Hello Journeyman and others,
I had the same problem with a HP Elitebook 8540w and a Samsung EVO 870 500GB. Here the chipset is one generation earlier, the Intel QM57.
The solution of changing the driver for the controller worked for me as well. However, this would not be a good solution if it led to noticeable performance hit on the drive.
The good news I have is that I measured the difference (the standard driver is provided by Microsoft) using the Samsung Magician and it seems to be insignificant.
However, for those wishing to keep the Intel drivers, here is what causes the problem and the way to avoid it.
The driver is lower layer of what Intel calls "Intel Rapid Storage Technology". This also consists of an application which - relying on the drivers - shows you the status of your SATA connected devices, monitors their "health" (and warns you if something is off) and gives you the ability to switch off power management of the controller. The version that applies to our generation (2010-14) of laptops is 12.9.0.1001 or thereabouts (same as the driver). This is how the GUI looks like.
The HELP part is reasonably well-written.
The problem is caused by the power management of the SATA controller by the Intel drivers. When this power management is on (a useful feature to save power during all those idle times, but comes with a little (unnoticeable for everyday use) latency when some disk operation starts again ), it seem to me that it switches the controller off when the computer is turning off and this state remains if power is not recycled (i.e. "warm" reboot).
You are able to switch off this power management at the "performance" part of the application:
If you switch it off, the reboot problem goes away with the Intel drivers (the above screenshot is when it is turned off).
Of course, the same trade-off question comes up: which is worth more, the ability to reboot without power recycle or the idle time power savings. This depends on your computer use habits (I guess during normal use you don't reboot that much) and how much power (--> carbon footprint) can be saved with the Intel power management on (to decide on this we would need data from a measurement by Intel in case of "typical" usage). Since windows also has its own power management (see the advanced part of the Power Options in Windows), I am hoping that there are still some savings when Intel's own power management is turned off.
Have a nice being-able-to-warm reboot-with intel drivers-day everyone:
UB
