drm/amd/powerplay: support hotspot/memory critical limit values

These new interfaces(temp2_crit, temp2_crit_hyst, temp3_crit,
temp3_crit_hyst) are supported on SOC15 dGPUs only.
Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h

@@ -75,6 +75,14 @@ struct amdgpu_dpm_thermal {
 	int                min_temp;
 	/* high temperature threshold */
 	int                max_temp;
+	/* hotspot low temperature threshold */
+	int                min_hotspot_temp;
+	/* hotspot high temperature critical threshold */
+	int                max_hotspot_crit_temp;
+	/* memory low temperature threshold */
+	int                min_mem_temp;
+	/* memory high temperature critical threshold */
+	int                max_mem_crit_temp;
 	/* was last interrupt low to high or high to low */
 	bool               high_to_low;
 	/* interrupt source */

drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c

@@ -1414,6 +1414,38 @@ static ssize_t amdgpu_hwmon_show_temp_thresh(struct device *dev,
 	return snprintf(buf, PAGE_SIZE, "%d\n", temp);
 }
 
+static ssize_t amdgpu_hwmon_show_hotspot_temp_thresh(struct device *dev,
+					     struct device_attribute *attr,
+					     char *buf)
+{
+	struct amdgpu_device *adev = dev_get_drvdata(dev);
+	int hyst = to_sensor_dev_attr(attr)->index;
+	int temp;
+
+	if (hyst)
+		temp = adev->pm.dpm.thermal.min_hotspot_temp;
+	else
+		temp = adev->pm.dpm.thermal.max_hotspot_crit_temp;
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", temp);
+}
+
+static ssize_t amdgpu_hwmon_show_mem_temp_thresh(struct device *dev,
+					     struct device_attribute *attr,
+					     char *buf)
+{
+	struct amdgpu_device *adev = dev_get_drvdata(dev);
+	int hyst = to_sensor_dev_attr(attr)->index;
+	int temp;
+
+	if (hyst)
+		temp = adev->pm.dpm.thermal.min_mem_temp;
+	else
+		temp = adev->pm.dpm.thermal.max_mem_crit_temp;
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", temp);
+}
+
 static ssize_t amdgpu_hwmon_get_pwm1_enable(struct device *dev,
 					    struct device_attribute *attr,
 					    char *buf)
@@ -1985,9 +2017,11 @@ static ssize_t amdgpu_hwmon_show_mclk_label(struct device *dev,
  *
  * - temp1_input: the on die GPU temperature in millidegrees Celsius
  *
- * - temp1_crit: temperature critical max value in millidegrees Celsius
+ * - temp[1-3]_crit: temperature critical max value in millidegrees Celsius
+ *   - temp2_crit and temp3_crit are supported on SOC15 dGPUs only
  *
- * - temp1_crit_hyst: temperature hysteresis for critical limit in millidegrees Celsius
+ * - temp[1-3]_crit_hyst: temperature hysteresis for critical limit in millidegrees Celsius
+ *   - temp2_crit_hyst and temp3_crit_hyst are supported on SOC15 dGPUs only
  *
  * hwmon interfaces for GPU voltage:
  *
@@ -2038,6 +2072,10 @@ static ssize_t amdgpu_hwmon_show_mclk_label(struct device *dev,
 static SENSOR_DEVICE_ATTR(temp1_input, S_IRUGO, amdgpu_hwmon_show_temp, NULL, 0);
 static SENSOR_DEVICE_ATTR(temp1_crit, S_IRUGO, amdgpu_hwmon_show_temp_thresh, NULL, 0);
 static SENSOR_DEVICE_ATTR(temp1_crit_hyst, S_IRUGO, amdgpu_hwmon_show_temp_thresh, NULL, 1);
+static SENSOR_DEVICE_ATTR(temp2_crit, S_IRUGO, amdgpu_hwmon_show_hotspot_temp_thresh, NULL, 0);
+static SENSOR_DEVICE_ATTR(temp2_crit_hyst, S_IRUGO, amdgpu_hwmon_show_hotspot_temp_thresh, NULL, 1);
+static SENSOR_DEVICE_ATTR(temp3_crit, S_IRUGO, amdgpu_hwmon_show_mem_temp_thresh, NULL, 0);
+static SENSOR_DEVICE_ATTR(temp3_crit_hyst, S_IRUGO, amdgpu_hwmon_show_mem_temp_thresh, NULL, 1);
 static SENSOR_DEVICE_ATTR(pwm1, S_IRUGO | S_IWUSR, amdgpu_hwmon_get_pwm1, amdgpu_hwmon_set_pwm1, 0);
 static SENSOR_DEVICE_ATTR(pwm1_enable, S_IRUGO | S_IWUSR, amdgpu_hwmon_get_pwm1_enable, amdgpu_hwmon_set_pwm1_enable, 0);
 static SENSOR_DEVICE_ATTR(pwm1_min, S_IRUGO, amdgpu_hwmon_get_pwm1_min, NULL, 0);
@@ -2064,6 +2102,10 @@ static struct attribute *hwmon_attributes[] = {
 	&sensor_dev_attr_temp1_input.dev_attr.attr,
 	&sensor_dev_attr_temp1_crit.dev_attr.attr,
 	&sensor_dev_attr_temp1_crit_hyst.dev_attr.attr,
+	&sensor_dev_attr_temp2_crit.dev_attr.attr,
+	&sensor_dev_attr_temp2_crit_hyst.dev_attr.attr,
+	&sensor_dev_attr_temp3_crit.dev_attr.attr,
+	&sensor_dev_attr_temp3_crit_hyst.dev_attr.attr,
 	&sensor_dev_attr_pwm1.dev_attr.attr,
 	&sensor_dev_attr_pwm1_enable.dev_attr.attr,
 	&sensor_dev_attr_pwm1_min.dev_attr.attr,
@@ -2186,6 +2228,15 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj,
 	     attr == &sensor_dev_attr_freq2_label.dev_attr.attr))
 		return 0;
 
+	/* only SOC15 dGPUs support hotspot and mem temperatures */
+	if (((adev->flags & AMD_IS_APU) ||
+	     adev->asic_type < CHIP_VEGA10) &&
+	    (attr == &sensor_dev_attr_temp2_crit.dev_attr.attr ||
+	     attr == &sensor_dev_attr_temp2_crit_hyst.dev_attr.attr ||
+	     attr == &sensor_dev_attr_temp3_crit.dev_attr.attr ||
+	     attr == &sensor_dev_attr_temp3_crit_hyst.dev_attr.attr))
+		return 0;
+
 	return effective_mode;
 }
 

drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c

@@ -225,7 +225,13 @@ int phm_register_irq_handlers(struct pp_hwmgr *hwmgr)
 int phm_start_thermal_controller(struct pp_hwmgr *hwmgr)
 {
 	int ret = 0;
-	struct PP_TemperatureRange range = {TEMP_RANGE_MIN, TEMP_RANGE_MAX};
+	struct PP_TemperatureRange range = {
+		TEMP_RANGE_MIN,
+		TEMP_RANGE_MAX,
+		TEMP_RANGE_MIN,
+		TEMP_RANGE_MAX,
+		TEMP_RANGE_MIN,
+		TEMP_RANGE_MAX};
 	struct amdgpu_device *adev = hwmgr->adev;
 
 	if (hwmgr->hwmgr_func->get_thermal_temperature_range)
@@ -239,6 +245,10 @@ int phm_start_thermal_controller(struct pp_hwmgr *hwmgr)
 
 	adev->pm.dpm.thermal.min_temp = range.min;
 	adev->pm.dpm.thermal.max_temp = range.max;
+	adev->pm.dpm.thermal.min_hotspot_temp = range.hotspot_min;
+	adev->pm.dpm.thermal.max_hotspot_crit_temp = range.hotspot_crit_max;
+	adev->pm.dpm.thermal.min_mem_temp = range.mem_min;
+	adev->pm.dpm.thermal.max_mem_crit_temp = range.mem_crit_max;
 
 	return ret;
 }

drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c

@@ -4852,12 +4852,16 @@ static int vega10_notify_cac_buffer_info(struct pp_hwmgr *hwmgr,
 static int vega10_get_thermal_temperature_range(struct pp_hwmgr *hwmgr,
 		struct PP_TemperatureRange *thermal_data)
 {
-	struct phm_ppt_v2_information *table_info =
-			(struct phm_ppt_v2_information *)hwmgr->pptable;
+	struct vega10_hwmgr *data = hwmgr->backend;
+	PPTable_t *pp_table = &(data->smc_state_table.pp_table);
 
 	memcpy(thermal_data, &SMU7ThermalWithDelayPolicy[0], sizeof(struct PP_TemperatureRange));
 
-	thermal_data->max = table_info->tdp_table->usSoftwareShutdownTemp *
+	thermal_data->max = pp_table->TedgeLimit *
+		PP_TEMPERATURE_UNITS_PER_CENTIGRADES;
+	thermal_data->hotspot_crit_max = pp_table->ThotspotLimit *
+		PP_TEMPERATURE_UNITS_PER_CENTIGRADES;
+	thermal_data->mem_crit_max = pp_table->ThbmLimit *
 		PP_TEMPERATURE_UNITS_PER_CENTIGRADES;
 
 	return 0;

drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c

@@ -2526,12 +2526,17 @@ static int vega12_notify_cac_buffer_info(struct pp_hwmgr *hwmgr,
 static int vega12_get_thermal_temperature_range(struct pp_hwmgr *hwmgr,
 		struct PP_TemperatureRange *thermal_data)
 {
-	struct phm_ppt_v3_information *pptable_information =
-		(struct phm_ppt_v3_information *)hwmgr->pptable;
+	struct vega12_hwmgr *data =
+			(struct vega12_hwmgr *)(hwmgr->backend);
+	PPTable_t *pp_table = &(data->smc_state_table.pp_table);
 
 	memcpy(thermal_data, &SMU7ThermalWithDelayPolicy[0], sizeof(struct PP_TemperatureRange));
 
-	thermal_data->max = pptable_information->us_software_shutdown_temp *
+	thermal_data->max = pp_table->TedgeLimit *
+		PP_TEMPERATURE_UNITS_PER_CENTIGRADES;
+	thermal_data->hotspot_crit_max = pp_table->ThotspotLimit *
+		PP_TEMPERATURE_UNITS_PER_CENTIGRADES;
+	thermal_data->mem_crit_max = pp_table->ThbmLimit *
 		PP_TEMPERATURE_UNITS_PER_CENTIGRADES;
 
 	return 0;

drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c

@@ -3974,12 +3974,17 @@ static int vega20_notify_cac_buffer_info(struct pp_hwmgr *hwmgr,
 static int vega20_get_thermal_temperature_range(struct pp_hwmgr *hwmgr,
 		struct PP_TemperatureRange *thermal_data)
 {
-	struct phm_ppt_v3_information *pptable_information =
-		(struct phm_ppt_v3_information *)hwmgr->pptable;
+	struct vega20_hwmgr *data =
+			(struct vega20_hwmgr *)(hwmgr->backend);
+	PPTable_t *pp_table = &(data->smc_state_table.pp_table);
 
 	memcpy(thermal_data, &SMU7ThermalWithDelayPolicy[0], sizeof(struct PP_TemperatureRange));
 
-	thermal_data->max = pptable_information->us_software_shutdown_temp *
+	thermal_data->max = pp_table->TedgeLimit *
+		PP_TEMPERATURE_UNITS_PER_CENTIGRADES;
+	thermal_data->hotspot_crit_max = pp_table->ThotspotLimit *
+		PP_TEMPERATURE_UNITS_PER_CENTIGRADES;
+	thermal_data->mem_crit_max = pp_table->ThbmLimit *
 		PP_TEMPERATURE_UNITS_PER_CENTIGRADES;
 
 	return 0;

drivers/gpu/drm/amd/powerplay/inc/power_state.h

@@ -124,6 +124,10 @@ struct PP_StateSoftwareAlgorithmBlock {
 struct PP_TemperatureRange {
 	int min;
 	int max;
+	int hotspot_min;
+	int hotspot_crit_max;
+	int mem_min;
+	int mem_crit_max;
 };
 
 struct PP_StateValidationBlock {

drivers/gpu/drm/amd/powerplay/inc/pp_thermal.h

@@ -27,14 +27,14 @@
 
 static const struct PP_TemperatureRange SMU7ThermalWithDelayPolicy[] =
 {
-	{-273150,  99000},
-	{ 120000, 120000},
+	{-273150,  99000, -273150, 99000, -273150, 99000},
+	{ 120000, 120000, 120000, 120000, 120000, 120000},
 };
 
 static const struct PP_TemperatureRange SMU7ThermalPolicy[] =
 {
-	{-273150,  99000},
-	{ 120000, 120000},
+	{-273150,  99000, -273150, 99000, -273150, 99000},
+	{ 120000, 120000, 120000, 120000, 120000, 120000},
 };
 
 #endif