Commit 38a1f0e9 authored by Jean Delvare's avatar Jean Delvare Committed by Mark M. Hoffman

hwmon: (adm1031) Fix register overwrite in set_fan_div()

Don't rely on the register cache when setting a new fan clock divider.
For one thing, the cache might not have been initialized at all if the
driver has just been loaded. For another, the cached values may be old
and you never know what can happen in the driver's back.

Also invalidate the cache instead of trying to adjust the measured fan
speed: the whole point of changing the clock divider is to get a better
reading.
Signed-off-by: default avatarJean Delvare <khali@linux-fr.org>
Signed-off-by: default avatarMark M. Hoffman <mhoffman@lightlink.com>
parent d5b0b5d6
......@@ -542,18 +542,26 @@ set_fan_div(struct device *dev, const char *buf, size_t count, int nr)
return -EINVAL;
mutex_lock(&data->update_lock);
/* Get fresh readings */
data->fan_div[nr] = adm1031_read_value(client,
ADM1031_REG_FAN_DIV(nr));
data->fan_min[nr] = adm1031_read_value(client,
ADM1031_REG_FAN_MIN(nr));
/* Write the new clock divider and fan min */
old_div = FAN_DIV_FROM_REG(data->fan_div[nr]);
data->fan_div[nr] = (tmp & 0xC0) | (0x3f & data->fan_div[nr]);
new_min = data->fan_min[nr] * old_div /
FAN_DIV_FROM_REG(data->fan_div[nr]);
data->fan_min[nr] = new_min > 0xff ? 0xff : new_min;
data->fan[nr] = data->fan[nr] * old_div /
FAN_DIV_FROM_REG(data->fan_div[nr]);
adm1031_write_value(client, ADM1031_REG_FAN_DIV(nr),
data->fan_div[nr]);
adm1031_write_value(client, ADM1031_REG_FAN_MIN(nr),
data->fan_min[nr]);
/* Invalidate the cache: fan speed is no longer valid */
data->valid = 0;
mutex_unlock(&data->update_lock);
return count;
}
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment