Age | Commit message (Collapse) | Author |
|
13231349, 13231352, 13231353, 13231355, 13231438, 13231440, 13231357, 13231442, 13231359, 13231031, 13231456, 13230694, 13231724, 13231033, 13230839, 13231035, 13230841, 13230843, 13231362, 13231744, 13231746, 13231459, 13231748, 13231750, 13231752, 13229581, 13229583, 13231590, 13231592, 13231765, 13230401, 13230403, 13231785, 13231787, 13231789, 13231791] into rvc-qpr1-d-release
Change-Id: I879efe5ae91b5b675ccbb733a503ae52278c5e39
|
|
Change-Id: I906ae5bed4b340817a52f7968276dea035f2e65a
Fixes: 169256435
Merged-In: I4d87e90793a88a646aaa1ae5806f118f1ae51e30
(cherry picked from commit 60f3ab275ef3ddf3afcdfdce4eb09b59024fec51)
|
|
13083737, 13083756, 13083777, 13083779, 13083758, 13083548, 13083531, 13083718, 13083781, 13083738, 13083550, 13083740, 13083743, 13083591, 13083592, 13083595, 13083817, 13083819, 13083821, 13083823, 13083825, 13083826, 13083829, 13083551, 13083761, 13083763, 13083765, 13083783, 13083785, 13083833, 13083786, 13083767, 13083856, 13083744, 13083876, 13083721, 13083490, 13083492, 13083273, 13083860, 13083896] into rvc-qpr1-d-release
Change-Id: I1389d662b355f8fbdae2be4178f7082be9ea5dcc
|
|
This extra layer of indirection below ClientCounterCallback fixes a shared pointer ownership issue between LazyServiceRegistrar and ServiceManager. It also allows for implementation changes (like this one) without changing headers and breaking VNDK.
Bug: 170212632
Test: Manual (Went through reproduction steps in bug on cf_x86_phone-userdebug)
Test: atest aidl_lazy_test
Change-Id: I4164a6d44e567c752726953e85aee0e91c6b525e
Merged-In: I4164a6d44e567c752726953e85aee0e91c6b525e
(cherry picked from commit 7c227cc333b85938a1ad0f860655bb83567ca755)
|
|
Since there is no check to see if SensorEventConnection has been
destroyed, the mEventCache pointer can still be used even after it
was freed.
Bug: 168211968
Test: Run test code that attempts to enable a sensor after destroying
the SensorEventConnection, and verify no system_server crash occurs.
Change-Id: Ia9275b7cc574df371cdb2e1b80c6699df193b580
Merged-In: Ia9275b7cc574df371cdb2e1b80c6699df193b580
(cherry picked from commit 3e9afc163256db661b9039120d07501b3a8a7d99)
(cherry picked from commit 930b583557e46f14b6d0a0dc1700eb33f18a7714)
|
|
13045624, 13045638, 13045639, 13045640, 13045641, 13045642, 13045643, 13046336, 13046337, 13046338, 13045644, 13045758, 13045759, 13045761, 13045762, 13045763, 13045609, 13045610, 13045611, 13045612, 13045764, 13046417, 13045645, 13046456, 13045632, 13045646, 13046457, 13046458, 13046459, 13046460, 13046461, 13046420, 13046346, 13046421, 13046347, 13046422, 13046496, 13046349, 13046423, 13046424, 13046497, 13044632, 13046462, 13045613, 13045647, 13045347, 13045614, 13045615, 13046425, 13046357, 13046358, 13046516, 13046517, 13046518, 13045633, 13045634, 13045635, 13045648, 13046359, 13045649, 13046519, 13046520, 13046521, 13046427, 13046428, 13046429, 13046430, 13046431, 13046432, 13046433, 13046435, 13046536, 13046537, 13046538, 13046539, 13046540, 13046541, 13046543, 13045766, 13045767, 13045768, 13046556, 13046557, 13046558, 13046522, 13046523, 13046524, 13046525, 13046526, 13046091, 13046527, 13046528, 13046529, 13046530, 13046531, 13046532, 13046360, 13046361, 13046362, 13046559, 13046560, 13046353, 13046465, 13046544, 13046533, 13045769, 13046363, 13046466, 13046561, 13046562, 13046364, 13045652, 13046534, 13046563] into rvc-qpr1-d-release
Change-Id: I960bcbf12cfe434c0a89bf84db9a11a7255fa250
|
|
Make sure that the layer still exists before trying to set
the frame rate for it.
Bug: 170027345
Change-Id: Ice1edd7c5320f4d1d915643e30430919de258dbc
Test: manual
(cherry picked from commit 162f643dded48b230888139d1e4c8a5e2013a7cd)
|
|
rvc-qpr1-d-release
Change-Id: I96fc3863a8828793d09ba565dd541a158bdfdfda
|
|
12952587, 12952588, 12952589, 12952107, 12952470, 12952641] into rvc-qpr1-release
Change-Id: I8e259c5da4a4241f8e80d5504963fad4a701e243
|
|
Make sure that the layer still exists before trying to set
the frame rate for it.
Bug: 170027345
Change-Id: Ice1edd7c5320f4d1d915643e30430919de258dbc
Test: manual
(cherry picked from commit 162f643dded48b230888139d1e4c8a5e2013a7cd)
|
|
rvc-qpr1-release
Change-Id: Ib970f7c29150ecf2b5c2cb0fb878be05a921efd0
|
|
rvc-qpr1-release
Change-Id: I179742a0bc89c255eb106f4ede002e62f3cf6aa1
|
|
We have a bunch of bug-reports where we are seeing an abort inside
Rect::inset during/after rotation animations. I can't think of any
reason Rect::inset would abort besides integer overflow. I'm also not
sure what exact state the layer was in, because at the time of the
report SurfaceFlinger has already crashed. I also can't reproduce. So
it's a little difficult to figure out how we actually ended up
overflowing here. I guess some math must have gone very wrong during
the rotation animation calculation, though apparently on a non-visible layer
or only for one frame or something, because no-one reports any visible
issues. For lack of anything better to do now, I'm going to sanitize
the function against integer overflow on the theory that SurfaceFlinger
should at least be resistant against invalid input. I also wonder if
something has gone wrong in the SF math, because we may be applying
a non rect preserving transform to the rect...but when I work through
examples I can't see exactly how it would lead to overflow.
Bug: 166217947
Test: Existing tests pass
Change-Id: I7d989d5be77595f88b2c52fa69c3ed29d17d3931
(cherry picked from commit a724230a325aca0e85dfd874e75f9631ac731093)
|
|
|
|
rvc-qpr1-release
Change-Id: If2307a063cc1d43dd4381ea8bf6434463c598dc2
|
|
Reason: the Backup service doesn't support secondary users.
Test: adb shell pm list features | grep backup || echo NOT_FOUND
Test: atest CtsBackupHostTestCases
Bug: 166339795
Bug: 169341308
Change-Id: I9bf370ac7364725c540e0b9c3acf2d597ea6ab5a
|
|
|
|
rvc-qpr1-release
Change-Id: Ib31f4267587a93af0e88679ea3dc6be032735998
|
|
If a defer transaction is applied on a detached layer, the barrier layer
cannot dequeue its buffer since it incorrectly waits for the detached
layer to commit its dependent pending states. When a layer is detached
we remove all remove sync points. This fix ensures that an already
detached layer does not add any pending states or sync points.
Test: atest SurfaceFlinger_test:DetachChildren#DeferredTransactionOnDetachedChildren
Fixes: 150924737
Change-Id: Iefe4248b1c2f4273d07a0324697a82d0fe9c5e52
(cherry picked from commit edbe6d264ccd2a277b87577b931aa22ad7d9a90f)
|
|
rvc-qpr1-release
Change-Id: Ifeca626065ce2d5240a4e8543aa7d1770a146b39
|
|
Introduce a new function to use BINDER_GET_FROZEN_INFO and related
tests.
Bug: 143717177
Test: atest testBinderLib
Change-Id: I8a6c550fadef90b73a36a7931c2a8b2d00a64a04
Merged-In: I8a6c550fadef90b73a36a7931c2a8b2d00a64a04
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/frameworks/native/+/12495227
Change-Id: I86147f77de84f22a5d3823854a46573edc579cf8
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/frameworks/native/+/12495227
Change-Id: I5317dfcfb4f0ca599e2f8ac9fbc243752dc21093
|
|
Bug: 166163010
Signed-off-by: Jenny Ho <hsiufangho@google.com>
Change-Id: Icec3984fe51dee411e8413cd6b44d67a88eaef9c
(cherry picked from commit 05a6e79f3f919e1e223445a98deb6160b064aae6)
|
|
rvc-qpr1-release
Change-Id: I16fe3485403080d9447d74606a3e27ca8ce914a3
|
|
|
|
The BINDER_FREEZE ioctl and related structures have been added to the
driver to notify binder of the frozen state of a process. This patch
introduces low-level user space handling for libbinder.
Bug: 143717177
Test: verified that binder transactions to a frozen app return ERR_FROZEN
Test: verified that transactions are correctly received by non-frozen apps
Co-developed-by: Todd Kjos <tkjos@google.com>
Change-Id: I31fed5ecb040f5ba5b8e27ab6a20c441964f32b4
Merged-In: I31fed5ecb040f5ba5b8e27ab6a20c441964f32b4
|
|
The target SDK version is used to avoid a permission check for apps
targeting an earlier version of Android. VNDK clients don't provide a
valid op package name which causes the target SDK version check to fail.
This CL removes that target SDK version for those clients (or anyone
that fakes their VNDK op package name) and instead *always* performs the
permission check for them.
Bug: 165055606
Test: Run VNDK client and verify error is no longer printed. Also, try
popular step counter and verify it still can get step data.
Change-Id: I71c0dfa95bd4f753f6819ad46886541605a24213
|
|
Bug: http://b/153609531
Test: m libnative_bridge_guest_libRSSupport.native_bridge
Change-Id: I864ad01c5496c51f562eb4ed25d0225221c91064
|
|
rvc-qpr1-release
Change-Id: I39f41c28a375b98306c0dcced933dbb38fe09c93
|
|
am: 8d49c3fbae am: 59ffb3c974 -s ours am: c2e1fb6a74 -s ours am: a61f909ff1 -s ours
am skip reason: Change-Id I1f6909531bc640097f3f48c4a558fd03f2fa62cb with SHA-1 deb5346761 is in history
Original change: https://googleplex-android-review.googlesource.com/c/platform/frameworks/native/+/12371860
Change-Id: Ia220dd4aceb3ab4949fa3535ebc124d53b8d48d3
|
|
am: 8d49c3fbae am: 59ffb3c974 -s ours am: c2e1fb6a74 -s ours
am skip reason: Change-Id I1f6909531bc640097f3f48c4a558fd03f2fa62cb with SHA-1 deb5346761 is in history
Original change: https://googleplex-android-review.googlesource.com/c/platform/frameworks/native/+/12371860
Change-Id: I550ae75832cd7cac9d3ab0d16dfea926a68d7ded
|
|
am: 8d49c3fbae am: 59ffb3c974 -s ours
am skip reason: Change-Id I1f6909531bc640097f3f48c4a558fd03f2fa62cb with SHA-1 deb5346761 is in history
Original change: https://googleplex-android-review.googlesource.com/c/platform/frameworks/native/+/12371860
Change-Id: I97cbb620a2d8839af2c2c618db9abd7842b80103
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/frameworks/native/+/12371860
Change-Id: I3c6546c3358724a46793203737759f765cd03807
|
|
rvc-qpr1-release
Change-Id: I7a12f999564ee0ecc8aba9236aec012e001b09f0
|
|
6674db2de2
Original change: https://googleplex-android-review.googlesource.com/c/platform/frameworks/native/+/12371718
Change-Id: I083a231ec00cc39391552f18b8612f4e1e75a40b
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/frameworks/native/+/12371718
Change-Id: Ie4e2b1886c5d20be313660a3d10238761e887f1d
|
|
rvc-qpr1-release
Change-Id: I1457239f7dc8c8d1d8ad898135d9a4f491bc3e16
|
|
|
|
Currently RenderEngine does not rebind texture memory after it is done
drawing, so buffers may not be freed until the corresponding texture is
deleted. For SurfaceFlinger this is more acceptable since layers tend to
reuse buffers and they own their own texture name, so it's more natural
to implicitly tie memory lifecycle to layer lifecycle since the buffers
won't need to be free until the layer is deleted anyways.
But for mediaserver it's not acceptable, since thumbnail generation is
rare and texture generation is expensive, so if we don't clean up
texture resources properly then mediaserver hangs on to a video frame
until the next time a thumbnail is generated, which is very bad for HDR
frames since it's a multi-megabyte overhead.
This patch will solve this for mediaserver by adding a toggle to
cleanupPostRender. It's likely that SurfaceFlinger can benefit from this
as well since rebinding memory is fast, but more investigation would be
needed to see if that is the case, and this patch is less risky to be
applied to other branches.
Bug: 162202251
Test: librenderengine_test
Test: Launch custom apk with thumbnail generation
Test: adb shell dumpsys meminfo mediaserver
Change-Id: Iec7f65c4911b5cbf2189fa33929bbcfa97bb4c14
Merged-In: Iec7f65c4911b5cbf2189fa33929bbcfa97bb4c14
|
|
Also include other recently added ops that are currently missing.
Bug: 162547999
Test: make
Change-Id: Iea7b09f8efb0873bc521a36e9086dcec61c99d3d
|
|
People directly using libbinder_ndk functions who didn't create a debug
dump function function would fail to initialize that pointer, and
potentially crash. Those who didn't create a shell function were
guaranteed to crash. This wasn't noticed because the C++ wrappers which
are the recommended way to use libbinder_ndk always set these functions.
Bug: 161812320
Test: unit tests
Merged-In: I1f6909531bc640097f3f48c4a558fd03f2fa62cb
Change-Id: I1f6909531bc640097f3f48c4a558fd03f2fa62cb
|
|
People directly using libbinder_ndk functions who didn't create a debug
dump function function would fail to initialize that pointer, and
potentially crash. Those who didn't create a shell function were
guaranteed to crash. This wasn't noticed because the C++ wrappers which
are the recommended way to use libbinder_ndk always set these functions.
Bug: 161812320
Test: unit tests
Merged-In: I1f6909531bc640097f3f48c4a558fd03f2fa62cb
Change-Id: I1f6909531bc640097f3f48c4a558fd03f2fa62cb
|
|
rvc-qpr1-release
Change-Id: I75f4f5df44b51b4942c3d3d9c88ed597db4a060c
|
|
into rvc-qpr-dev
|
|
We want to overwrite an old service with the same name as a new service
that is being added. The old one will be cleaned up as servicemanager
loses the reference to it.
Test: atest servicemanager_test && additional tests outlined in Bug
Bug: 162553096
Merged-In: I8835d9f7505367d83a0f29f1d928a07b3d57e56a
Change-Id: I8835d9f7505367d83a0f29f1d928a07b3d57e56a
|
|
rvc-qpr1-release
Change-Id: I014714831a51f5f99ddd6eaa11986ade42bba01d
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/frameworks/native/+/12329047
Change-Id: Ifca14afce16a6c2d169638780577c011fee46b25
|
|
* AChoreographer receives refresh rates from DisplayManager
already, so there's no need to default-enable them - AChoreographer only
needs to pump an event with the latest refresh rate to wake up looper. This
also ensures that AChoreographer's callbacks are entirely in-sync with
DisplayManager since there's no raciness.
* Instead of re-requesting a config change from SF, instead inject it in
AChoreographer correctly to save on binder.
Bug: 154874011
Bug: 158680912
Bug: 161406626
Test: while [ true ]; do adb shell service call SurfaceFlinger 1035 i32
1; adb shell service call SurfaceFlinger 1035 i32 0; and repeatedly
rotate the home screen with auto-rotate off.
Change-Id: I66abc2e28e60f06987ce3a54be294c94b77524fc
Merged-In: I66abc2e28e60f06987ce3a54be294c94b77524fc
|
|
rvc-qpr1-release
Change-Id: I36aa6e7fee5808a9afc61f5e353c7d0a22ee9422
|