It used to be in system server, but this is for a class in boot
classpath so we can move it to AndroidRuntime.cpp and avoid system
server dirtying memory for these methods.
Test: m
Bug: 162110941
Change-Id: Ifda5957aab757fe47c97b1aee0a6a8b7a059ac3d
Previously, ANRs for gesture monitors were reported using the pid of the
owner, and ANRs for all windows were reported using its input channel,
which was tracked by WM.
Now, there can be input windows that are not tracked by WM. We unify the
ANR reporting pipeline so that we first try to report an ANR using the
window's input channel. If the ANR reporting fails because the input
channel was not tracked by WM, we fall back on reporting ANR via the
pid of the window owner.
Bug: 210978621
Test: manual: Create a gesture monitor that sleeps while processing
input events, ensure there is an ANR.
Change-Id: If368ccf4a5cc676f64f3ed9de4f765b0abe63284
Merged-In: If368ccf4a5cc676f64f3ed9de4f765b0abe63284
Reorganize networkstats JNI so it can be included in the connectivity
module.
Bug: 197717846
Test: m
Change-Id: I815ea379f3d9915041c80ac5e743ce62e05f4cf8
Merged-In: I815ea379f3d9915041c80ac5e743ce62e05f4cf8
* changes:
Improve compaction to abort when system changes to awake state
Improve compaction by skipping bad VMAs instead of fully bailing out
Fix for compaction bailing out when MAX_RW_COUNT bytes are sent to compaction
Avoid using nested callback classes, move the Executor to register
function and update the Javadoc to emphasize unregister call is always
expected even when a task is already destoryed.
Bug: b/216628683
Test: atest TaskFpsCallbackTest
Change-Id: I801898d082a6eeb6a6f991e3765795a270943d42
While the screen is off we sometimes issue compactions as they are not
disruptive to the user. However, it is possible to wake up and have
compactions running or scheduled so this may lead to jank. This patch
cancels any pending or running compactions that were scheduled to happen
during this non interactive type when the system becomes awake to avoid
jank.
Test: Manual
Bug: 214654755
Change-Id: Ie9b2cbaa7093d63e77f8666d1e6f050ab5610a1d
Previously when a VMA failed compaction for any reason the system would
stop trying to compact all the rest of the VMAs and bail out. However,
there are multiple reasons that a VMA can fail due to not being
reclaimable with -EINVAL, in such instances we will just skip the VMA
and keep going with the rest of the VMAs and any other error will still
cause compaction to bail out as it would likely be irrecoverable.
Test: Manual. Verified that once a VMA errors with -EINVAL, it continues
with the rest of the VMAs and other errors bail out.
Bug: 205658049
Change-Id: Ifc190e371c4dce0eaa6dbab104aa0b666e06027d
Compaction uses process_madvise which internally makes use of iovec
which imposes certain restrictions on the amount of data that can be
sent, previously we had added a constraint of the total vmas sent within
a single syscall determined by UIO_MAXIOV. However, in this CL we also
added a cap on the amount of bytes sent per syscall up to the maximum
supported MAX_RW_COUNT and split the VMA into chunks which end up
being sent in different syscalls.
Test: Verified ZRAM changes during compaction and added logging to
verify that all VMAs where processed
Bug: 205658049
Change-Id: Ib478397a1199d31a85606b0a38ab1b673b2333fc
- Add additional javadoc for vendor and product ids
- Enumerate full allowed key code set for VirtualKeyEvent
- On cloud devices, CTS fails with 'bad address' for VirtualTouchscreen.
Pass by address rather than by value to solve.
Bug: 216554655
Test: atest CtsHardwareTestCases:android.hardware.input.cts.tests
Change-Id: Ifb233d2b6c4eee1ffdcba218ed17312e67bac6f6
Reorganize networkstats JNI so it can be included in the connectivity
module.
Bug: 197717846
Test: m
Change-Id: I815ea379f3d9915041c80ac5e743ce62e05f4cf8
This HAL was changed to InputProcessor, use it here in the jni code in
order to link the InputManager code.
Bug: 210158587
Test: m libinputflinger
Change-Id: Iff33f089862422081ab8a0ddb06f12019e89d943
getInterfaceVersion() returns -1 if it is unknown.
V1 of the HAL doesn't implement getInterfaceVersion
due to ABI reasons for regular, unversioned, AIDL
interfaces.
Bug: 206670536
Test: on device
Change-Id: I358fe2748b4a8450b68d0954dee0814fe246eb18
CallObjectMethod returns a local reference that's causing memory
leakage. We need to make sure all these local refs are deleted.
This change combines I7462bd9293b214d75df5196ae42b4da1da904e1a
(ag/16752598) and I85d68039a5f3951cb629fd4286b182c93a7293e1
(ag/16765005) from the master branch.
Bug: 214967397
Test: manual
Change-Id: Idb222eab7a498b23a194272543b2dab235eaf002
