lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Sun, 14 May 2017 21:23:08 +0000
From:   Chris Packham <Chris.Packham@...iedtelesis.co.nz>
To:     Guenter Roeck <linux@...ck-us.net>,
        "linux-hwmon@...r.kernel.org" <linux-hwmon@...r.kernel.org>
CC:     Tobi Wulff <Tobi.Wulff@...iedtelesis.co.nz>,
        Jean Delvare <jdelvare@...e.com>,
        Jonathan Corbet <corbet@....net>,
        "linux-doc@...r.kernel.org" <linux-doc@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v3 3/4] hwmon: (adt7475) temperature smoothing

On 15/05/17 03:40, Guenter Roeck wrote:
> On 05/10/2017 08:45 PM, Chris Packham wrote:
>> When enabled temperature smoothing allows ramping the fan speed over a
>> configurable period of time instead of jumping to the new speed
>> instantaneously.
>>
>> Signed-off-by: Chris Packham <chris.packham@...iedtelesis.co.nz>
>> ---
>>
>> Changes in v2:
>> - use a single tempN_smoothing attribute
>
> This is a bit confusing. tempN suggests that the attribute would be associated
> with a given temperature, not with fan control. Not that I have a better idea
> for an attribute name, though, so unless you find a better name I am ok with it.
>

The datasheet is a bit confusing in this respect.

 From the description of register 0x62:

"Assuming that PWMx is associated with the Remote 1 temperature channel, 
these bits define the maximum rate of change of the PWMx output for 
Remote 1 temperature related changes. Instead of the fan speed jumping 
instantaneously to its newly determined speed, it ramps
gracefully at the rate determined by these bits. This feature ultimately 
enhances the acoustics of the fan."

Based on my reading it's a property of the temperature input not of the 
PWM. If you changed pwmN_auto_channels_temp this setting would stay with 
the temperature sensor not the PWM.

>> Changes in v3:
>> - change enh_acou to enh_acoustics
>> - simplify show_temp_st()
>>
>>  Documentation/hwmon/adt7475 |  4 ++
>>  drivers/hwmon/adt7475.c     | 93 +++++++++++++++++++++++++++++++++++++++++++++
>>  2 files changed, 97 insertions(+)
>>
>> diff --git a/Documentation/hwmon/adt7475 b/Documentation/hwmon/adt7475
>> index 3990bae60e78..e82b24ec4b07 100644
>> --- a/Documentation/hwmon/adt7475
>> +++ b/Documentation/hwmon/adt7475
>> @@ -114,6 +114,10 @@ minimum (i.e. auto_point1_pwm). This behaviour can be configured using the
>>  pwm[1-*]_stall_dis sysfs attribute. A value of 0 means the fans will shut off.
>>  A value of 1 means the fans will run at auto_point1_pwm.
>>
>> +The responsiveness of the ADT747x to temperature changes can be configured.
>> +This allows smoothing of the fan speed transition. To set the transition time
>> +set the value in ms in the temp[1-*]_smoothing sysfs attribute.
>> +
>>  Notes
>>  -----
>>
>> diff --git a/drivers/hwmon/adt7475.c b/drivers/hwmon/adt7475.c
>> index 4d6c625fec70..f7322330789c 100644
>> --- a/drivers/hwmon/adt7475.c
>> +++ b/drivers/hwmon/adt7475.c
>> @@ -526,6 +526,90 @@ static ssize_t set_temp(struct device *dev, struct device_attribute *attr,
>>  	return count;
>>  }
>>
>> +/* Assuming CONFIG6[SLOW] is 0 */
>> +static const int ad7475_st_map[] = {
>> +	37500, 18800, 12500, 7500, 4700, 3100, 1600, 800,
>> +};
>> +
>> +static ssize_t show_temp_st(struct device *dev, struct device_attribute *attr,
>> +				  char *buf)
>> +{
>> +	struct sensor_device_attribute_2 *sattr = to_sensor_dev_attr_2(attr);
>> +	struct i2c_client *client = to_i2c_client(dev);
>> +	struct adt7475_data *data = i2c_get_clientdata(client);
>> +	long val;
>> +
>> +	switch (sattr->index) {
>> +	case 0:
>> +		val = data->enh_acoustics[0] & 0xf;
>> +		break;
>> +	case 1:
>> +		val = (data->enh_acoustics[1] >> 4) & 0xf;
>> +		break;
>> +	case 2:
>> +		val = data->enh_acoustics[1] & 0xf;
>> +		break;
>> +	default:
>> +		return -EINVAL;
>
> This will never happen and, if it does, would indicate a bug, not invalid input.
> I kind of dislike dead code; it just bloats the kernel. Please either use
> default: for or together with case 2:, or make it if/else.
>

Will combine default and case 2.

>> +	}
>> +
>> +	if (val & 0x8)
>> +		return sprintf(buf, "%d\n", ad7475_st_map[val & 0x7]);
>> +	else
>> +		return sprintf(buf, "0\n");
>> +}
>> +
>> +static ssize_t set_temp_st(struct device *dev, struct device_attribute *attr,
>> +				 const char *buf, size_t count)
>> +{
>> +	struct sensor_device_attribute_2 *sattr = to_sensor_dev_attr_2(attr);
>> +	struct i2c_client *client = to_i2c_client(dev);
>> +	struct adt7475_data *data = i2c_get_clientdata(client);
>> +	unsigned char reg;
>> +	int shift, idx;
>> +	ulong val;
>> +
>> +	if (kstrtoul(buf, 10, &val))
>> +		return -EINVAL;
>> +
>> +	switch (sattr->index) {
>> +	case 0:
>> +		reg = REG_ENHANCE_ACOUSTICS1;
>> +		shift = 0;
>> +		idx = 0;
>> +		break;
>> +	case 1:
>> +		reg = REG_ENHANCE_ACOUSTICS2;
>> +		shift = 4;
>> +		idx = 1;
>> +		break;
>> +	case 2:
>> +		reg = REG_ENHANCE_ACOUSTICS2;
>> +		shift = 0;
>> +		idx = 1;
>> +		break;
>
> Is this correct ? It associates
> 	temp1_smoothing -> remote 1
> 	temp2_smoothing -> remote 2
> 	temp3_smoothing -> local

Yes

> which, unless I am missing something, doesn't match temp1/2/3.
>

So it should be

temp1 -> Remote 1
temp2 -> Local
temp3 -> Remote 2

I'll rework accordingly.

>> +	default:
>> +		return -EINVAL;
>
> Same as above.
>

Done.

>> +	}
>> +
>> +	if (val > 0) {
>> +		val = find_closest_descending(val, ad7475_st_map,
>> +					      ARRAY_SIZE(ad7475_st_map));
>> +		val |= 0x8;
>> +	}
>> +
>> +	mutex_lock(&data->lock);
>> +
>> +	data->enh_acoustics[idx] &= ~(0xf << shift);
>> +	data->enh_acoustics[idx] |= (val << shift);
>> +
>> +	i2c_smbus_write_byte_data(client, reg, data->enh_acoustics[idx]);
>> +
>> +	mutex_unlock(&data->lock);
>> +
>> +	return count;
>> +}
>> +
>>  /*
>>   * Table of autorange values - the user will write the value in millidegrees,
>>   * and we'll convert it
>> @@ -1008,6 +1092,8 @@ static SENSOR_DEVICE_ATTR_2(temp1_crit, S_IRUGO | S_IWUSR, show_temp, set_temp,
>>  			    THERM, 0);
>>  static SENSOR_DEVICE_ATTR_2(temp1_crit_hyst, S_IRUGO | S_IWUSR, show_temp,
>>  			    set_temp, HYSTERSIS, 0);
>> +static SENSOR_DEVICE_ATTR_2(temp1_smoothing, S_IRUGO | S_IWUSR, show_temp_st,
>> +			    set_temp_st, 0, 0);
>>  static SENSOR_DEVICE_ATTR_2(temp2_input, S_IRUGO, show_temp, NULL, INPUT, 1);
>>  static SENSOR_DEVICE_ATTR_2(temp2_alarm, S_IRUGO, show_temp, NULL, ALARM, 1);
>>  static SENSOR_DEVICE_ATTR_2(temp2_max, S_IRUGO | S_IWUSR, show_temp, set_temp,
>> @@ -1024,6 +1110,8 @@ static SENSOR_DEVICE_ATTR_2(temp2_crit, S_IRUGO | S_IWUSR, show_temp, set_temp,
>>  			    THERM, 1);
>>  static SENSOR_DEVICE_ATTR_2(temp2_crit_hyst, S_IRUGO | S_IWUSR, show_temp,
>>  			    set_temp, HYSTERSIS, 1);
>> +static SENSOR_DEVICE_ATTR_2(temp2_smoothing, S_IRUGO | S_IWUSR, show_temp_st,
>> +			    set_temp_st, 0, 1);
>>  static SENSOR_DEVICE_ATTR_2(temp3_input, S_IRUGO, show_temp, NULL, INPUT, 2);
>>  static SENSOR_DEVICE_ATTR_2(temp3_alarm, S_IRUGO, show_temp, NULL, ALARM, 2);
>>  static SENSOR_DEVICE_ATTR_2(temp3_fault, S_IRUGO, show_temp, NULL, FAULT, 2);
>> @@ -1041,6 +1129,8 @@ static SENSOR_DEVICE_ATTR_2(temp3_crit, S_IRUGO | S_IWUSR, show_temp, set_temp,
>>  			    THERM, 2);
>>  static SENSOR_DEVICE_ATTR_2(temp3_crit_hyst, S_IRUGO | S_IWUSR, show_temp,
>>  			    set_temp, HYSTERSIS, 2);
>> +static SENSOR_DEVICE_ATTR_2(temp3_smoothing, S_IRUGO | S_IWUSR, show_temp_st,
>> +			    set_temp_st, 0, 2);
>>  static SENSOR_DEVICE_ATTR_2(fan1_input, S_IRUGO, show_tach, NULL, INPUT, 0);
>>  static SENSOR_DEVICE_ATTR_2(fan1_min, S_IRUGO | S_IWUSR, show_tach, set_tach,
>>  			    MIN, 0);
>> @@ -1125,6 +1215,7 @@ static struct attribute *adt7475_attrs[] = {
>>  	&sensor_dev_attr_temp1_auto_point2_temp.dev_attr.attr,
>>  	&sensor_dev_attr_temp1_crit.dev_attr.attr,
>>  	&sensor_dev_attr_temp1_crit_hyst.dev_attr.attr,
>> +	&sensor_dev_attr_temp1_smoothing.dev_attr.attr,
>>  	&sensor_dev_attr_temp2_input.dev_attr.attr,
>>  	&sensor_dev_attr_temp2_alarm.dev_attr.attr,
>>  	&sensor_dev_attr_temp2_max.dev_attr.attr,
>> @@ -1134,6 +1225,7 @@ static struct attribute *adt7475_attrs[] = {
>>  	&sensor_dev_attr_temp2_auto_point2_temp.dev_attr.attr,
>>  	&sensor_dev_attr_temp2_crit.dev_attr.attr,
>>  	&sensor_dev_attr_temp2_crit_hyst.dev_attr.attr,
>> +	&sensor_dev_attr_temp2_smoothing.dev_attr.attr,
>>  	&sensor_dev_attr_temp3_input.dev_attr.attr,
>>  	&sensor_dev_attr_temp3_fault.dev_attr.attr,
>>  	&sensor_dev_attr_temp3_alarm.dev_attr.attr,
>> @@ -1144,6 +1236,7 @@ static struct attribute *adt7475_attrs[] = {
>>  	&sensor_dev_attr_temp3_auto_point2_temp.dev_attr.attr,
>>  	&sensor_dev_attr_temp3_crit.dev_attr.attr,
>>  	&sensor_dev_attr_temp3_crit_hyst.dev_attr.attr,
>> +	&sensor_dev_attr_temp3_smoothing.dev_attr.attr,
>>  	&sensor_dev_attr_fan1_input.dev_attr.attr,
>>  	&sensor_dev_attr_fan1_min.dev_attr.attr,
>>  	&sensor_dev_attr_fan1_alarm.dev_attr.attr,
>>
>
>

Powered by blists - more mailing lists