Message ID | 1944274.r5ID26zhWd@wuerfel (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Fri, 07 Jun 2013 12:02:31 +0200 Arnd Bergmann <arnd@arndb.de> wrote: > On Friday 07 June 2013 10:39:20 Jingoo Han wrote: > > Add CONFIG_PM_SLEEP to suspend/resume functions to fix the following > > build warning when CONFIG_PM_SLEEP is not selected. This is because > > sleep PM callbacks defined by SIMPLE_DEV_PM_OPS are only used when > > the CONFIG_PM_SLEEP is enabled. > > > > drivers/video/backlight/backlight.c:211:12: warning: 'backlight_suspend' defined but not used [-Wunused-function] > > drivers/video/backlight/backlight.c:225:12: warning: 'backlight_resume' defined but not used [-Wunused-function] > > > > Signed-off-by: Jingoo Han <jg1.han@samsung.com> > > --- > > drivers/video/backlight/backlight.c | 2 ++ > > 1 file changed, 2 insertions(+) > > Your patch looks ok, but I find it extremely annoying to have new warnings > like this one come up every single day in linux-next. It really shouldn't > be this hard to use a macro called SIMPLE_DEV_PM_OPS() correctly. > > Below is an implementation of SIMPLE_DEV_PM_OPS and UNIVERSAL_DEV_PM_OPS > that avoids this issue by introducing an unused reference to the suspend > and resume functions. gcc is smart enough to leave out that unused code > by itself, and it would actually improve compile-time coverage to have > something like this, besides being harder to misuse. > > This would be a better approach if we didn't already have all the "#ifdef > CONFIG_PM_SLEEP" in place that hide the functions now. Unfortunately we > already have over 300 uses of SIMPLE_DEV_PM_OPS/UNIVERSAL_DEV_PM_OPS > in the kernel today, so removing all the #ifdef atomically without > creating more build errors is rather hard to do. > > Maybe someone has an idea how to extend my approach so it works with > and without the #ifdef, to let us transition to a situation that no > longer needs them. You could create new macros, and add a checkpatch rule to remind people to not use the old ones. Then people can migrate over from the old macros at a leisurely pace. The problem will be in thinking up decent names for the new macros. -- To unsubscribe from this list: send the line "unsubscribe linux-fbdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 16:31 Mon 10 Jun , Andrew Morton wrote: > On Fri, 07 Jun 2013 12:02:31 +0200 Arnd Bergmann <arnd@arndb.de> wrote: > > > On Friday 07 June 2013 10:39:20 Jingoo Han wrote: > > > Add CONFIG_PM_SLEEP to suspend/resume functions to fix the following > > > build warning when CONFIG_PM_SLEEP is not selected. This is because > > > sleep PM callbacks defined by SIMPLE_DEV_PM_OPS are only used when > > > the CONFIG_PM_SLEEP is enabled. > > > > > > drivers/video/backlight/backlight.c:211:12: warning: 'backlight_suspend' defined but not used [-Wunused-function] > > > drivers/video/backlight/backlight.c:225:12: warning: 'backlight_resume' defined but not used [-Wunused-function] > > > > > > Signed-off-by: Jingoo Han <jg1.han@samsung.com> > > > --- > > > drivers/video/backlight/backlight.c | 2 ++ > > > 1 file changed, 2 insertions(+) > > > > Your patch looks ok, but I find it extremely annoying to have new warnings > > like this one come up every single day in linux-next. It really shouldn't > > be this hard to use a macro called SIMPLE_DEV_PM_OPS() correctly. > > > > Below is an implementation of SIMPLE_DEV_PM_OPS and UNIVERSAL_DEV_PM_OPS > > that avoids this issue by introducing an unused reference to the suspend > > and resume functions. gcc is smart enough to leave out that unused code > > by itself, and it would actually improve compile-time coverage to have > > something like this, besides being harder to misuse. > > > > This would be a better approach if we didn't already have all the "#ifdef > > CONFIG_PM_SLEEP" in place that hide the functions now. Unfortunately we > > already have over 300 uses of SIMPLE_DEV_PM_OPS/UNIVERSAL_DEV_PM_OPS > > in the kernel today, so removing all the #ifdef atomically without > > creating more build errors is rather hard to do. > > > > Maybe someone has an idea how to extend my approach so it works with > > and without the #ifdef, to let us transition to a situation that no > > longer needs them. > > You could create new macros, and add a checkpatch rule to remind people > to not use the old ones. Then people can migrate over from the old > macros at a leisurely pace. > > The problem will be in thinking up decent names for the new macros. Agreed Best Regards, J. -- To unsubscribe from this list: send the line "unsubscribe linux-fbdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/include/linux/pm.h b/include/linux/pm.h index a224c7f..5a801bd 100644 --- a/include/linux/pm.h +++ b/include/linux/pm.h @@ -320,6 +320,9 @@ struct dev_pm_ops { #define SET_RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) #endif +#define UNUSED_DEV_PM_OPS(name, func...) \ +static const int (*__unused_ ## name[])(struct device*) __maybe_unused = { func }; + /* * Use this if you want to use the same suspend and resume callbacks for suspend * to RAM and hibernation. @@ -327,7 +332,8 @@ struct dev_pm_ops { #define SIMPLE_DEV_PM_OPS(name, suspend_fn, resume_fn) \ const struct dev_pm_ops name = { \ SET_SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) \ -} +}; \ +UNUSED_DEV_PM_OPS(name, suspend_fn, resume_fn) /* * Use this for defining a set of PM operations to be used in all situations @@ -346,7 +352,8 @@ const struct dev_pm_ops name = { \ const struct dev_pm_ops name = { \ SET_SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) \ SET_RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \ -} +}; \ +UNUSED_DEV_PM_OPS(name, suspend_fn, resume_fn, idle_fn) /** * PM_EVENT_ messages