svn commit: r264250 - head/sys/dev/acpica

[Adrian Chadd]

On 8 April 2014 06:14, John Baldwin <jhb at freebsd.org> wrote:
> On Monday, April 07, 2014 10:36:27 pm Adrian Chadd wrote:
>> Author: adrian
>> Date: Tue Apr  8 02:36:27 2014
>> New Revision: 264250
>> URL: http://svnweb.freebsd.org/changeset/base/264250
>>
>> Log:
>>   Add a basic set of data points which count the number of sleep entries
>>   that are being done by the OS.
>>
>>   For now this'll match up with the "wakeups"; although I'll dig deeper into
>>   this to see if we can determine which sleep state the CPU managed to get
>>   into.  Most things I've seen these days only expose up to C2 or C3 via
>>   ACPI even though the CPU goes all the way down to C6 or C7.
>
> No, those are actually the same thing.  ACPI and Intel both use C-states for
> the same thing, but the numbers don't line up.  That is, Intel's C6/C7 gets
> exposed to the OS as C2/C3 via ACPI.  The 6/7 does matter, (I think) if you
> are using monitor/mwait as I believe the value you configure for an mwait
> sleep has to use Intel's number (6/7) whereas the ACPI number (2/3) is
> assigned by the results of _CST or whichever object it is ACPI queries.
>
> All that to say that ACPI is already using Intel's C6/C7 if you have
> configured your BIOS to expose it.

I know that the C2/C3 states are whatever ACPI and the CPU decides to
map things to. What I'm curious about is whether we can get some kind
of feedback from the CPU when it wakes up about what sleep state it
actually made it down to.


-a