From: Likun Gao <Likun.Gao@xxxxxxx> For Sienna_Cichlid, GFXOFF state puts gfx dpm into standby mode, then the gfxclk can't be retireved. Signed-off-by: Likun Gao <Likun.Gao@xxxxxxx> Reviewed-by: Kenneth Feng <kenneth.feng@xxxxxxx> Signed-off-by: Alex Deucher <alexander.deucher@xxxxxxx> --- .../drm/amd/powerplay/sienna_cichlid_ppt.c | 71 ++++++++++++++++--- 1 file changed, 60 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c index ef8532ff8e30..2ac30d2ddf5e 100644 --- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c +++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c @@ -706,6 +706,9 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu, uint32_t mark_index = 0; uint32_t gen_speed, lane_width; + if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK)) + amdgpu_gfx_off_ctrl(adev, false); + switch (clk_type) { case SMU_GFXCLK: case SMU_SCLK: @@ -716,20 +719,20 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu, case SMU_DCEFCLK: ret = smu_get_current_clk_freq(smu, clk_type, &cur_value); if (ret) - return size; + goto print_clk_out; /* 10KHz -> MHz */ cur_value = cur_value / 100; ret = smu_get_dpm_level_count(smu, clk_type, &count); if (ret) - return size; + goto print_clk_out; if (!sienna_cichlid_is_support_fine_grained_dpm(smu, clk_type)) { for (i = 0; i < count; i++) { ret = smu_get_dpm_freq_by_index(smu, clk_type, i, &value); if (ret) - return size; + goto print_clk_out; size += sprintf(buf + size, "%d: %uMhz %s\n", i, value, cur_value == value ? "*" : ""); @@ -737,10 +740,10 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu, } else { ret = smu_get_dpm_freq_by_index(smu, clk_type, 0, &freq_values[0]); if (ret) - return size; + goto print_clk_out; ret = smu_get_dpm_freq_by_index(smu, clk_type, count - 1, &freq_values[2]); if (ret) - return size; + goto print_clk_out; freq_values[1] = cur_value; mark_index = cur_value == freq_values[0] ? 0 : @@ -783,19 +786,26 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu, break; } +print_clk_out: + if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK)) + amdgpu_gfx_off_ctrl(adev, true); + return size; } static int sienna_cichlid_force_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, uint32_t mask) { - + struct amdgpu_device *adev = smu->adev; int ret = 0, size = 0; uint32_t soft_min_level = 0, soft_max_level = 0, min_freq = 0, max_freq = 0; soft_min_level = mask ? (ffs(mask) - 1) : 0; soft_max_level = mask ? (fls(mask) - 1) : 0; + if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK)) + amdgpu_gfx_off_ctrl(adev, false); + switch (clk_type) { case SMU_GFXCLK: case SMU_SCLK: @@ -812,20 +822,24 @@ static int sienna_cichlid_force_clk_levels(struct smu_context *smu, ret = smu_get_dpm_freq_by_index(smu, clk_type, soft_min_level, &min_freq); if (ret) - return size; + goto forec_level_out; ret = smu_get_dpm_freq_by_index(smu, clk_type, soft_max_level, &max_freq); if (ret) - return size; + goto forec_level_out; ret = smu_set_soft_freq_range(smu, clk_type, min_freq, max_freq, false); if (ret) - return size; + goto forec_level_out; break; default: break; } +forec_level_out: + if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK)) + amdgpu_gfx_off_ctrl(adev, true); + return size; } @@ -1265,6 +1279,7 @@ static int sienna_cichlid_get_profiling_clk_mask(struct smu_context *smu, uint32_t *mclk_mask, uint32_t *soc_mask) { + struct amdgpu_device *adev = smu->adev; int ret = 0; uint32_t level_count = 0; @@ -1276,7 +1291,9 @@ static int sienna_cichlid_get_profiling_clk_mask(struct smu_context *smu, *mclk_mask = 0; } else if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) { if(sclk_mask) { + amdgpu_gfx_off_ctrl(adev, false); ret = smu_get_dpm_level_count(smu, SMU_SCLK, &level_count); + amdgpu_gfx_off_ctrl(adev, true); if (ret) return ret; *sclk_mask = level_count - 1; @@ -1698,6 +1715,38 @@ static int sienna_cichlid_update_pcie_parameters(struct smu_context *smu, return 0; } +int sienna_cichlid_get_dpm_ultimate_freq(struct smu_context *smu, + enum smu_clk_type clk_type, + uint32_t *min, uint32_t *max) +{ + struct amdgpu_device *adev = smu->adev; + int ret; + + if (clk_type == SMU_GFXCLK) + amdgpu_gfx_off_ctrl(adev, false); + ret = smu_v11_0_get_dpm_ultimate_freq(smu, clk_type, min, max); + if (clk_type == SMU_GFXCLK) + amdgpu_gfx_off_ctrl(adev, true); + + return ret; +} + +int sienna_cichlid_set_soft_freq_limited_range(struct smu_context *smu, + enum smu_clk_type clk_type, + uint32_t min, uint32_t max) +{ + struct amdgpu_device *adev = smu->adev; + int ret; + + if (clk_type == SMU_GFXCLK) + amdgpu_gfx_off_ctrl(adev, false); + ret = smu_v11_0_set_soft_freq_limited_range(smu, clk_type, min, max); + if (clk_type == SMU_GFXCLK) + amdgpu_gfx_off_ctrl(adev, true); + + return ret; +} + static void sienna_cichlid_dump_pptable(struct smu_context *smu) { struct smu_table_context *table_context = &smu->smu_table; @@ -2452,8 +2501,8 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = { .baco_set_state = smu_v11_0_baco_set_state, .baco_enter = smu_v11_0_baco_enter, .baco_exit = smu_v11_0_baco_exit, - .get_dpm_ultimate_freq = smu_v11_0_get_dpm_ultimate_freq, - .set_soft_freq_limited_range = smu_v11_0_set_soft_freq_limited_range, + .get_dpm_ultimate_freq = sienna_cichlid_get_dpm_ultimate_freq, + .set_soft_freq_limited_range = sienna_cichlid_set_soft_freq_limited_range, .override_pcie_parameters = smu_v11_0_override_pcie_parameters, .get_pptable_power_limit = sienna_cichlid_get_pptable_power_limit, }; -- 2.25.4 _______________________________________________ amd-gfx mailing list amd-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/amd-gfx