RE: [PATCH v3] cpuidle: poll_state: Add time limit to poll_idle()

[Doug Smythies]

On 2018.03.14 07:09 Rafael J. Wysocki wrote:

... [snip]...

> v2 -> v3: Use local_clock() for time measurements and drop the
>          counter, since that should be lightweight enough (as
>          suggested by Peter).

I have been testing the latest of everything for a couple of days
now, and everything continues to be great.

Note that I was using a POLL_IDLE_TIME_CHECK_COUNT of 1 anyhow, because
I specifically wanted to test the worst case time through the loop.
i.e. I wanted any potential issue to be 1000 times more likely to find.
My problem is that I don't know of a good test for this specifically.

I'll switch to this V3, along with V4 of the "sched/cpuidle: Idle loop
rework" 7 patch set.

As for energy savings for just this patch only, I would refer readers
to my previous test results from late November, [1], as I haven't
re-done those Phoronix tests yet, but I don't expect the results to
differ much.

[1] https://marc.info/?l=linux-pm&m=151154499710125&w=2

... Doug