Rely on the disk to spin itself down
In other words, instead of playing games with watching
/sys/block/sdX/stat files and sending a STANDBY IMMEDIATE command when
these haven't changed for a while, just use the STANDBY command itself
with the timeout value.
This avoids having to wake up the daemon every five seconds. In
addition, this change ensures that device specific minimums are
respected.
We didn't do this in the original commit because while testing the
feature, some of disks, notably ones from WD, appeared to not respect
the STANDBY command. However, when moved to another enclosure (the
disks was initially in a SATA PM enclosure, it got moved to a SAS
enclosure) things started working again. Also turns out the WD disk in
question has a device specific timeout of 10 minutes.