diff mbox series

[v2,1/3] PM: introduce pm_ptr() macro

Message ID 20200413123207.74552-1-paul@crapouillou.net (mailing list archive)
State RFC
Headers show
Series [v2,1/3] PM: introduce pm_ptr() macro | expand

Commit Message

Paul Cercueil April 13, 2020, 12:32 p.m. UTC
This macro is analogous to the infamous of_match_ptr(). If CONFIG_PM
is enabled, this macro will resolve to its argument, otherwise to NULL.

Signed-off-by: Paul Cercueil <paul@crapouillou.net>
---

Notes:
    v2: Remove pm_sleep_ptr() macro

 include/linux/pm.h | 6 ++++++
 1 file changed, 6 insertions(+)

Comments

Rafael J. Wysocki April 26, 2020, 3:40 p.m. UTC | #1
On Monday, April 13, 2020 2:32:05 PM CEST Paul Cercueil wrote:
> This macro is analogous to the infamous of_match_ptr(). If CONFIG_PM
> is enabled, this macro will resolve to its argument, otherwise to NULL.

Well, this is going to result in quite a bit of unused code being
added to the kernels built with CONFIG_PM unset.

Is there any way to avoid that somehow?

> Signed-off-by: Paul Cercueil <paul@crapouillou.net>
> ---
> 
> Notes:
>     v2: Remove pm_sleep_ptr() macro
> 
>  include/linux/pm.h | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/include/linux/pm.h b/include/linux/pm.h
> index e057d1fa2469..1c0eec06905d 100644
> --- a/include/linux/pm.h
> +++ b/include/linux/pm.h
> @@ -374,6 +374,12 @@ const struct dev_pm_ops name = { \
>  	SET_RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
>  }
>  
> +#ifdef CONFIG_PM
> +#define pm_ptr(_ptr) (_ptr)
> +#else
> +#define pm_ptr(_ptr) NULL
> +#endif
> +
>  /*
>   * PM_EVENT_ messages
>   *
>
Paul Cercueil April 26, 2020, 5:06 p.m. UTC | #2
Hi Rafael,

Le dim. 26 avril 2020 à 17:40, Rafael J. Wysocki <rjw@rjwysocki.net> a 
écrit :
> On Monday, April 13, 2020 2:32:05 PM CEST Paul Cercueil wrote:
>>  This macro is analogous to the infamous of_match_ptr(). If CONFIG_PM
>>  is enabled, this macro will resolve to its argument, otherwise to 
>> NULL.
> 
> Well, this is going to result in quite a bit of unused code being
> added to the kernels built with CONFIG_PM unset.
> 
> Is there any way to avoid that somehow?

Using __maybe_unused on the dev_pm_ops struct and the callbacks, as 
long as they are static, they should be dropped by the compiler when 
CONFIG_PM is not set.

-Paul

> 
>>  Signed-off-by: Paul Cercueil <paul@crapouillou.net>
>>  ---
>> 
>>  Notes:
>>      v2: Remove pm_sleep_ptr() macro
>> 
>>   include/linux/pm.h | 6 ++++++
>>   1 file changed, 6 insertions(+)
>> 
>>  diff --git a/include/linux/pm.h b/include/linux/pm.h
>>  index e057d1fa2469..1c0eec06905d 100644
>>  --- a/include/linux/pm.h
>>  +++ b/include/linux/pm.h
>>  @@ -374,6 +374,12 @@ const struct dev_pm_ops name = { \
>>   	SET_RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
>>   }
>> 
>>  +#ifdef CONFIG_PM
>>  +#define pm_ptr(_ptr) (_ptr)
>>  +#else
>>  +#define pm_ptr(_ptr) NULL
>>  +#endif
>>  +
>>   /*
>>    * PM_EVENT_ messages
>>    *
>> 
> 
> 
> 
>
diff mbox series

Patch

diff --git a/include/linux/pm.h b/include/linux/pm.h
index e057d1fa2469..1c0eec06905d 100644
--- a/include/linux/pm.h
+++ b/include/linux/pm.h
@@ -374,6 +374,12 @@  const struct dev_pm_ops name = { \
 	SET_RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
 }
 
+#ifdef CONFIG_PM
+#define pm_ptr(_ptr) (_ptr)
+#else
+#define pm_ptr(_ptr) NULL
+#endif
+
 /*
  * PM_EVENT_ messages
  *