Hi,
On Saturday, January 21, 2012, Arve Hjønnevåg wrote:
> Add an ioctl, EVIOCSSUSPENDBLOCK, to block suspend while the event
> queue is not empty. This allows userspace code to process input
> events while the device appears to be asleep.
I have two questions to start with, if you don't mind:
(1) Does Android user space use an analogous interface right now or is it
a prototype?
(2) What exactly are the use cases for it (ie. what problem does it address
that cannot be addressed in a different way)?
Rafael
> Signed-off-by: Arve Hjønnevåg <[email protected]>
> ---
> drivers/input/evdev.c | 63 +++++++++++++++++++++++++++++++++++++++++++++++++
> include/linux/input.h | 3 ++
> 2 files changed, 66 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/input/evdev.c b/drivers/input/evdev.c
> index 76457d5..e212757 100644
> --- a/drivers/input/evdev.c
> +++ b/drivers/input/evdev.c
> @@ -43,6 +43,7 @@ struct evdev_client {
> unsigned int tail;
> unsigned int packet_head; /* [future] position of the first element of next packet */
> spinlock_t buffer_lock; /* protects access to buffer, head and tail */
> + struct wakeup_source *wakeup_source;
> struct fasync_struct *fasync;
> struct evdev *evdev;
> struct list_head node;
> @@ -75,10 +76,14 @@ static void evdev_pass_event(struct evdev_client *client,
> client->buffer[client->tail].value = 0;
>
> client->packet_head = client->tail;
> + if (client->wakeup_source)
> + __pm_relax(client->wakeup_source);
> }
>
> if (event->type == EV_SYN && event->code == SYN_REPORT) {
> client->packet_head = client->head;
> + if (client->wakeup_source)
> + __pm_stay_awake(client->wakeup_source);
> kill_fasync(&client->fasync, SIGIO, POLL_IN);
> }
>
> @@ -255,6 +260,10 @@ static int evdev_release(struct inode *inode, struct file *file)
> mutex_unlock(&evdev->mutex);
>
> evdev_detach_client(evdev, client);
> + if (client->wakeup_source) {
> + __pm_relax(client->wakeup_source);
> + wakeup_source_unregister(client->wakeup_source);
> + }
> kfree(client);
>
> evdev_close_device(evdev);
> @@ -373,6 +382,9 @@ static int evdev_fetch_next_event(struct evdev_client *client,
> if (have_event) {
> *event = client->buffer[client->tail++];
> client->tail &= client->bufsize - 1;
> + if (client->wakeup_source &&
> + client->packet_head == client->tail)
> + __pm_relax(client->wakeup_source);
> }
>
> spin_unlock_irq(&client->buffer_lock);
> @@ -623,6 +635,48 @@ static int evdev_handle_set_keycode_v2(struct input_dev *dev, void __user *p)
> return input_set_keycode(dev, &ke);
> }
>
> +static int evdev_enable_suspend_block(struct evdev *evdev,
> + struct evdev_client *client)
> +{
> + struct wakeup_source *ws;
> + char name[28];
> +
> + if (client->wakeup_source)
> + return 0;
> +
> + snprintf(name, sizeof(name), "%s-%d",
> + dev_name(&evdev->dev), task_tgid_vnr(current));
> +
> + ws = wakeup_source_register(name);
> + if (!ws)
> + return -ENOMEM;
> +
> + spin_lock_irq(&client->buffer_lock);
> + client->wakeup_source = ws;
> + if (client->packet_head != client->tail)
> + __pm_stay_awake(client->wakeup_source);
> + spin_unlock_irq(&client->buffer_lock);
> + return 0;
> +}
> +
> +static int evdev_disable_suspend_block(struct evdev *evdev,
> + struct evdev_client *client)
> +{
> + struct wakeup_source *ws;
> +
> + spin_lock_irq(&client->buffer_lock);
> + ws = client->wakeup_source;
> + client->wakeup_source = NULL;
> + spin_unlock_irq(&client->buffer_lock);
> +
> + if (ws) {
> + __pm_relax(ws);
> + wakeup_source_unregister(ws);
> + }
> +
> + return 0;
> +}
> +
> static long evdev_do_ioctl(struct file *file, unsigned int cmd,
> void __user *p, int compat_mode)
> {
> @@ -696,6 +750,15 @@ static long evdev_do_ioctl(struct file *file, unsigned int cmd,
>
> case EVIOCSKEYCODE_V2:
> return evdev_handle_set_keycode_v2(dev, p);
> +
> + case EVIOCGSUSPENDBLOCK:
> + return put_user(!!client->wakeup_source, ip);
> +
> + case EVIOCSSUSPENDBLOCK:
> + if (p)
> + return evdev_enable_suspend_block(evdev, client);
> + else
> + return evdev_disable_suspend_block(evdev, client);
> }
>
> size = _IOC_SIZE(cmd);
> diff --git a/include/linux/input.h b/include/linux/input.h
> index 3862e32..daf0177 100644
> --- a/include/linux/input.h
> +++ b/include/linux/input.h
> @@ -129,6 +129,9 @@ struct input_keymap_entry {
>
> #define EVIOCGRAB _IOW('E', 0x90, int) /* Grab/Release device */
>
> +#define EVIOCGSUSPENDBLOCK _IOR('E', 0x91, int) /* get suspend block enable */
> +#define EVIOCSSUSPENDBLOCK _IOW('E', 0x91, int) /* set suspend block enable */
> +
> /*
> * Device properties and quirks
> */
>
2012/2/11 Rafael J. Wysocki <[email protected]>:
> Hi,
>
> On Saturday, January 21, 2012, Arve Hj?nnev?g wrote:
>> Add an ioctl, EVIOCSSUSPENDBLOCK, to block suspend while the event
>> queue is not empty. This allows userspace code to process input
>> events while the device appears to be asleep.
>
> I have two questions to start with, if you don't mind:
>
> (1) Does Android user space use an analogous interface right now or is it
> ? ?a prototype?
>
Yes (for the next version), but with a wakelock based implementation.
> (2) What exactly are the use cases for it (ie. what problem does it address
> ? ?that cannot be addressed in a different way)?
>
The reason for adding an ioctl versus the old android version with
used a wakelock for all input events, is that not all input events are
wakeup events. Specifically some sensors generate input events at such
a high rate that they prevent suspend while the sensor is on even
though the driver has a suspend hook that turns the sensor off.
If you are asking why input events need to block suspend at all, this
was the example used in the suspend blocker documentation:
- The Keypad driver gets an interrupt. It then calls suspend_block on the
keypad-scan suspend_blocker and starts scanning the keypad matrix.
- The keypad-scan code detects a key change and reports it to the input-event
driver.
- The input-event driver sees the key change, enqueues an event, and calls
suspend_block on the input-event-queue suspend_blocker.
- The keypad-scan code detects that no keys are held and calls suspend_unblock
on the keypad-scan suspend_blocker.
- The user-space input-event thread returns from select/poll, calls
suspend_block on the process-input-events suspend_blocker and then calls read
on the input-event device.
- The input-event driver dequeues the key-event and, since the queue is now
empty, it calls suspend_unblock on the input-event-queue suspend_blocker.
- The user-space input-event thread returns from read. If it determines that
the key should be ignored, it calls suspend_unblock on the
process_input_events suspend_blocker and then calls select or poll. The
system will automatically suspend again, since now no suspend blockers are
active.
--
Arve Hj?nnev?g
Hi,
On Tuesday, February 14, 2012, Arve Hj?nnev?g wrote:
> 2012/2/11 Rafael J. Wysocki <[email protected]>:
> > Hi,
> >
> > On Saturday, January 21, 2012, Arve Hj?nnev?g wrote:
> >> Add an ioctl, EVIOCSSUSPENDBLOCK, to block suspend while the event
> >> queue is not empty. This allows userspace code to process input
> >> events while the device appears to be asleep.
> >
> > I have two questions to start with, if you don't mind:
> >
> > (1) Does Android user space use an analogous interface right now or is it
> > a prototype?
> >
>
> Yes (for the next version), but with a wakelock based implementation.
That's exactly what I wanted to know. :-)
> > (2) What exactly are the use cases for it (ie. what problem does it address
> > that cannot be addressed in a different way)?
> >
>
> The reason for adding an ioctl versus the old android version with
> used a wakelock for all input events, is that not all input events are
> wakeup events. Specifically some sensors generate input events at such
> a high rate that they prevent suspend while the sensor is on even
> though the driver has a suspend hook that turns the sensor off.
>
> If you are asking why input events need to block suspend at all, this
> was the example used in the suspend blocker documentation:
> - The Keypad driver gets an interrupt. It then calls suspend_block on the
> keypad-scan suspend_blocker and starts scanning the keypad matrix.
> - The keypad-scan code detects a key change and reports it to the input-event
> driver.
> - The input-event driver sees the key change, enqueues an event, and calls
> suspend_block on the input-event-queue suspend_blocker.
> - The keypad-scan code detects that no keys are held and calls suspend_unblock
> on the keypad-scan suspend_blocker.
> - The user-space input-event thread returns from select/poll, calls
> suspend_block on the process-input-events suspend_blocker and then calls read
> on the input-event device.
> - The input-event driver dequeues the key-event and, since the queue is now
> empty, it calls suspend_unblock on the input-event-queue suspend_blocker.
> - The user-space input-event thread returns from read. If it determines that
> the key should be ignored, it calls suspend_unblock on the
> process_input_events suspend_blocker and then calls select or poll. The
> system will automatically suspend again, since now no suspend blockers are
> active.
That looks reasonable to me.
Still, I have one issue with adding those ioctls. Namely, wouldn't it be
simpler to treat all events coming from wakeup devices as wakeup events?
With the new ioctls user space can "mark" a queue of events coming out of
a device that's not a wakeup one as a "wakeup source", which doesn't seem to
be correct.
Conversely, with those ioctls user space can effectively turn events coming
out of a wakeup device into non-wakeup ones, which doesn't seem to be correct
either.
So, I think evdev client queue's wakeup source (or wakelock) should stay active
if there are any events in the queue and the underlying device is a wakeup one,
i.e. device_may_wakeup() returns true for it. Then, we won't need the new
ioctls at all and user space will still be able to control which events are to
be regarded as wakeup ones by changing the wakeup settings of the devices that
generate them.
Thanks,
Rafael
2012/2/15 Rafael J. Wysocki <[email protected]>:
...
> Still, I have one issue with adding those ioctls. ?Namely, wouldn't it be
> simpler to treat all events coming from wakeup devices as wakeup events?
>
User-space needs to block suspend between select/poll and read for
this to work, so an explicit call to enable this is useful.
> With the new ioctls user space can "mark" a queue of events coming out of
> a device that's not a wakeup one as a "wakeup source", which doesn't seem to
> be correct.
>
OK, but I don't think this is a big problem.
> Conversely, with those ioctls user space can effectively turn events coming
> out of a wakeup device into non-wakeup ones, which doesn't seem to be correct
> either.
>
I don't agree with this. There can be multiple readers on an input
device. Only the reader that is responsible for acting on the wakeup
event should call this ioctl.
> So, I think evdev client queue's wakeup source (or wakelock) should stay active
> if there are any events in the queue and the underlying device is a wakeup one,
I don't want additional readers of the device to prevent suspend.
> i.e. device_may_wakeup() returns true for it. ?Then, we won't need the new
> ioctls at all and user space will still be able to control which events are to
> be regarded as wakeup ones by changing the wakeup settings of the devices that
> generate them.
>
I don't think this is currently hooked up for most of the devices we
have. If we do add a dynamic wakeup settings I would prefer to have an
ioctl to set which events on a device should be enabled for wakeup (or
just enabled) instead of switching the entire drive. That way we could
turn off unneeded rows and columns in a key matrix.
--
Arve Hj?nnev?g
On Thursday, February 16, 2012, Arve Hj?nnev?g wrote:
> 2012/2/15 Rafael J. Wysocki <[email protected]>:
> ...
> > Still, I have one issue with adding those ioctls. Namely, wouldn't it be
> > simpler to treat all events coming from wakeup devices as wakeup events?
> >
>
> User-space needs to block suspend between select/poll and read for
> this to work, so an explicit call to enable this is useful.
>
> > With the new ioctls user space can "mark" a queue of events coming out of
> > a device that's not a wakeup one as a "wakeup source", which doesn't seem to
> > be correct.
> >
>
> OK, but I don't think this is a big problem.
>
> > Conversely, with those ioctls user space can effectively turn events coming
> > out of a wakeup device into non-wakeup ones, which doesn't seem to be correct
> > either.
> >
>
> I don't agree with this. There can be multiple readers on an input
> device. Only the reader that is responsible for acting on the wakeup
> event should call this ioctl.
Ah, OK. So you envision the new ioctls as the way of saying "I'm the one
who's going to take care of those wakeup events"? If that's the case, may
I suggest changing the names?
> > So, I think evdev client queue's wakeup source (or wakelock) should stay active
> > if there are any events in the queue and the underlying device is a wakeup one,
>
> I don't want additional readers of the device to prevent suspend.
OK
> > i.e. device_may_wakeup() returns true for it. Then, we won't need the new
> > ioctls at all and user space will still be able to control which events are to
> > be regarded as wakeup ones by changing the wakeup settings of the devices that
> > generate them.
> >
>
> I don't think this is currently hooked up for most of the devices we
> have. If we do add a dynamic wakeup settings I would prefer to have an
> ioctl to set which events on a device should be enabled for wakeup (or
> just enabled) instead of switching the entire drive. That way we could
> turn off unneeded rows and columns in a key matrix.
Can you actually select what keys may wake up the system from sleep
(I mean "real sleep", when the whole system is entirely off except for wakeup
devices)?
Rafael
On Thu, Feb 16, 2012 at 11:31:06PM +0100, Rafael J. Wysocki wrote:
> Can you actually select what keys may wake up the system from sleep
> (I mean "real sleep", when the whole system is entirely off except for wakeup
> devices)?
Yes, that's often possible. In suspend mode you can get a two stage
process where the wakeup itself is triggered by simply looking for any
electrical activity on a set of pins. This generates a wakeup and an
interrupt for the keypad which then powers up the actual keypad circuit
and looks to see what the keypress was. The Samsung keypad driver in
mainline does this in runtime PM, while no key is pressed the keypad
controller is in suspend mode.