on_power_down(adev, false); - } - - if (adev->pg_flags & AMD_PG_SUPPORT_CP) - gfx_v9_0_enable_cp_power_gating(adev, true); - else - gfx_v9_0_enable_cp_power_gating(adev, false); - } + WREG32(mmRLC_JUMP_TABLE_RESTORE, + adev->gfx.rlc.cp_table_gpu_addr >> 8); + gfx_v9_0_init_gfx_power_gating(adev); } } @@ -3061,6 +3137,9 @@ static int gfx_v9_0_hw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; int i; + amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_GFX, + AMD_PG_STATE_UNGATE); + amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); @@ -3279,6 +3358,11 @@ static int gfx_v9_0_late_init(void *handle) if (r) return r; + r = amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_GFX, + AMD_PG_STATE_GATE); + if (r) + return r; + return 0; } @@ -3339,8 +3423,7 @@ static void gfx_v9_0_exit_rlc_safe_mode(struct amdgpu_device *adev) static void gfx_v9_0_update_gfx_cg_power_gating(struct amdgpu_device *adev, bool enable) { - /* TODO: double check if we need to perform under safe mdoe */ - /* gfx_v9_0_enter_rlc_safe_mode(adev); */ + gfx_v9_0_enter_rlc_safe_mode(adev); if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) && enable) { gfx_v9_0_enable_gfx_cg_power_gating(adev, true); @@ -3351,7 +3434,7 @@ static void gfx_v9_0_update_gfx_cg_power_gating(struct amdgpu_device *adev, gfx_v9_0_enable_gfx_pipeline_powergating(adev, false); } - /* gfx_v9_0_exit_rlc_safe_mode(adev); */ + gfx_v9_0_exit_rlc_safe_mode(adev); } static void gfx_v9_0_update_gfx_mg_power_gating(struct amdgpu_device *adev, @@ -3742,7 +3825,7 @@ static void gfx_v9_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, } amdgpu_ring_write(ring, header); -BUG_ON(ib->gpu_addr & 0x3); /* Dword align */ + BUG_ON(ib->gpu_addr & 0x3); /* Dword align */ amdgpu_ring_write(ring, #ifdef __BIG_ENDIAN (2 << 0) | @@ -3774,13 +3857,16 @@ static void gfx_v9_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, { bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT; bool int_sel = flags & AMDGPU_FENCE_FLAG_INT; + bool writeback = flags & AMDGPU_FENCE_FLAG_TC_WB_ONLY; /* RELEASE_MEM - flush caches, send int */ amdgpu_ring_write(ring, PACKET3(PACKET3_RELEASE_MEM, 6)); - amdgpu_ring_write(ring, (EOP_TCL1_ACTION_EN | - EOP_TC_ACTION_EN | - EOP_TC_WB_ACTION_EN | - EOP_TC_MD_ACTION_EN | + amdgpu_ring_write(ring, ((writeback ? (EOP_TC_WB_ACTION_EN | + EOP_TC_NC_ACTION_EN) : + (EOP_TCL1_ACTION_EN | + EOP_TC_ACTION_EN | + EOP_TC_WB_ACTION_EN | + EOP_TC_MD_ACTION_EN)) | EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5))); amdgpu_ring_write(ring, DATA_SEL(write64bit ? 2 : 1) | INT_SEL(int_sel ? 2 : 0)); @@ -4137,6 +4223,20 @@ static void gfx_v9_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg, gfx_v9_0_wait_reg_mem(ring, 0, 0, 0, reg, 0, val, mask, 0x20); } +static void gfx_v9_0_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring, + uint32_t reg0, uint32_t reg1, + uint32_t ref, uint32_t mask) +{ + int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX); + + if (amdgpu_sriov_vf(ring->adev)) + gfx_v9_0_wait_reg_mem(ring, usepfp, 0, 1, reg0, reg1, + ref, mask, 0x20); + else + amdgpu_ring_emit_reg_write_reg_wait_helper(ring, reg0, reg1, + ref, mask); +} + static void gfx_v9_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev, enum amdgpu_interrupt_state state) { @@ -4458,6 +4558,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = { .emit_tmz = gfx_v9_0_ring_emit_tmz, .emit_wreg = gfx_v9_0_ring_emit_wreg, .emit_reg_wait = gfx_v9_0_ring_emit_reg_wait, + .emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait, }; static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = { @@ -4492,6 +4593,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = { .set_priority = gfx_v9_0_ring_set_priority_compute, .emit_wreg = gfx_v9_0_ring_emit_wreg, .emit_reg_wait = gfx_v9_0_ring_emit_reg_wait, + .emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait, }; static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = { @@ -4522,6 +4624,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = { .emit_rreg = gfx_v9_0_ring_emit_rreg, .emit_wreg = gfx_v9_0_ring_emit_wreg, .emit_reg_wait = gfx_v9_0_ring_emit_reg_wait, + .emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait, }; static void gfx_v9_0_set_ring_funcs(struct amdgpu_device *adev) |