Systems that initiate frequent suspend/resume from userspace
can make the kernel aware by enabling PM_USERSPACE_AUTOSLEEP
config.
This allows for certain sleep-sensitive code (wireguard/rng) to
decide on what preparatory work should be performed (or not) in
their pm_notification callbacks.
This patch was prompted by the discussion at [1] which attempts
to remove CONFIG_ANDROID that currently guards these code paths.
[1] https://lore.kernel.org/r/[email protected]/
Suggested-by: Jason A. Donenfeld <[email protected]>
Signed-off-by: Kalesh Singh <[email protected]>
---
drivers/char/random.c | 4 ++--
drivers/net/wireguard/device.c | 3 ++-
kernel/power/Kconfig | 20 ++++++++++++++++++++
3 files changed, 24 insertions(+), 3 deletions(-)
diff --git a/drivers/char/random.c b/drivers/char/random.c
index e3dd1dd3dd22..8c90f535d149 100644
--- a/drivers/char/random.c
+++ b/drivers/char/random.c
@@ -755,8 +755,8 @@ static int random_pm_notification(struct notifier_block *nb, unsigned long actio
spin_unlock_irqrestore(&input_pool.lock, flags);
if (crng_ready() && (action == PM_RESTORE_PREPARE ||
- (action == PM_POST_SUSPEND &&
- !IS_ENABLED(CONFIG_PM_AUTOSLEEP) && !IS_ENABLED(CONFIG_ANDROID)))) {
+ (action == PM_POST_SUSPEND && !IS_ENABLED(CONFIG_PM_AUTOSLEEP) &&
+ !IS_ENABLED(CONFIG_PM_USERSPACE_AUTOSLEEP)))) {
crng_reseed();
pr_notice("crng reseeded on system resumption\n");
}
diff --git a/drivers/net/wireguard/device.c b/drivers/net/wireguard/device.c
index aa9a7a5970fd..d58e9f818d3b 100644
--- a/drivers/net/wireguard/device.c
+++ b/drivers/net/wireguard/device.c
@@ -69,7 +69,8 @@ static int wg_pm_notification(struct notifier_block *nb, unsigned long action, v
* its normal operation rather than as a somewhat rare event, then we
* don't actually want to clear keys.
*/
- if (IS_ENABLED(CONFIG_PM_AUTOSLEEP) || IS_ENABLED(CONFIG_ANDROID))
+ if (IS_ENABLED(CONFIG_PM_AUTOSLEEP) ||
+ IS_ENABLED(CONFIG_PM_USERSPACE_AUTOSLEEP))
return 0;
if (action != PM_HIBERNATION_PREPARE && action != PM_SUSPEND_PREPARE)
diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig
index a12779650f15..60a1d3051cc7 100644
--- a/kernel/power/Kconfig
+++ b/kernel/power/Kconfig
@@ -143,6 +143,26 @@ config PM_AUTOSLEEP
Allow the kernel to trigger a system transition into a global sleep
state automatically whenever there are no active wakeup sources.
+config PM_USERSPACE_AUTOSLEEP
+ bool "Userspace opportunistic sleep"
+ depends on PM_SLEEP
+ help
+ Notify kernel of aggressive userspace autosleep power management policy.
+
+ This option changes the behavior of various sleep-sensitive code to deal
+ with frequent userspace-initiated transitions into a global sleep state.
+
+ Saying Y here, disables code paths that most users really should keep
+ enabled. In particular, only enable this if it is very common to be
+ asleep/awake for very short periods of time (<= 2 seconds).
+
+ Only platforms, such as Android, that implement opportunistic sleep from
+ a userspace power manager service should enable this option; and not
+ other machines. Therefore, you should say N here, unless you are
+ extremely certain that this is what you want. The option otherwise has
+ bad, undesirable effects, and should not be enabled just for fun.
+
+
config PM_WAKELOCKS
bool "User space wakeup sources interface"
depends on PM_SLEEP
base-commit: 03c765b0e3b4cb5063276b086c76f7a612856a9a
--
2.37.0.rc0.161.g10f37bed90-goog
Hi Kalesh,
On Thu, Jun 30, 2022 at 07:12:29PM +0000, Kalesh Singh wrote:
> Systems that initiate frequent suspend/resume from userspace
> can make the kernel aware by enabling PM_USERSPACE_AUTOSLEEP
> config.
>
> This allows for certain sleep-sensitive code (wireguard/rng) to
> decide on what preparatory work should be performed (or not) in
> their pm_notification callbacks.
>
> This patch was prompted by the discussion at [1] which attempts
> to remove CONFIG_ANDROID that currently guards these code paths.
>
> [1] https://lore.kernel.org/r/[email protected]/
>
> Suggested-by: Jason A. Donenfeld <[email protected]>
> Signed-off-by: Kalesh Singh <[email protected]>
Thanks, looks good to me. Do you have a corresponding Gerrit link to the
change adding this to the base Android kernel config? If so, have my
Ack:
Acked-by: Jason A. Donenfeld <[email protected]>
Jason
On Thu, Jun 30, 2022 at 12:49 PM Jason A. Donenfeld <[email protected]> wrote:
>
> Hi Kalesh,
>
> On Thu, Jun 30, 2022 at 07:12:29PM +0000, Kalesh Singh wrote:
> > Systems that initiate frequent suspend/resume from userspace
> > can make the kernel aware by enabling PM_USERSPACE_AUTOSLEEP
> > config.
> >
> > This allows for certain sleep-sensitive code (wireguard/rng) to
> > decide on what preparatory work should be performed (or not) in
> > their pm_notification callbacks.
> >
> > This patch was prompted by the discussion at [1] which attempts
> > to remove CONFIG_ANDROID that currently guards these code paths.
> >
> > [1] https://lore.kernel.org/r/[email protected]/
> >
> > Suggested-by: Jason A. Donenfeld <[email protected]>
> > Signed-off-by: Kalesh Singh <[email protected]>
>
> Thanks, looks good to me. Do you have a corresponding Gerrit link to the
> change adding this to the base Android kernel config? If so, have my
> Ack:
>
> Acked-by: Jason A. Donenfeld <[email protected]>
Hi Jason,
Our latest supported kernels in Android are based on 5.15 so the
config change isn't yet needed. Once there are newer versions with the
CONFIG_ANDROID removed I will add this to the defconfig.
Thanks,
Kalesh
>
> Jason
>
> --
> To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
>
On Thu, Jun 30, 2022 at 01:41:40PM -0700, Kalesh Singh wrote:
> Our latest supported kernels in Android are based on 5.15 so the
> config change isn't yet needed. Once there are newer versions with the
> CONFIG_ANDROID removed I will add this to the defconfig.
Okay. It might be still worth getting something uploaded to gerrit so
that it's easy to remember and submit whenever the time comes.
Also, what about android running on mainline? Where does that base
config live?
Jason
Hi Kalesh,
On Thu, Jun 30, 2022 at 03:02:46PM -0700, Kalesh Singh wrote:
> I've uploaded the changes on android-mainline [1]. We'll submit there
> once the upstream changes are finalized.
>
> [1] https://android-review.googlesource.com/c/kernel/common/+/2142693/1
Excellent. I think everything is all set then, at least from my
perspective. There's a viable replacement for this usage of
CONFIG_ANDROID, there are patches ready to go both in the kernel and on
Android's configs, and now all we do is wait for Rafael. Great!
Maybe people will have opinions on the naming
(CONFIG_PM_RAPID_USERSPACE_AUTOSLEEP vs
CONFIG_PM_ANDROID_USERAPCE_AUTO_SLEEP vs what you have vs something else
vs who knows), but whatever is chosen seems probably fine, as this is a
pretty low key change since it can always be tweaked further later (it's
not ABI).
Jason
On Thu, Jun 30, 2022 at 2:14 PM Jason A. Donenfeld <[email protected]> wrote:
>
> On Thu, Jun 30, 2022 at 01:41:40PM -0700, Kalesh Singh wrote:
> > Our latest supported kernels in Android are based on 5.15 so the
> > config change isn't yet needed. Once there are newer versions with the
> > CONFIG_ANDROID removed I will add this to the defconfig.
>
> Okay. It might be still worth getting something uploaded to gerrit so
> that it's easy to remember and submit whenever the time comes.
>
> Also, what about android running on mainline? Where does that base
> config live?
I've uploaded the changes on android-mainline [1]. We'll submit there
once the upstream changes are finalized.
[1] https://android-review.googlesource.com/c/kernel/common/+/2142693/1
Thanks,
Kalesh
>
> Jason
On Thu, Jun 30, 2022 at 09:49:24PM +0200, Jason A. Donenfeld wrote:
> Hi Kalesh,
>
> On Thu, Jun 30, 2022 at 07:12:29PM +0000, Kalesh Singh wrote:
> > Systems that initiate frequent suspend/resume from userspace
> > can make the kernel aware by enabling PM_USERSPACE_AUTOSLEEP
> > config.
> >
> > This allows for certain sleep-sensitive code (wireguard/rng) to
> > decide on what preparatory work should be performed (or not) in
> > their pm_notification callbacks.
> >
> > This patch was prompted by the discussion at [1] which attempts
> > to remove CONFIG_ANDROID that currently guards these code paths.
> >
> > [1] https://lore.kernel.org/r/[email protected]/
> >
> > Suggested-by: Jason A. Donenfeld <[email protected]>
> > Signed-off-by: Kalesh Singh <[email protected]>
>
> Thanks, looks good to me. Do you have a corresponding Gerrit link to the
> change adding this to the base Android kernel config? If so, have my
> Ack:
>
> Acked-by: Jason A. Donenfeld <[email protected]>
Cool, I'll queue this up and also the CONFIG_ANDROID removal into my
tree now, thanks all for working it out!
greg k-h