linux-kernel - Re: [intel-pstate driver regression] processor frequency very high even if in idle

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 linux-cve-announce 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:	Mon, 4 Apr 2016 07:13:51 +0200
From:	Sedat Dilek <sedat.dilek@...il.com>
To:	Doug Smythies <dsmythies@...us.net>
Cc:	Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>,
	"Rafael J. Wysocki" <rafael.j.wysocki@...el.com>,
	Linux PM list <linux-pm@...r.kernel.org>,
	LKML <linux-kernel@...r.kernel.org>,
	Jörg Otte <jrg.otte@...il.com>
Subject: Re: [intel-pstate driver regression] processor frequency very high
 even if in idle

On Sun, Apr 3, 2016 at 8:59 PM, Doug Smythies <dsmythies@...us.net> wrote:
> On 2016.04.02 11:21 Sedat Dilek wrote:
>> On Sat, Apr 2, 2016 at 7:19 PM, Jörg Otte wrote:
>>> 2016-04-02 17:28 GMT+02:00 Srinivas Pandruvada wrote:
>>>>

Hi Doug,

are you involved in the Ubuntu-OS? Developer for Canonical?

>>>> If you are using Ubuntu, the OS has a script which will automatically
>>>> change from performance.
>>>> Doug can give more information on this script.
>>
>>> maybe:
>>> /etc/init.d/ondemand
>
> Yes.
>
>> With CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y (linux-pm.git#linux-next) I get...
>>
>> $ cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_driver
>> acpi-cpufreq
>> acpi-cpufreq
>> acpi-cpufreq
>> acpi-cpufreq
>>
>> $ cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
>> ondemand
>> ondemand
>> ondemand
>> ondemand
>
> Yes, those are the expected results for the acpi-cpufreq CPU frequency scaling driver.
> You should be able to observe the governor set to sched util for the first minute
> after re-boot and/or if you set it yourself after the /etc/init.d/ondemand script
> has finished (i.e. more than 1 minute after re-boot.)
>

I have hardcoded to use "schedutil" driver after one minute in
/etc/init.d/ondemand for testing-purposes and not to fall back to
"ondemand".

>> ...is there a difference when using intel_pstate as scaling_driver?
>
> Yes, but only because there are different available governors for the two drivers.
>
>> Are the scripts of Ubuntu working properly with acpi-cpufreq (only)?
>
> As far as I know the /etc/init.d/ondemand is working properly. It sets the acpi-cpufreq
> driver to use the "ondemand" governor and it sets the intel_pstate driver to use the
> "powersave" governor.
>

I haven't looked at this exactly.
The ondemand-script is not saying to fall back to "powersave" in case
of intel_pstate-driver (here on Ubuntu/precise).

- Sedat -