After the bind to the FUL service is complete, an
onServiceConnected() callback is received. This callback is
asynchronous - bindService() does not block while we are waiting for
the service to finish binding. Therefore, when rapidly turning the
screen on and off, it is possible to call bindService() and then call
unbindService() before the onServiceConnected() callback is received.
When onServiceConnected() is received, startUi() is called. If the
service is no longer bound, a runtime restart occurs when calling
startUi().
Note that onServiceConnected() actually has its work done via a
handler. The delay of calling the handler increases the possibility
of unbindService() being called before trying to call startUi(). But
since this problem still happens without using the handler,
eliminating using the handler would not solve the problem and would
just create the problems that come with performing operations on
different threads since onServiceConnected() is not called on the main
thread.
Also note that a new instance of FaceUnlock is created in
LockPatternKeyguardView with each iteration. So, if we bind/stop/bind
before getting onServiceConnected(), the second bind happens in a new
instance of FaceUnlock and therefore does not lead to a problem when
onServiceConnected() returns as a result of the first bind.
This fixes some occurrences of bug 6409767. However, this fixes the
problem when turned the device on and off rapidly. It seems there
are some reports of bug 6409767 where this is not the case, so I
can't be sure this has any affect on those cases.
This change also cleans up some debugging and modifies other
debugging to try to get just the information that is useful for
tracking down the bug.
Change-Id: Ifa59107b9974acaa8a18b74b5d47e4cf3a794b8e
Ensure that the shown panel view is not currently attached to a parent
before adding it to the panel decor view.
Bug 6430928
Change-Id: Ic64ec4222db4754e64afdf06d7d2b77fb5ef825a
This is part of the change to remove the blink checkbox. Since the
blink checkbox won't be visible anymore, we want to set liveliness to
off instead of checking the current value.
Change-Id: Iaa68cea8ec0a6012eaaaac77cea0f50575b7e660
After an unrecognized face occurs 5 times in a row, we disable FUL
until the user unlocks via the backup lock. This prevents attacks
where someone tries a bunch of different photos, hoping for a good
enough match to the device's owner.
This value was previously set to 15, which is much higher than
necessary. This change sets it to 5. We've been holding off on
this change because it makes our testing more difficult, but we
want this in there for factory ROM this week.
Change-Id: I4e1acc5b1dcc2c0629e0c0fe97a837d6edc44d5d
The biometric unlock initializeView() function is called every time
the lockscreen is recreated. Since this normally happens when the
device turns off, initializeView() was covering the backup lock so the
backup lock is not exposed when the device turns back on. However,
initializeView() is also called when lockscreen is recreated due to an
orientation change.
With this change, the show() call to cover the backup lock has been
moved out of initializeView(), and the backup lock is now only covered
when the screen is turning off, preventing the backup lock from being
covered on an orientation change.
This also includes changes to prevent biometric unlock function calls
from occurring when SIM or Account unlock is in use. In fact, in any
situation where we know FUL won't be used, we don't even construct it.
This is not only more efficient, but it also cuts down on the
possibility of errors where FUL is being used when it shouldn't be.
Change-Id: Ie97761840df8de5701703d9b9b991726fb601064
The window manager now performs the crop internally, evaluating
it every animation from, to be able to update it along with
the surface position.
Change-Id: I960a2161b9defb6fba4840fa35aee4e411c39b32
This refactoring sets the stage for a follow-on change that
will make use additional functions of the power HAL.
Moved functionality from android.os.Power into PowerManagerService.
None of these functions make sense being called outside of the
system server. Moving them to the PowerManagerService makes it
easier to ensure that the power HAL is initialized exactly once.
Similarly, moved ShutdownThread out of the policy package and into
the services package where it can tie into the PowerManagerService
as needed.
Bug: 6435382
Change-Id: I958241bb124fb4410d96f5d5eb00ed68d60b29e5
This fixes a rare crash that could happen when the device orientation
changes while the handle is held over a target. LockScreen.cleanUp()
was being called which set mCallback to null which then causes it
to crash in MultiWaveViewMethods.onTrigger().
The solution is to also remove OnTriggerListeners in LockScreen.cleanup().
Change-Id: I487c7c9dbbe40417e685b39f2e79b1c43b30fa00
This change allows more flexibility in target placement in MultiWaveView.
It now supports a new way of specifying chevron drawables that allows
them to be moved in directions corresponding to more than the four basic
directions (right, up, left, down).
Chevron drawables can now be updated in an overlay.
This change also adds a place holder and makes a minor tweak to the layout
on 720dp devices where the navbar buttons weren't centered.
Change-Id: Icd319ec5f276870380e27737c873e78f599ff751
Changed the pokeWakelock() call back to take one argument - the duration to stay awake in ms. This
change was needed in order to poke the wakelock for the duration of the watchdog timeout. This
must be done in the service because the duration of the watchdog timeout is unknown at this point.
Moved pokeWakelock() from start() to handleServiceConnected() to make sure that this poke happens
before the poke in the service. This poke is still needed to account for when devices are rotated.
Change-Id: I19d62df1489514de0588ebb937678358e70ffc95
If a fallback key is generated using a key plus a modifier,
then it's possible we might get a different fallback key
generated if the modifier has changed. PhoneWindowManager
needs to remember which fallback is last generated for a
given key code so that it can apply the same fallback action.
When generating cancellation events, it's important to have
preserved the policyFlags of the original event. Otherwise
we may not dispatch the cancellation properly. For example,
some actions are not performed if the POLICY_FLAG_TRUSTED
is not specified.
Remember the metaState associated with a key event so we can
include it when canceled.
Tell the policy when a fallback is being cancelled so that it
can clean up its state.
After a SEARCH shortcut is invoked, clear the flag indicating
that a shortcut is pending. This is to prevent SEARCH from
getting stuck down in the case where we might forget to send
the up. (Shouldn't happen anymore after the prior fixes.)
Bug: 5616255
Change-Id: I68f0a9679c7af464eaf31c099f2aa50b53fecf1f
There are three functions in FaceUnlock.java that have the requirement
that they are to be called on the UI thread. I added checks to log
an error if they are ever called off of the UI thread.
Change-Id: I581968e8138b7561b7ad75a1ac6945bf218e2bcf
Clean up related to new Vibration policy.
The vibrate behavior is now only derived from:
- presence of a vibrator on the device
- current ringer mode selected: NORMAL, VIBRATE or SILENT
If no vibrator is present the ringer mode can only be NORMAL or SILENT.
The control of ringer mode via volume keys when volume keys control the "master"
stream type (RING on phones, MUSIC on tablets) is as follows:
If a vibrator is present:
VOL- and volume equals 1: NORMAL => VIBRATE
VOL- and volume equals 0 and not continuous press: VIBRATE => SILENT
VOL+ and in SILENT mode: SILENT => VIBRATE
VOL+ and in VIBRATE mode: VIBRATE => NORMAL, volume = 1
If no vibrator is present:
VOL- and volume equals 0 and not continuous press: NORMAL => SILENT
VOL+ and in SILENT mode: SILENT => NORMAL, volume = 0
VIBRATE_ON and VIBRATE_IN_SILENT settings are not stored/retreived any more.
AudioService checks and corrects ringer mode and stream volumes if necessary when reading from
DB at boot time.
Also:
Added dump for stream volumes in AudioService.
Added device names missing in AudioSystem for USB accessory and USB device.
Issue: 6036529
Issue: 6414950
Issue: 6448163
Change-Id: I77fb821ec63e4e566320cac2701b4ac466e86aef
This fixes several layout issues on phones and tablets with MultiWaveView.
It adds a new background scrim to be shown behind the MultiWaveView
on the navigation bar.
MultiWaveView is updated to handle gravity constraints which makes layouts
much more flexible across devices.
Change-Id: I64068f4eaa81ef6c464247049117d53432fcacbd
activities
Bug 6413342
Clear extra attributes in the base theme so that they properly overlay
a ContextThemeWrapper over a fixed-size dialog theme.
Change-Id: Ia0d49f0746ca01516afa1eaea7091e5da60d14ae
Inform user "Unusable SIM card." when SIM is permanently disabled
due to too many wrong PUK inputs.
bug:5548405
Change-Id: Id82c7681f58f09709c630b4bc455facb549938ed
The window manager policy made some incorrect assumptions about the
meaning of the Configuration.keyboard field. We need to be more
careful about distinguishing between built-in and external keyboards.
Most of this change is to move the determination of the parts of
the Configuration related to input devices into the WindowManagerService
leveraging new features of the InputManagerService to good effect.
Then we plumb through the flag that indicates whether a device
is internal or external so that we can be more particular about
how the lid switch effects changes to the Configuration.
Bug: 6424373
Change-Id: I36a1c22ade35e578955465a25940a33f227b9763
Uses a static to keep track whether this is the first pass through the lockscreen
This will be true from when the process is instantiated until the lockscreen
is constructed for the first time, and then false afterwards.
When it is true, we suppress the biometric unlock (Face Unlock).
This avoids the problem of unpredictable and slow startup behavior for Face Unlock
on the first unlock after boot.
Change-Id: I70033af17ba3163d5fe07fc58096f01afaaaafd0
AudioService maintains a stack of registered media button event
receivers.
This change modifies the broadcasters of ACTION_MEDIA_BUTTON intents
let AudioService directly handle the corresponding key event instead
of trapping the intent sent by PhoneWindowManager, KeyguardViewBase
and PhoneFallbackEventHandler.
Because the key event may be sent through a PendingIntent,
AudioService now also implements the OnFinished interface to be
notified when the event was consumed so it can release the wake
lock held if it was held when the key event needed to be sent
(see where PassHeadsetKey was instanciated in PhoneWindowManager).
Change-Id: I2e8614df94af9d54edbf714ef443cc372d21827a
Added a config option to allow the lid switch to turn off the
screen. This is a closer match to what a lid switch should be
doing.
Removed an old feature to bypass keyguard when keyboard is visible
because the way it was plumbed in made bad assumptions about
the meaning of the lid switch. Also, the last product we shipped
that had a physical keyboard turned this config option off.
So away it goes. We can bring it back someday if we really want it.
It's questionable how useful the feature is anyhow, since it only
works when the keyguard is unsecure and when the lid switch is
unlikely to be jostled in the user's pocket.
Fixed a bug where we would tell the power manager that the keyboard
was visible even if the lid switch did not control the keyboard.
This used to cause the power manager to try to set the keyboard
brightness, which doesn't work.
Bug: 6377115
Bug: 6406726
Change-Id: Ic84b71d09563d51c92cd1cf132fa8bdee6509103
This will be used to determine which parts of a window a completely
hidden by system UI elements (status bar, nav bar, system bar) so
that they can be clipped out from rendering.
Change-Id: I2c6c6ac67dbdfeed82d2c089ef806fb483165bd9
If turning the power off while FUL was closing (due to a timeout or a
cancel for example), it would be possible for unbind() to get called
twice due to a race condition. Turning the phone off calls unbind()
from the UI thread, while the other close comes from a binder thread
since it is coming from the service.
PATCH SET 1 attempted to solve the problem by adding a lock, but
having a lock around the bind is a bit scary.
PATCH SET 2 takes a new approach by having all calls coming from a
binder thread to be sent as messages to be handled on the UI thread.
Having all events occur on the UI thread removes the possibility of
race conditions and makes the code stronger by making everything
happen in a deterministic order.
This commit also cleans up the logcat logging a bit. A couple of the
log messages are now printing without DEBUG being set to true. This
is by no means spamming the log and they of course are only logged
when FUL is being used. But it serves to give us some meaningful
information from bug reports that are currently showing nothing. The
statements that are now logged would have made this particular bug
easy to track down.
Change-Id: I25a65c0808d88cb941439e5bf1f989dba8608be4
