[PATCH] drm/amd/pm: consistent approach for smartshift
Lazar, Lijo
lijo.lazar at amd.com
Wed May 11 09:10:35 UTC 2022
On 5/11/2022 2:18 PM, Sathishkumar S wrote:
> always create smartshift attributes from dgpu device even on SS1.0.
> consider units of power in metrics table and convert if necessary.
> powershift value is in percentage and values to range between 0-100.
>
> Suggested-by: Alex Deucher <alexander.deucher at amd.com>
> Signed-off-by: Sathishkumar S <sathishkumar.sundararaju at amd.com>
> Acked-by: Alex Deucher <alexander.deucher at amd.com>
Can you split the RN, YC and generic changes to 3 separate patches?
Would be easier to review and revert specifically for an ASIC.
Thanks,
Lijo
> ---
> drivers/gpu/drm/amd/pm/amdgpu_pm.c | 146 +++++++++++-------
> .../gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c | 54 +++++--
> .../drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c | 54 +++++--
> 3 files changed, 166 insertions(+), 88 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/pm/amdgpu_pm.c b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
> index d3228216b2da..c2406baeef93 100644
> --- a/drivers/gpu/drm/amd/pm/amdgpu_pm.c
> +++ b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
> @@ -1734,22 +1734,11 @@ static ssize_t amdgpu_get_gpu_metrics(struct device *dev,
> return size;
> }
>
> -/**
> - * DOC: smartshift_apu_power
> - *
> - * The amdgpu driver provides a sysfs API for reporting APU power
> - * share if it supports smartshift. The value is expressed as
> - * the proportion of stapm limit where stapm limit is the total APU
> - * power limit. The result is in percentage. If APU power is 130% of
> - * STAPM, then APU is using 30% of the dGPU's headroom.
> - */
> -
> -static ssize_t amdgpu_get_smartshift_apu_power(struct device *dev, struct device_attribute *attr,
> - char *buf)
> +static int amdgpu_read_powershift_percent(struct amdgpu_device *adev,
> + uint32_t *ss_power, bool dgpu_share)
> {
> - struct drm_device *ddev = dev_get_drvdata(dev);
> - struct amdgpu_device *adev = drm_to_adev(ddev);
> - uint32_t ss_power, size;
> + struct drm_device *ddev = adev_to_drm(adev);
> + uint32_t size;
> int r = 0;
>
> if (amdgpu_in_reset(adev))
> @@ -1763,28 +1752,64 @@ static ssize_t amdgpu_get_smartshift_apu_power(struct device *dev, struct device
> return r;
> }
>
> - r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_SS_APU_SHARE,
> - (void *)&ss_power, &size);
> - if (r)
> - goto out;
> -
> - r = sysfs_emit(buf, "%u%%\n", ss_power);
> + if (dgpu_share)
> + r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_SS_DGPU_SHARE,
> + (void *)ss_power, &size);
> + else
> + r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_SS_APU_SHARE,
> + (void *)ss_power, &size);
>
> -out:
> pm_runtime_mark_last_busy(ddev->dev);
> pm_runtime_put_autosuspend(ddev->dev);
> return r;
> }
> +/**
> + * DOC: smartshift_apu_power
> + *
> + * The amdgpu driver provides a sysfs API for reporting APU power
> + * shift in percentage if platform supports smartshift. Value 0 means that
> + * there is no powershift and values between [1-100] means that the power
> + * is shifted to APU, the percentage of boost is with respect to APU power
> + * limit on the platform.
> + */
> +
> +static ssize_t amdgpu_get_smartshift_apu_power(struct device *dev, struct device_attribute *attr,
> + char *buf)
> +{
> + struct drm_device *ddev = dev_get_drvdata(dev);
> + struct amdgpu_device *adev = drm_to_adev(ddev);
> + uint32_t ss_power = 0;
> + int r = 0, i;
> +
> + r = amdgpu_read_powershift_percent(adev, &ss_power, false);
> + if (!r)
> + r = sysfs_emit(buf, "%u%%\n", ss_power);
> + else if (r == -EOPNOTSUPP) {
> + /* sensor not available on dGPU, try to read from APU */
> + adev = NULL;
> + mutex_lock(&mgpu_info.mutex);
> + for (i = 0; i < mgpu_info.num_gpu; i++) {
> + if (mgpu_info.gpu_ins[i].adev->flags & AMD_IS_APU) {
> + adev = mgpu_info.gpu_ins[i].adev;
> + break;
> + }
> + }
> + mutex_unlock(&mgpu_info.mutex);
> + if (adev && !amdgpu_read_powershift_percent(adev, &ss_power, false))
> + r = sysfs_emit(buf, "%u%%\n", ss_power);
> + }
> +
> + return r;
> +}
>
> /**
> * DOC: smartshift_dgpu_power
> *
> - * The amdgpu driver provides a sysfs API for reporting the dGPU power
> - * share if the device is in HG and supports smartshift. The value
> - * is expressed as the proportion of stapm limit where stapm limit
> - * is the total APU power limit. The value is in percentage. If dGPU
> - * power is 20% higher than STAPM power(120%), it's using 20% of the
> - * APU's power headroom.
> + * The amdgpu driver provides a sysfs API for reporting dGPU power
> + * shift in percentage if platform supports smartshift. Value 0 means that
> + * there is no powershift and values between [1-100] means that the power is
> + * shifted to dGPU, the percentage of boost is with respect to dGPU power
> + * limit on the platform.
> */
>
> static ssize_t amdgpu_get_smartshift_dgpu_power(struct device *dev, struct device_attribute *attr,
> @@ -1792,31 +1817,27 @@ static ssize_t amdgpu_get_smartshift_dgpu_power(struct device *dev, struct devic
> {
> struct drm_device *ddev = dev_get_drvdata(dev);
> struct amdgpu_device *adev = drm_to_adev(ddev);
> - uint32_t ss_power, size;
> - int r = 0;
> + uint32_t ss_power = 0;
> + int r = 0, i;
>
> - if (amdgpu_in_reset(adev))
> - return -EPERM;
> - if (adev->in_suspend && !adev->in_runpm)
> - return -EPERM;
> -
> - r = pm_runtime_get_sync(ddev->dev);
> - if (r < 0) {
> - pm_runtime_put_autosuspend(ddev->dev);
> - return r;
> + r = amdgpu_read_powershift_percent(adev, &ss_power, true);
> + if (!r)
> + r = sysfs_emit(buf, "%u%%\n", ss_power);
> + else if (r == -EOPNOTSUPP) {
> + /* sensor not available on dGPU, try to read from APU */
> + adev = NULL;
> + mutex_lock(&mgpu_info.mutex);
> + for (i = 0; i < mgpu_info.num_gpu; i++) {
> + if (mgpu_info.gpu_ins[i].adev->flags & AMD_IS_APU) {
> + adev = mgpu_info.gpu_ins[i].adev;
> + break;
> + }
> + }
> + mutex_unlock(&mgpu_info.mutex);
> + if (adev && !amdgpu_read_powershift_percent(adev, &ss_power, true))
> + r = sysfs_emit(buf, "%u%%\n", ss_power);
> }
>
> - r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_SS_DGPU_SHARE,
> - (void *)&ss_power, &size);
> -
> - if (r)
> - goto out;
> -
> - r = sysfs_emit(buf, "%u%%\n", ss_power);
> -
> -out:
> - pm_runtime_mark_last_busy(ddev->dev);
> - pm_runtime_put_autosuspend(ddev->dev);
> return r;
> }
>
> @@ -1884,18 +1905,23 @@ static ssize_t amdgpu_set_smartshift_bias(struct device *dev,
> static int ss_power_attr_update(struct amdgpu_device *adev, struct amdgpu_device_attr *attr,
> uint32_t mask, enum amdgpu_device_attr_states *states)
> {
> - uint32_t ss_power, size;
> + struct pci_dev *pdev = NULL;
> + unsigned short devices = 0;
>
> - if (!amdgpu_acpi_is_power_shift_control_supported())
> - *states = ATTR_STATE_UNSUPPORTED;
> - else if ((adev->flags & AMD_IS_PX) &&
> - !amdgpu_device_supports_smart_shift(adev_to_drm(adev)))
> - *states = ATTR_STATE_UNSUPPORTED;
> - else if (amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_SS_APU_SHARE,
> - (void *)&ss_power, &size))
> + if (!amdgpu_device_supports_smart_shift(adev_to_drm(adev)))
> *states = ATTR_STATE_UNSUPPORTED;
> - else if (amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_SS_DGPU_SHARE,
> - (void *)&ss_power, &size))
> +
> + while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, pdev)) != NULL) {
> + if (pdev->vendor == 0x1002)
> + devices++;
> + }
> + pdev = NULL;
> + while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_OTHER << 8, pdev)) != NULL) {
> + if (pdev->vendor == 0x1002)
> + devices++;
> + }
> +
> + if (devices < 2)
> *states = ATTR_STATE_UNSUPPORTED;
>
> return 0;
> diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c
> index fd6c44ece168..a781c63b10aa 100644
> --- a/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c
> +++ b/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c
> @@ -1119,6 +1119,39 @@ static int renoir_get_power_profile_mode(struct smu_context *smu,
> return size;
> }
>
> +static void renoir_get_ss_power_percent(SmuMetrics_t *metrics,
> + uint32_t *apu_percent, uint32_t *dgpu_percent)
> +{
> + uint32_t apu_boost = 0;
> + uint32_t dgpu_boost = 0;
> + uint16_t apu_limit = 0;
> + uint16_t dgpu_limit = 0;
> + uint16_t apu_power = 0;
> + uint16_t dgpu_power = 0;
> +
> + apu_power = metrics->ApuPower;
> + apu_limit = metrics->StapmOriginalLimit;
> + if (apu_power > apu_limit && apu_limit != 0)
> + apu_boost = ((apu_power - apu_limit) * 100) / apu_limit;
> + apu_boost = (apu_boost > 100) ? 100 : apu_boost;
> +
> + dgpu_power = metrics->dGpuPower;
> + if (metrics->StapmCurrentLimit > metrics->StapmOriginalLimit)
> + dgpu_limit = metrics->StapmCurrentLimit - metrics->StapmOriginalLimit;
> + if (dgpu_power > dgpu_limit && dgpu_limit != 0)
> + dgpu_boost = ((dgpu_power - dgpu_limit) * 100) / dgpu_limit;
> + dgpu_boost = (dgpu_boost > 100) ? 100 : dgpu_boost;
> +
> + if (dgpu_boost >= apu_boost)
> + apu_boost = 0;
> + else
> + dgpu_boost = 0;
> +
> + *apu_percent = apu_boost;
> + *dgpu_percent = dgpu_boost;
> +}
> +
> +
> static int renoir_get_smu_metrics_data(struct smu_context *smu,
> MetricsMember_t member,
> uint32_t *value)
> @@ -1127,6 +1160,9 @@ static int renoir_get_smu_metrics_data(struct smu_context *smu,
>
> SmuMetrics_t *metrics = (SmuMetrics_t *)smu_table->metrics_table;
> int ret = 0;
> + uint32_t apu_percent = 0;
> + uint32_t dgpu_percent = 0;
> +
>
> ret = smu_cmn_get_metrics_table(smu,
> NULL,
> @@ -1172,25 +1208,15 @@ static int renoir_get_smu_metrics_data(struct smu_context *smu,
> break;
> case METRICS_SS_APU_SHARE:
> /* return the percentage of APU power with respect to APU's power limit.
> - * percentage is reported, this isn't boost value. Smartshift power
> - * boost/shift is only when the percentage is more than 100.
> */
> - if (metrics->StapmOriginalLimit > 0)
> - *value = (metrics->ApuPower * 100) / metrics->StapmOriginalLimit;
> - else
> - *value = 0;
> + renoir_get_ss_power_percent(metrics, &apu_percent, &dgpu_percent);
> + *value = apu_percent;
> break;
> case METRICS_SS_DGPU_SHARE:
> /* return the percentage of dGPU power with respect to dGPU's power limit.
> - * percentage is reported, this isn't boost value. Smartshift power
> - * boost/shift is only when the percentage is more than 100.
> */
> - if ((metrics->dGpuPower > 0) &&
> - (metrics->StapmCurrentLimit > metrics->StapmOriginalLimit))
> - *value = (metrics->dGpuPower * 100) /
> - (metrics->StapmCurrentLimit - metrics->StapmOriginalLimit);
> - else
> - *value = 0;
> + renoir_get_ss_power_percent(metrics, &apu_percent, &dgpu_percent);
> + *value = dgpu_percent;
> break;
> default:
> *value = UINT_MAX;
> diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c
> index e2d099409123..ac3d4b0863d0 100644
> --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c
> +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c
> @@ -276,6 +276,40 @@ static int yellow_carp_mode2_reset(struct smu_context *smu)
> return yellow_carp_mode_reset(smu, SMU_RESET_MODE_2);
> }
>
> +
> +static void yellow_carp_get_ss_power_percent(SmuMetrics_t *metrics,
> + uint32_t *apu_percent, uint32_t *dgpu_percent)
> +{
> + uint32_t apu_boost = 0;
> + uint32_t dgpu_boost = 0;
> + uint16_t apu_limit = 0;
> + uint16_t dgpu_limit = 0;
> + uint16_t apu_power = 0;
> + uint16_t dgpu_power = 0;
> +
> + apu_power = metrics->ApuPower/1000;
> + apu_limit = metrics->StapmOpnLimit;
> + if (apu_power > apu_limit && apu_limit != 0)
> + apu_boost = ((apu_power - apu_limit) * 100) / apu_limit;
> + apu_boost = (apu_boost > 100) ? 100 : apu_boost;
> +
> + dgpu_power = metrics->dGpuPower/1000;
> + if (metrics->StapmCurrentLimit > metrics->StapmOpnLimit)
> + dgpu_limit = metrics->StapmCurrentLimit - metrics->StapmOpnLimit;
> + if (dgpu_power > dgpu_limit && dgpu_limit != 0)
> + dgpu_boost = ((dgpu_power - dgpu_limit) * 100) / dgpu_limit;
> + dgpu_boost = (dgpu_boost > 100) ? 100 : dgpu_boost;
> +
> + if (dgpu_boost >= apu_boost)
> + apu_boost = 0;
> + else
> + dgpu_boost = 0;
> +
> + *apu_percent = apu_boost;
> + *dgpu_percent = dgpu_boost;
> +
> +}
> +
> static int yellow_carp_get_smu_metrics_data(struct smu_context *smu,
> MetricsMember_t member,
> uint32_t *value)
> @@ -284,6 +318,8 @@ static int yellow_carp_get_smu_metrics_data(struct smu_context *smu,
>
> SmuMetrics_t *metrics = (SmuMetrics_t *)smu_table->metrics_table;
> int ret = 0;
> + uint32_t apu_percent = 0;
> + uint32_t dgpu_percent = 0;
>
> ret = smu_cmn_get_metrics_table(smu, NULL, false);
> if (ret)
> @@ -333,25 +369,15 @@ static int yellow_carp_get_smu_metrics_data(struct smu_context *smu,
> break;
> case METRICS_SS_APU_SHARE:
> /* return the percentage of APU power with respect to APU's power limit.
> - * percentage is reported, this isn't boost value. Smartshift power
> - * boost/shift is only when the percentage is more than 100.
> */
> - if (metrics->StapmOpnLimit > 0)
> - *value = (metrics->ApuPower * 100) / metrics->StapmOpnLimit;
> - else
> - *value = 0;
> + yellow_carp_get_ss_power_percent(metrics, &apu_percent, &dgpu_percent);
> + *value = apu_percent;
> break;
> case METRICS_SS_DGPU_SHARE:
> /* return the percentage of dGPU power with respect to dGPU's power limit.
> - * percentage is reported, this isn't boost value. Smartshift power
> - * boost/shift is only when the percentage is more than 100.
> */
> - if ((metrics->dGpuPower > 0) &&
> - (metrics->StapmCurrentLimit > metrics->StapmOpnLimit))
> - *value = (metrics->dGpuPower * 100) /
> - (metrics->StapmCurrentLimit - metrics->StapmOpnLimit);
> - else
> - *value = 0;
> + yellow_carp_get_ss_power_percent(metrics, &apu_percent, &dgpu_percent);
> + *value = dgpu_percent;
> break;
> default:
> *value = UINT_MAX;
>
More information about the amd-gfx
mailing list