* [PATCH] PM: runtime: add system sleep callbacks to EXPORT_*_RUNTIME_DEV_PM_OPS()
@ 2026-03-01 1:22 Neel Bullywon
2026-03-06 18:30 ` Rafael J. Wysocki
0 siblings, 1 reply; 2+ messages in thread
From: Neel Bullywon @ 2026-03-01 1:22 UTC (permalink / raw)
To: rafael; +Cc: jic23, paul, linux-pm, linux-kernel, Neel Bullywon
DEFINE_RUNTIME_DEV_PM_OPS() sets pm_runtime_force_suspend() and
pm_runtime_force_resume() as system sleep callbacks via
_DEFINE_DEV_PM_OPS(), but the exported variants
(EXPORT_RUNTIME_DEV_PM_OPS, EXPORT_GPL_RUNTIME_DEV_PM_OPS,
EXPORT_NS_RUNTIME_DEV_PM_OPS, EXPORT_NS_GPL_RUNTIME_DEV_PM_OPS)
only set RUNTIME_PM_OPS() and omit SYSTEM_SLEEP_PM_OPS(). This
means drivers using the export macros do not get system sleep
support, contrary to what the documentation suggests.
Add SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
pm_runtime_force_resume) to all four export variants to match the
behavior of DEFINE_RUNTIME_DEV_PM_OPS().
All current users of these macros are IIO drivers (bme680, bmp280,
kxsd9, mpl115, bmi088-accel) and none implement custom system sleep
callbacks, so this change is safe for all existing callers.
Suggested-by: Jonathan Cameron <jic23@kernel.org>
Suggested-by: Paul Cercueil <paul@crapouillou.net>
Signed-off-by: Neel Bullywon <neelb2403@gmail.com>
---
include/linux/pm_runtime.h | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h
index 41037c513f06..c6a6a0569b11 100644
--- a/include/linux/pm_runtime.h
+++ b/include/linux/pm_runtime.h
@@ -42,18 +42,26 @@
#define EXPORT_RUNTIME_DEV_PM_OPS(name, suspend_fn, resume_fn, idle_fn) \
EXPORT_DEV_PM_OPS(name) = { \
+ SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, \
+ pm_runtime_force_resume) \
RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
}
#define EXPORT_GPL_RUNTIME_DEV_PM_OPS(name, suspend_fn, resume_fn, idle_fn) \
EXPORT_GPL_DEV_PM_OPS(name) = { \
+ SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, \
+ pm_runtime_force_resume) \
RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
}
#define EXPORT_NS_RUNTIME_DEV_PM_OPS(name, suspend_fn, resume_fn, idle_fn, ns) \
EXPORT_NS_DEV_PM_OPS(name, ns) = { \
+ SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, \
+ pm_runtime_force_resume) \
RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
}
#define EXPORT_NS_GPL_RUNTIME_DEV_PM_OPS(name, suspend_fn, resume_fn, idle_fn, ns) \
EXPORT_NS_GPL_DEV_PM_OPS(name, ns) = { \
+ SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, \
+ pm_runtime_force_resume) \
RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
}
--
2.44.0
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] PM: runtime: add system sleep callbacks to EXPORT_*_RUNTIME_DEV_PM_OPS()
2026-03-01 1:22 [PATCH] PM: runtime: add system sleep callbacks to EXPORT_*_RUNTIME_DEV_PM_OPS() Neel Bullywon
@ 2026-03-06 18:30 ` Rafael J. Wysocki
0 siblings, 0 replies; 2+ messages in thread
From: Rafael J. Wysocki @ 2026-03-06 18:30 UTC (permalink / raw)
To: Neel Bullywon; +Cc: rafael, jic23, paul, linux-pm, linux-kernel
On Sun, Mar 1, 2026 at 2:22 AM Neel Bullywon <neelb2403@gmail.com> wrote:
>
> DEFINE_RUNTIME_DEV_PM_OPS() sets pm_runtime_force_suspend() and
> pm_runtime_force_resume() as system sleep callbacks via
> _DEFINE_DEV_PM_OPS(),
Isn't that a mistake?
I'm not actually sure if it would work for everybody. Stuff like USB
may want to use the exact same callbacks for runtime PM and
system-wide PM.
> but the exported variants
> (EXPORT_RUNTIME_DEV_PM_OPS, EXPORT_GPL_RUNTIME_DEV_PM_OPS,
> EXPORT_NS_RUNTIME_DEV_PM_OPS, EXPORT_NS_GPL_RUNTIME_DEV_PM_OPS)
> only set RUNTIME_PM_OPS() and omit SYSTEM_SLEEP_PM_OPS(). This
> means drivers using the export macros do not get system sleep
> support, contrary to what the documentation suggests.
>
> Add SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
> pm_runtime_force_resume) to all four export variants to match the
> behavior of DEFINE_RUNTIME_DEV_PM_OPS().
>
> All current users of these macros are IIO drivers (bme680, bmp280,
> kxsd9, mpl115, bmi088-accel) and none implement custom system sleep
> callbacks, so this change is safe for all existing callers.
So maybe let IIO have their own set of macros?
> Suggested-by: Jonathan Cameron <jic23@kernel.org>
> Suggested-by: Paul Cercueil <paul@crapouillou.net>
> Signed-off-by: Neel Bullywon <neelb2403@gmail.com>
> ---
> include/linux/pm_runtime.h | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h
> index 41037c513f06..c6a6a0569b11 100644
> --- a/include/linux/pm_runtime.h
> +++ b/include/linux/pm_runtime.h
> @@ -42,18 +42,26 @@
>
> #define EXPORT_RUNTIME_DEV_PM_OPS(name, suspend_fn, resume_fn, idle_fn) \
> EXPORT_DEV_PM_OPS(name) = { \
> + SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, \
> + pm_runtime_force_resume) \
> RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
> }
> #define EXPORT_GPL_RUNTIME_DEV_PM_OPS(name, suspend_fn, resume_fn, idle_fn) \
> EXPORT_GPL_DEV_PM_OPS(name) = { \
> + SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, \
> + pm_runtime_force_resume) \
> RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
> }
> #define EXPORT_NS_RUNTIME_DEV_PM_OPS(name, suspend_fn, resume_fn, idle_fn, ns) \
> EXPORT_NS_DEV_PM_OPS(name, ns) = { \
> + SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, \
> + pm_runtime_force_resume) \
> RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
> }
> #define EXPORT_NS_GPL_RUNTIME_DEV_PM_OPS(name, suspend_fn, resume_fn, idle_fn, ns) \
> EXPORT_NS_GPL_DEV_PM_OPS(name, ns) = { \
> + SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, \
> + pm_runtime_force_resume) \
> RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
> }
>
> --
> 2.44.0
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-03-06 18:30 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-03-01 1:22 [PATCH] PM: runtime: add system sleep callbacks to EXPORT_*_RUNTIME_DEV_PM_OPS() Neel Bullywon
2026-03-06 18:30 ` Rafael J. Wysocki
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®