Hi,
On 10/14/2013 07:02 PM, Peter Zijlstra wrote:
On Fri, Oct 11, 2013 at 06:19:10PM +0100, Morten Rasmussen wrote:
Hi,
I have revised the previous power scheduler proposal[1] trying to address as many of the comments as possible. The overall idea was discussed at LPC[2,3]. The revised design has removed the power scheduler and replaced it with a high level power driver interface. An interface that allows the scheduler to query the power driver for information and provide hints to guide power management decisions in the power driver.
The power driver is going to be a unified platform power driver that can replace cpufreq and cpuidle drivers. Generic power policies will be optional helper functions called from the power driver. Platforms may choose to implement their own policies as part of their power driver.
This RFC series prototypes a part of the power driver interface (cpu capacity hints) and shows how they can be used from the scheduler. More extensive use of the power driver hints and queries is left for later. The focus for now is the power driver interface. The patch series includes a power driver/cpufreq governor that can use existing cpufreq drivers as backend. It has been tested (not thoroughly) on ARM TC2. The cpufreq governor power driver implementation is rather horrible, but it illustrates how the power driver interface can be used. Native power drivers is on the todo list.
The power driver interface is still missing quite a few calls to handle: Idle, adding extra information to the sched_domain hierarchy to guide scheduling decisions (packing), and possibly scaling of tracked load to compensate for frequency changes and asymmetric systems (big.LITTLE).
This set is based on 3.11. I have done ARM TC2 testing based on linux-linaro 2013.08[4] to get cpufreq support for TC2.
What I'm missing is a general overview of why what and how.
I agree that the "why" needs to be mentioned very clearly since the patchset revolves around it. As far as I understand we need a single controller for deciding the power efficiency of the kernel, who is exposed to all the user policies and the frequency+idle states stats of the CPU to begin with. These stats are being supplied by the power driver.
Having these details and decision making in multiple places like we do today in cpuidle, cpu-frequency and scheduler will probably cause problems. For example, when the power efficiency of the kernel goes wrong we have trouble point out the reason behind it. Where did the problem arise from among the above three power policy decision makers? This is a maintainability concern. Another reason is the power saving decisions made by say cpuidle may not complement the power saving decisions made by cpufreq. This can lead to inconsistent results across different workloads.
Thus having a single policy maker for power savings we are hoping to solve the primary concerns of consistent behaviour from the kernel in terms of power efficiency and improved maintainability.
In particular; how does this proposal lead to power savings. Is there a mathematical model that supports this framework? Something where if you give it a task set with global utilisation < 1 (ie. there's idle time), it results in less power used.
AFAIK, this patchset is an attempt to achieve consistency in the power efficiency of the kernel across workloads with the existing algorithms, in addition to a cleanup involving integration of the power policy making in one place as explained above. In an attempt to do so, *maybe* better power numbers can be obtained or at-least the default power efficiency of the kernel will show up.
However adding the new patchsets like packing small tasks, heterogeneous scheduling, power aware scheduling etc.. *should* then yield good and consistent power savings since they now stand on top of an integrated stable power driver.
Regards Preeti U Murthy
Also, how does this proposal deal with cpufreq's fundamental broken approach to SMP? Afaict nothing considers the effect of one cpu upon another -- something which isn't true at all.
In fact, I don't see anything except a random bunch of hooks without an over-all picture of how to get less power used.