commit efa6f355b06675aa4d0879fd279e22c16d5c046c
Author: Mikhail Naganov <mnaganov@google.com>
Date: Wed Aug 10 12:25:13 2016 -0700
MIDI: Use server-side socket in blocking mode for virtual devices
Since virtual MIDI servers may misbehave, blocking mode will throttle
them if clients are not coping with their sending speed.
Bug: 29413812
Change-Id: I9c4a2a7a7ea3ea060c93fedc7d0f033427c557c9
commit 755dfb5f83749d3963c63d98d692307f8271c804
Author: Mikhail Naganov <mnaganov@google.com>
Date: Fri Jul 8 13:26:19 2016 -0700
Protect MIDI framework against client blocks in MidiReceiver.onSend
Make the server-side socket non-blocking when creating MidiOutputPort
for clients. Thus if a client ceases to read from its side of the
socket pair, the server will just fail to write instead of blocking.
One drawback is that the MidiOutputPort on the client can't indicate
that it has become dysfunctional, but it's not possible without
changing the API.
Bug: 29413812
Change-Id: I9dfcbdd214a815cea8fd1365324fd78ca459268a
commit c740b13953761f58233ac651a0b5227733b1bdcc
Author: Mikhail Naganov <mnaganov@google.com>
Date: Fri Jun 17 04:11:25 2016 -0700
UsbMidiDevice: Clean up terminology and fix comments
When working with physical MIDI devices, an *input* stream is used
for reading from *output* port of the device, and vice versa. Thus,
using "input" and "output" without specifying whether it's a stream
or a port is confusing.
Clarify names of counter variables, and fix a couple of comments
that were incorrect due to this confusion. No functional changes.
Change-Id: If561eaca4bade94e9296d2c703c9fcebc91296e2
commit 4269c6417287737624f6165a8bbeb5aa427de9a0
Author: Glenn Kasten <gkasten@google.com>
Date: Thu May 5 18:49:16 2016 -0700
Update MIDI package summary
Bug: 28625060
Change-Id: If552ca8e1a0666d402b5f536699bf3fb09c1e324
commit 862d40b73168bde7d0be5280d997985c18061014
Author: Phil Burk <philburk@google.com>
Date: Tue Apr 19 15:56:24 2016 -0700
MidiDevice: do not open ports on closed device
Fix involves client side mIsDeviceClosed flag.
Bug: 24949216
Change-Id: I666284a787fbb9a710d2372fb424e8e54f6a2825
Signed-off-by: Phil Burk <philburk@google.com>
commit 6f1de358b9f2616e03f4655f01454770915ddd66
Author: Phil Burk <philburk@google.com>
Date: Mon Apr 18 16:05:28 2016 -0700
MidiService: fix resource leak
The proxy object was being used to match when adding or removing objects.
But they are different each time. So now we use an asBinder() object.
Bug: 28153736
Change-Id: I1bccebf1e9464668db757ff08b41902d0cf0e3a7
Signed-off-by: Phil Burk <philburk@google.com>
commit f7386bd535bb8a1d7f8df8f44a1748ab770c991a
Author: Phil Burk <philburk@google.com>
Date: Tue Apr 5 14:19:53 2016 -0700
MidiDevice: fix connectPorts for same Process
If connectPorts() was called for a device in the same process then
the connection would die when the ParcelFileDescriptor was closed.
Bug: 26406775
Change-Id: Id0538452593b4761ac2a93d366ade76d2e35ce73
Signed-off-by: Phil Burk <philburk@google.com>
Change-Id: I4dfc2a2cbaf04bf1a790ae2cb39bf74fb5bb16ac
Let RuntimeInit use an UncaughtExceptionPreHandler to log an exception
rather than relying on UncaughtHandler, which apps can replace. This
makes it easier to diagnose application death, especially during app
compatibility testing for a new version of Android.
Test: Verified manually, with the help of a small sample app (not
checked in), that stacktraces for RuntimeExceptions thrown on main
or background threads are logged even when the app set a default
UncaughtExceptionHandler that swallows the exception with no action.
Note that such an inappropriate UncaughtExceptionHandler will still
cause threads to die without the app being killed, which it should be.
In an exception then happens on the main thread, the app will freeze
until the ANR dialog kicks in after a few seconds. I have manually
verified that this behavior is unchanged from before this CL.
No new integration tests are included because the default system
behavior has not changed.
Bug: 29624607
Change-Id: Ie87377b0bcadc3ba4083a8ab1bedb8f3dd95a4bd
When "handling" an uncaught exception, make an attempt to stop
profiling. In case profiling was active, this will avoid losing
the profiling buffer.
This change is required as a base in order for
https://android-review.googlesource.com/#/c/249721/
to merge cleanly.
(Cherry picked from commit 4c79fea9ef)
Bug: 26291225
Change-Id: I35f352e5f28eafe4702da9eae587c3b65c360b3a
The when the timerfd alarm logic was added to the kernel, an oversight was made
and the interface does not check for the CAP_WAKE_ALARM permissions as required
via other kernel methods to trigger an alarm timer.
In v4.8-rc kernels, the change 2895a5e5b3a ("timerfd: Reject ALARM timerfds
without CAP_WAKE_ALARM") was added by Eric Caruso <ejcaruso@google.com>.
After this change (which may be backported to -stable), the AlarmManager will
fail on the first timerfd_create call, and will not be able to set the time
or handle other necessary functions.
The solution here is to add CAP_WAKE_ALARM to the system_server process.
Change-Id: Ifdb16f3ef42711e553f727165de3922d484b5be4
Signed-off-by: John Stultz <john.stultz@linaro.org>
Bug: 29586513
Also gives BackdropFrameRenderer a direct-destroy
of Choreographer since it's hammering on new Threads
and we don't want to wait for the GC to release
FDs.
Change-Id: Id2ec0af2ee4d5304961c4ab87a104ccb92f35fc2
The PowerProfile in BatteryStatsImpl may not be ready when
resetting stats early in the boot sequence.
Bug:29559031
Change-Id: I51bba762231a08804f1b68505bb1b0523476081d
Serbian in Latin script is only partially localized in Android N. We
should not show it as localized in either Setup Wizard or Settings.
Bug: 29460424
Change-Id: I6829a89c49c1b5e8768b1527daa3aa9508deb491
It turns out that the performance of
InputMethodManager#getCurrentInputMethodSubtype() is regressed from ~1ms
to ~20ms when
- Settings.Secure.SELECTED_INPUT_METHOD_SUBTYPE == -1 and
- The active IME supports many subtypes (~100)
because we try to find a fallback subtype based on the system locales
every time when IMM#getCurrentInputMethodSubtype() is called.
This could be contributing UI janks because spell checker clients
running in the UI thread indirectly depend on that method.
Fortunatelly the critical path is in
InputMethodUtils#getImplicitlyApplicableSubtypesLockedImpl(), which is
basically a state-less method. We can easily and safely cache its
result by using LocaleList and InputMethod as cache keys.
With this CL the performance basically recovers to the Android M level.
Bug: 28889203
Change-Id: I5ed16c7f14cc18052854f4fd6c9bae8550c332ee
Fixes a particularly bad transition in the case where there
is one incoming message to a group.
Change-Id: Ieddece4a496292a69e14cdcd74d94986938d8223
Fixes: 29043489
When app is resized in multi-window mode we preserve the window along
with the decor view. If action mode view was shown before such config
change, its view is detached, but the mPrimaryActionModeView variable
is not cleared. So when action mode view is shown again after that it
thinks that its view is still alive.
This CL adds additional check to inflate new mPrimaryActionModeView
if it is not attached to any window.
Bug: 28971666
Change-Id: Ia8c24d4322be32e8b2c8967301beb3a4d889d501
Also moves discharge coulomb data to BATTERY_DISCHARGE_DATA checkin line.
Bug:29264326
Bug:28743761
Change-Id: I78881e166c11f40b2397e04d2cab8d5200c71eea
For devices that report battery discharge via a coulomb counter,
record how much of the battery was discharged while the screen
was on/off.
Bug:28743761
Change-Id: Ie2d1708864352029ff466c1fed14fc057e19b93b
