Re: [PATCH v4 1/6] x86,sched: Add support for frequency invariance

From: Qais Yousef
Date: Tue Dec 24 2019 - 06:08:44 EST


On 12/23/19 17:16, Doug Smythies wrote:
> Yes, it fixes the schedutil governor behaving like the performance governor
> problem on my i7-2600K test system.
>
> I re-ran the tests several times, and re-booted back to the stock (problem)
> kernel to verify incorrect schedutil governor performance (i.e. I toggled
> back and forth, 2 times for each of 2 kernels, tests 1 and 2, total 8 tests).
> Kernel 5.5-rc2: 4 tests FAILED (as expected).
> Kernel 5.5-rc2 + this patch: 4 tests PASSED.

Great! Thanks for testing it, I'll send a proper patch shortly.

>
> Accidentally tested:
> Kernel 5.5-rc2 + this patch + command line "cgroup_no_v1=all": 1 test PASS.

I think this cgroup_no_v1 is a happy accident. It has nothing to do with the
fault, but for your case maybe helped observing things in a better way. FWIW,
I reproduced the issue on Juno Arm64 using Debian and Buildroot rootfs.

What is actually required to trigger the bug is to create a cpu controller and
add all system tasks to it to create some noise - which Ubuntu and Debian do by
default at boot time. In Buildroot when I did that manually I could see the
frequency going to max most of the time.

I'll add away to test this scenario.

Thanks for your detailed report.

Happy holidays!

--
Qais Yousef