Re: [PATCH 3/8] PM: core: Add EXPORT[_GPL]_SIMPLE_DEV_PM_OPS macros

[Jonathan Cameron]

On Wed, 5 Jan 2022 10:15:36 +0000
Paul Cercueil <paul@xxxxxxxxxxxxxxx> wrote:

> Hi Jonathan,
> 
> Le mer., janv. 5 2022 at 10:03:32 +0000, Jonathan Cameron 
> <Jonathan.Cameron@xxxxxxxxxx> a écrit :
> > On Tue, 4 Jan 2022 21:42:09 +0000
> > Paul Cercueil <paul@xxxxxxxxxxxxxxx> wrote:
> >   
> >>  These macros are defined conditionally, according to CONFIG_PM:
> >>  - if CONFIG_PM is enabled, these macros resolve to
> >>    DEFINE_SIMPLE_DEV_PM_OPS(), and the dev_pm_ops symbol will be
> >>    exported.
> >> 
> >>  - if CONFIG_PM is disabled, these macros will result in a dummy 
> >> static
> >>    dev_pm_ops to be created with the __maybe_unused flag. The 
> >> dev_pm_ops
> >>    will then be discarded by the compiler, along with the provided
> >>    callback functions if they are not used anywhere else.
> >> 
> >>  In the second case, the symbol is not exported, which should be
> >>  perfectly fine - users of the symbol should all use the pm_ptr() or
> >>  pm_sleep_ptr() macro, so the dev_pm_ops marked as "extern" in the
> >>  client's code will never be accessed.
> >> 
> >>  Signed-off-by: Paul Cercueil <paul@xxxxxxxxxxxxxxx>
> >>  ---
> >>   include/linux/pm.h | 33 ++++++++++++++++++++++++++++++---
> >>   1 file changed, 30 insertions(+), 3 deletions(-)
> >> 
> >>  diff --git a/include/linux/pm.h b/include/linux/pm.h
> >>  index 389e600df233..a1ce29566aea 100644
> >>  --- a/include/linux/pm.h
> >>  +++ b/include/linux/pm.h
> >>  @@ -8,6 +8,7 @@
> >>   #ifndef _LINUX_PM_H
> >>   #define _LINUX_PM_H
> >> 
> >>  +#include <linux/export.h>
> >>   #include <linux/list.h>
> >>   #include <linux/workqueue.h>
> >>   #include <linux/spinlock.h>
> >>  @@ -357,14 +358,40 @@ struct dev_pm_ops {
> >>   #define SET_RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn)
> >>   #endif
> >> 
> >>  +#define _DEFINE_DEV_PM_OPS(name, \
> >>  +			   suspend_fn, resume_fn, \
> >>  +			   runtime_suspend_fn, runtime_resume_fn, idle_fn) \
> >>  +const struct dev_pm_ops name = { \
> >>  +	SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) \
> >>  +	RUNTIME_PM_OPS(runtime_suspend_fn, runtime_resume_fn, idle_fn) \
> >>  +}
> >>  +  
> > 
> > one blank line probably enough.
> >   
> >>  +
> >>   /*
> >>    * Use this if you want to use the same suspend and resume 
> >> callbacks for suspend
> >>    * to RAM and hibernation.
> >>    */
> >>   #define DEFINE_SIMPLE_DEV_PM_OPS(name, suspend_fn, resume_fn) \
> >>  -const struct dev_pm_ops name = { \
> >>  -	SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) \
> >>  -}
> >>  +	_DEFINE_DEV_PM_OPS(name, suspend_fn, resume_fn, NULL, NULL, NULL)
> >>  +
> >>  +#ifdef CONFIG_PM
> >>  +#define _EXPORT_DEV_PM_OPS(name, suspend_fn, resume_fn, 
> >> runtime_suspend_fn, \
> >>  +			   runtime_resume_fn, idle_fn, sec) \
> >>  +	_DEFINE_DEV_PM_OPS(name, suspend_fn, resume_fn, 
> >> runtime_suspend_fn, \
> >>  +			   runtime_resume_fn, idle_fn); \
> >>  +	_EXPORT_SYMBOL(name, sec)
> >>  +#else
> >>  +#define _EXPORT_DEV_PM_OPS(name, suspend_fn, resume_fn, 
> >> runtime_suspend_fn, \
> >>  +			   runtime_resume_fn, idle_fn, sec) \
> >>  +static __maybe_unused _DEFINE_DEV_PM_OPS(__static_##name, 
> >> suspend_fn, \
> >>  +					 resume_fn, runtime_suspend_fn, \
> >>  +					 runtime_resume_fn, idle_fn)
> >>  +#endif
> >>  +
> >>  +#define EXPORT_SIMPLE_DEV_PM_OPS(name, suspend_fn, resume_fn) \
> >>  +	_EXPORT_DEV_PM_OPS(name, suspend_fn, resume_fn, NULL, NULL, NULL, 
> >> "")
> >>  +#define EXPORT_GPL_SIMPLE_DEV_PM_OPS(name, suspend_fn, resume_fn) \
> >>  +	_EXPORT_DEV_PM_OPS(name, suspend_fn, resume_fn, NULL, NULL, NULL, 
> >> "_gpl")  
> > 
> > So you can get away with these two cases because the 
> > SYSTEM_SLEEP_PM_OPS() all have
> > pm_sleep_ptr() wrappers.  However, _EXPORT_DEV_PM_OPS() could be used 
> > directly and
> > would require __maybe_unused for the RUNTIME_PM_OPS() parameters 
> > which isn't ideal.  
> 
> I don't see why. On both cases (CONFIG_PM enabled/disabled) the 
> runtime-PM callbacks are referenced directly, so at no point do they 
> appear as unused; therefore __maybe_unused is not needed.

Ah. I'd miss followed things through. Indeed the 'magic' __static_xxx_pm_ops
structure maintains a reference that the compiler can then remove.
On the plus side, turned out I'd not done a full set of tests with my
own patch set and found one bug in that :)

Acked-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>

> 
> Cheers,
> -Paul
> 
> > Maybe I'm missing some reason that isn't a problem though as easy to 
> > get lost in
> > these macros. :)
> > 
> > You could argue that the _ is meant to indicate that macro shouldn't 
> > be used directly
> > but I'm not that optimistic.
> > 
> > Jonathan
> > 
> > 
> >   
> >> 
> >>   /* Deprecated. Use DEFINE_SIMPLE_DEV_PM_OPS() instead. */
> >>   #define SIMPLE_DEV_PM_OPS(name, suspend_fn, resume_fn) \  
> >   
> 
>