Re: [PATCH 1/2] host: ehci-exynos: Convert to use the SET_SYSTEM_SLEEP_PM_OPS

[Anand Moon]

Hi Krzysztof,

On 10 October 2016 at 00:09, Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote:
> On Sun, Oct 09, 2016 at 11:57:59PM +0530, Anand Moon wrote:
>> hi Krzysztof,
>>
>> On 9 October 2016 at 22:57, Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote:
>> > On Sun, Oct 09, 2016 at 10:45:40PM +0530, Anand Moon wrote:
>> >> Hi Krzysztof,
>> >>
>> >> On 9 October 2016 at 22:04, Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote:
>> >> > On Sun, Oct 09, 2016 at 02:34:14PM +0000, Anand Moon wrote:
>> >> >> Move the ehci-exynos system PM callbacks within #ifdef CONFIG_PM_SLEEP
>> >> >> as to avoid them being build when not used. This also allows us to use the
>> >> >> SET_SYSTEM_SLEEP_PM_OPS macro which simplifies the code.
>> >> >>
>> >> >> Signed-off-by: Anand Moon <linux.amoon@xxxxxxxxx>
>> >> >> ---
>> >> >>  drivers/usb/host/ehci-exynos.c | 14 ++++++--------
>> >> >>  1 file changed, 6 insertions(+), 8 deletions(-)
>> >> >>
>> >> >> diff --git a/drivers/usb/host/ehci-exynos.c b/drivers/usb/host/ehci-exynos.c
>> >> >> index 42e5b66..1899900 100644
>> >> >> --- a/drivers/usb/host/ehci-exynos.c
>> >> >> +++ b/drivers/usb/host/ehci-exynos.c
>> >> >> @@ -251,7 +251,7 @@ static int exynos_ehci_remove(struct platform_device *pdev)
>> >> >>       return 0;
>> >> >>  }
>> >> >>
>> >> >> -#ifdef CONFIG_PM
>> >> >> +#ifdef CONFIG_PM_SLEEP
>> >> >
>> >> > Does not look like an equivalent change. How will it behave in a config
>> >> > with !SUSPEND && !HIBERNATE && PM?
>> >> >
>> >>
>> >> [snip]
>> >>
>> >> I just wanted to update suspend and resume callback to use
>> >> SET_SYSTEM_SLEEP_PM_OPS
>> >> as they are define under CONFIG_PM_SLEEP so I update above to avoid
>> >> compilation warning/error.
>> >
>> Apologize: for not understanding your question.
>>
>> > First of all you did not answer to my question, so let me rephrase into
>> > two:
>> > 1. Is the code equivalent?
>>
>> No CONFIG_PM and CONFIG_PM_SLEEP are different options.
>> But I could not disable CONFIG_PM_SLEEP option with either in exynos_defconfig
>
> So the code is not equivalent...

I might be wrong, below is the kconfig option for PM_SLEEP

 Symbol: PM_SLEEP [=y]
 Type  : boolean
    Defined at kernel/power/Kconfig
     Depends on: SUSPEND [=y] || HIBERNATE_CALLBACKS [=n]
     Selects: PM [=y]

So we cannot set CONFIG_PM_SLEEP=n and CONFIG_PM=y

I observed at many places were either CONFIG_PM or CONFIG_PM_SLEEP are used.

So I would like to use SIMPLE_DEV_PM_OPS macro to set struct
dev_pm_ops exynos_ohci_pm_ops to correct the code.

Best Regards
-Anand Moon

>
>>
>> CONFIG_PM_SLEEP=n or
>> # CONFIG_PM_SLEEP is not set
>>
>> > 2. What will be the output with !SUSPEND && !HIBERNATE && PM?
>>
>> #
>> # Power management options
>> #
>> # CONFIG_SUSPEND is not set
>> # CONFIG_HIBERNATION is not set
>> # CONFIG_PM is not set
>>
>> When CONFIG_SUSPEND and CONFIG_HIBERNATION are not set
>> CONFIG_PM is disabled and so is CONFIG_PM_SLEEP.
>
> In my config, the CONFIG_PM was enabled thus the code changes the
> functionality... Maybe this was intented but I really don't get it from
> the commit message or from your explanations here.
>
> Krzysztof