Fix for bug 2382830: new incoming SMS should not be rejected when
running low on internal phone storage.
Testing revealed that the /data partition should have at least 256 KiB
available in order to prevent random app crashes (including system apps)
due to SQLite transaction failures. With 256 KiB free, the device should
safely boot without storage full errors. This takes into account the
36-40 KiB that the YAFFS2 filesystem reports as available even after
the partition has been completely filled. I've set the default full
threshold to 1 MiB to provide a generous safety margin.
For this bug, I changed the DeviceStorageMonitorService demon to send
two new hidden notifications for device storage "full" and "not full",
when the free space falls below the full threshold (default 1 MiB,
but configurable as a system setting), in addition to the existing
storage low/okay notifications sent when the storage crosses the threshold
of 90% full (also configurable).
The SMS code was changed to use these new notifications so that it can
accept messages until the data partition has been filled to the maximum
safe capacity rather than stopping when it hits 90% full. There should
be no negative impact on battery life because the additional check in
the storage polling service should be offset by an optimization to cache
the free threshold values which were previously being computed every time
through the loop.
While testing this change, I discovered that SMSDispatcher was being
instantiated twice, the first time in GSMPhone/CDMAPhone, and the second
time in SimSmsInterfaceManager / RuimSmsInterfaceManager. Changed the code
to pass the original SMSDispatcher to the Sim/RuimSmsInterfaceManager
constructor.
Change-Id: Ie0c6d05294778ab6ee42e0fa01313af96d824c77
Added the MotionEvent.FLAG_WINDOW_IS_OBSCURED flag which is set by the
input manager whenever another visible window is partly or wholly obscured
the target of a touch event so that applications can filter touches
accordingly.
Added a "filterTouchesWhenObscured" attribute to View which can be used to
enable filtering of touches when the view's window is obscured.
Change-Id: I936d9c85013fd2d77fb296a600528d30a29027d2
Up the blocking queue maximum to twice the number of containers we
allow which should let "asec list" complete without blocking the thread
several times.
Change the threads for the NativeDaemonConnector to vold and netd to be
more descriptive instead of the same name.
Bug: 2501075
Change-Id: I7e7949845a6c70a3d619aec0fa79ceef3c17a4be
The buffer was being compacted in the read loop, but the start offset on
the compacted buffer wasn't being considered after the next read.
Bug: 2501075
Change-Id: I163297d751cf800d0bbc66df66b1a0fa0785de49
The *pReplyData argument of the command() function was left unitialized by EffectHandle::command()
when command was EFFECT_CMD_ENABLE, EFFECT_CMD_DISABLE and EFFECT_CMD_SET_PARAM_COMMIT.
Change-Id: I91a19817ead2a8cfbdd8e2d77ca270c7ce9d5bd4
A logic error would allow not deleting data when uninstalling a system
app upgrade that's of a newer release than the on on the system
partition. If the system app had a database upgrade and this happened,
you might be in trouble.
Luckily it appears no one ever does this in code. You'd have to manually
get there doing:
adb uninstall -k updated.system.package
Change-Id: I3110fedf2d147975a0635c71898d985dd642a2c1
A few lines down "ps" is run through a null check, so apparently there
could be some screwy system conditions where this is encountered. And,
in fact, this was encountered through an updated system pacakge that
managed to delete the package from /system/app during testing. It left
the PackageSetting in mDisabledSysPackages which would cause a deference
of a null variable.
Change-Id: Ic7faca569abd16a87a8a69c20df08c11a274a001
- Collect data at better times.
- Collect per-thread CPU usage as soon as possible after the ANR, and print
in log.
- Based on new per-thread CPU usage, limit the number of processes we
collect stacks from to not include inactive not interesting procs.
- Improve the way ProcessStats compute and reports its data.
Change-Id: I12b17fb47d593d175be69bb792c1f57179bf4fdf
- Fixed click when re-enabling effect during the turn off phase:
make sure the effect states where effect is processed are the same
where volume control is delegated to effect.
- Fixed click when effect is deleted while still active: do not apply
volume ramp if an effect having volume control was just removed from the
effect chain.
Also fixed a crash when PCM dump is enabled in effect bundle wrapper.
Change-Id: Ib562f5cf75c69af75df0e862536262e2514493e4
Finished the input device capability API.
Added a mechanism for calibrating touch devices to obtain more
accurate information about the touch contact area.
Improved pointer location to show new coordinates and capabilities.
Optimized pointer location display and formatting to avoid allocating large
numbers of temporary objects. The GC churn was causing the application to
stutter very badly when more than a couple of fingers were down).
Added more diagnostics.
Change-Id: Ie25380278ed6f16c5b04cd9df848015850383498
Also removes the artifical restriction that only one apply() can be in
flight at once. That was old from when I thought it'd end up being
required, but wasn't.
Change-Id: I3540ea8be6e0760d6a51d218186f71655c2f3f55
* Add native library tests during package move
* Fix omission that let PackageSettings and ApplicationInfo get out of
sync.
Change-Id: Ic4958f971cb2c9cfd270522fdc120b4031c1124a
The problem is that the audio policy manager does not handle the input devices
when forced use for telephony is changed.
The problem does not appear in a call over PSTN becasue only teh output devices drives the
routing of in call audio to/from the base band.
The fix consists in modifying AudioPolicyManagerBase::setForceUse() to check for active inputs
and update the input device if needed.
Change-Id: I0d36d1f5eef1cce527929180c29b025439902f10
This change moves the native library handling earlier in the package
installation process so that it may be inserted into ASEC containers
before they are finalized in the DefaultContainerService.
Note that native libraries on SD card requires that vold mount ASEC
containers without the "noexec" flag on the mount point.
Change-Id: Ib34b1886bf6f94b99bb7b3781db6e9b5a58807ba
See e.g. http://code.google.com/p/android/issues/detail?id=10042
This is also needed by the power control widget, which has no reliable
way otherwise of staying in-sync.
Change-Id: I8f2b6b79b1843329bae952a25ea56f15e3cf92aa
Adds a fire-and-forget save method (startCommit) to the
SharedPreferences.Editor, which is the way most people use it anyway.
This commit adds the implementation. The previous commit added the
interface and docs:
previous change: Idf9934b445da1fb72b79f0192218b47c0a7f5a34
git commit: edf32d01316bd3432c023f17747461b08ae36375
In addition, this change:
-- adds a generic "runPendingWorkFinishers" mechanism to
ActivityThread to wait on async operations that are still
in flight and use it for this.
-- ties runPendingWorkFinishers into Activity.onPause,
BroadcastReceiver, and Service.
-- makes sSharedPreferences keyed on name, not File, to avoid
unnnecessary allocations
-- documents and guarantees what thread
OnSharedPreferenceChangeListener callbacks run on
-- makes a few things in frameworks/base use startCommit(), notably
Preference.java (which was ignoring the return value anyway)
Change-Id: I1c8db60ad45643226fe6d246d3e513eeb7bd0ebd
Merge commit '797e688b3bebbf1f50a8ad08528740b064f79153' into gingerbread
* commit '797e688b3bebbf1f50a8ad08528740b064f79153':
Continue trying to make the turn off quickly after call feature work.
The previous implementation worked if the set timeout is short enough, which mine was. This one
carries the remaining amount of timeout override forward through all of the countdown steps.
Bug: 2844990
Change-Id: I040df22f9f9ddf98c355ac6845b7624e95f84f33
There was a flaw in the service management, when the same activity
is doing a bindService() for the same service IBinder. In this case
the activity would correctly keep a list of all generated connections,
however some other data structures would assume there is only one
connection per IBinder, and thus only remember the last.
When that last connection was unbound, the service would be destroyed
since it thought there were no more connections. Then when the
activity was finished, it would try to destroy the service again and
end up with an ANR because the service was already gone and would
not respond.
Change-Id: I59bde38bc24e78147b90b0a7cd525c2a1d20489f
