Messages in this thread | | | Date | Tue, 7 Oct 2014 13:30:58 -0700 | Subject | Re: [RFC PATCH] pwm: Add the concept of an "active" pinctrl | From | Doug Anderson <> |
| |
Linus,
On Tue, Oct 7, 2014 at 6:15 AM, Linus Walleij <linus.walleij@linaro.org> wrote: > On Thu, Sep 25, 2014 at 10:26 PM, Doug Anderson <dianders@chromium.org> wrote: >> On Thu, Sep 25, 2014 at 12:31 AM, Linus Walleij >> <linus.walleij@linaro.org> wrote: >>> On Wed, Sep 24, 2014 at 1:17 AM, Doug Anderson <dianders@chromium.org> wrote: > >>> The device core will handle three states for you: "default", "sleep", "idle". >> >> Except it will handle only one state unless "CONFIG_PM" is defined, >> right? In other words "sleep" and "idle" totally don't exist in the >> "!CONFIG_PM" case. > > Yes. Because the intent of these settings is usually to save power. > So by disabling the PM config, you can test if the system works > without PM. Even hairy stuff like power-related pin control. It has > this nice orthogonal debug feature to it... > >> Normally it should be considered "safe" (though high power) >> to disable CONFIG_PM, right? > > Yeah. If your usecase makes it unsafe to have just "default", > you should not rely on these state names but invent new ones.
I thought about it a bit more and I decided that a new state made the most sense. I put another RFC up there at <https://patchwork.kernel.org/patch/5049041/>. Feel free to have a look at it.
-Doug
| |