Age | Commit message (Collapse) | Author |
|
|
|
Add startBugreport() to Dumpstate interface. This is part of
proper binderization of Dumpstate.
The function is not implemented yet.
BUG: 111441001
Test: m -j dumpstate
Test: adb shell /data/nativetest64/dumpstate_test/dumpstate_test
Change-Id: I484215067e5653d2e2f4e34bd9a9b569d426f411
|
|
|
|
When a binder is local, process death means that the client calling
linkToDeath will also die. In this case, there is no need to use a
death recipient, so an error is returned.
Bug: 112664205
Test: atest android.binder.cts
Change-Id: I675637e937c1ac7a584a2d97bb1238e5b9f3e030
|
|
|
|
|
|
- require instantiation of SharedRefBase
- try to detect race in SpAIBinder set
Bug: 112664205
Test: atest android.binder.cts
Test: manually cause crash by instantiating and deleting SharedRefbase
Change-Id: I0f5e23de5bd620274801ea7c7c1b9765f05c6de2
|
|
|
|
|
|
|
|
Add a move constructor to ParcelFileDescriptor. This is required to be
able to something like
vector<ParcelFileDescriptor> fds;
fds.push_back(ParcelFileDescriptor(std::move(unique_fd(open("some_file")))));
Bug: 115607973
Test: system/tools/aidl/runtests.sh
Change-Id: If8186500b528984c3718310c37694a3936905dfb
|
|
The comments in this file were not consistent with the rest of the
module.
Bug: 112664205
Test: N/A
Change-Id: Ie95a8d34f297628c49bcc6810f432ade61dcd62a
|
|
|
|
|
|
(Manual cherrypick to aosp)
Some of the values passed to HWC2::Display by FakeHwcDisplayInjector as
used by the test were passed by reference, with the reference being
owned by the injector.
However, the lifetime of the injector was effectively temporary, and so
the constructed Display instance ended up holding a reference to a dead
instance.
This patch uses statically constructed default values for those
parameters instead, so that the values will not be destroyed. The test
code can also still override the values used if non-default values are
ever desired.
It also adds a comment warning of the possibility so it hopefully does
not happen again.
Bug: 111454533
Bug: 117451794
Test: atest libsurfaceflinger_unittest # On aosp_cf_x86_phone-userdebug
Change-Id: I8dfe6e31111a558a7892f0eef91bd96b8dd71f59
Merged-In: I8dfe6e31111a558a7892f0eef91bd96b8dd71f59
|
|
The plan is to replace this type with std::unique_resource, but for
the time being, it should have a more appropriate name.
Bug: 112664205
Test: atest android.binder.cts
Change-Id: I8d18a3dc4bd76f1b383459fc778d647756d868b1
|
|
|
|
Both to make it easier to change the underlying API in the future and
to make it look nicer.
Bug: 112664205
Test: android.binder.cts
Change-Id: I61ae78949055a4dde655dd54f5dad86ea354e6b4
|
|
|
|
Avoids object allocation that would occur in local case if someone
called 'AIBinder_isRemote(iface->asBinder())'.
Bug: 112664205
Test: atest android.binder.cts
Change-Id: I45cb02bf626a94e633d8e26a14714f60622d7bd6
|
|
|
|
|
|
Bug: 111454530
Test: runng this change via forrest
Change-Id: I7c1b2ae81feba7d9798cf7dcb999499b5a2cb27f
Merged-In: I7c1b2ae81feba7d9798cf7dcb999499b5a2cb27f
(cherry picked from commit 2e702c31452b27b757632544d0c479beee81b21b)
|
|
C++ wrapper namespace 'android' -> 'ndk'
AutoAIBinder -> SpAIBinder (to reflect usage, disambiguate with auto_ptr)
AutoA* -> ScopedA (to reflect usage, disambiguate with auto_ptr
Bug: 112664205
Test: atest android.binder.cts
Change-Id: I84b22572cd9ffcd4e6107337b83290f35d3ad4a7
|
|
* changes:
libbinder_ndk: Add C++ wrapper code.
libbinder_ndk: add read/write string methods.
|
|
When EX_TRANSACTION_FAILED is set, mError will never be okay.
Bug: 116618418
Test: atest android.binder.cts
Change-Id: If44ba1fa82cce836f4ee4faad117a1e6b24bf8de
|
|
This library wraps libbinder_ndk with types similar to those found in
libbinder. This makes implementations more familiar and reduces the
amount of auto-generated code that has to be repeated in AIDL.
In order to simplify the distribution of the this code, it is
header-only in the NDK headers. This allows NDK users to use AIDL
against the NDK with no additional dependencies.
Bug: 112664205
Test: atest android.binder.cts
Change-Id: I6e6ef646226d474d538639f740e340c4e37fa7f7
|
|
utf16 is left on the wire, but utf8 is exposed in the NDK.
Bug: 111445392
Test: atest android.binder.cts
Change-Id: Ie4f19ba52dc2323b1c6cf0c4b5e60c7bf2802cf8
|
|
|
|
Otherwise, they could escape into another process.
Bug: 111445392
Test: atest android.binder.cts
Change-Id: Id7bfb63c20fb8ec4f0ca9ce51abb530a961e8dfe
|
|
|
|
|
|
std::string::c_str returns memory that's freed when the string dies. So,
this method should be returning the string, rather than a pointer to
it's freshly-deleted contents.
Caught by our static analyzer.
Bug: None
Test: Builds
Change-Id: I3dce7b60c6cfb4e3bfbd6c0dd6fe30dfe7f4313c
|
|
Corresponds to ART commit a497a39fe8830fe7b1c1c23b413753fe4e12923b.
Bug: 77856493
Test: installd_otapreopt_test
Test: manual
Change-Id: I0c1b37f78570f306a8bbc50a9a9c54126bb3865e
|
|
|
|
|
|
|
|
Key changes since 1.1.85:
* Added VK_KHR_shader_atomic_int64
* Added VK_KHR_driver_properties
Test: build
Change-Id: I94595b642be6daf8325b20763fe20c0115f6d47e
|
|
is being served."""
|
|
|
|
Bug: 110490179
Fix: 110490179
Test: Manual test dumpsys activity
adb shell dumpsys activity service all-platform-non-critical
adb shell dumpsys activity service all-platform
adb shell dumpsys activity service
Take bugreport and check contents
Test: mmm -j56 frameworks/native/cmds/dumpstate && \
mmm -j56 frameworks/native/cmds/dumpsys && adb sync data && \
adb shell /data/nativetest64/dumpstate_test/dumpstate_test && \
adb shell /data/nativetest64/dumpsys_test/dumpsys_test && \
adb shell /data/nativetest64/dumpstate_smoke_test/dumpstate_smoke_test && \
printf "\n\n#### ALL TESTS PASSED ####\n"
Change-Id: I3736823f72ccb42e0b2603575e827a1647d02ef1
Merged-In: I3736823f72ccb42e0b2603575e827a1647d02ef1
(cherry picked from commit e3694e314e14a696c056ab627c8a3b9066a4e87d)
|
|
being served.""
This reverts commit d2e4de2a0c99031fd7227f162517de313fd29262.
Reason for revert: Dependencies which broke tests due to exclusion from
LOCAL_JNI_SHARED_LIBS have been added.
Change-Id: If400a48214274d121e79135ac5d573fcb1a4a4c2
|
|
Earlier commit broke VelocityTracker::getEstimate, which also broke
pointer location. In additional VelocityTracker::getEstimate is an @hide
api that could be potentially accessed by app.
Add native tests for getEstimate (and specifically, for the currently
used LeastSquares lsq2 strategy) to prevent future regressions.
Bug: 114017699
Test: atest libinput_tests
Change-Id: I69d0c40283b9589b8559e698b66d758a6c838f19
|
|
Bug: 116614593
Test: build with WITH_TIDY=1
Change-Id: I47101c362198665194f8b3248751caaa7da32505
|
|
|
|
|
|
|
|
This is so the code is more modular before we can make a proper binder
service.
Also made the reference to options a constant.
BUG: 111441001
Test: mmm -j frameworks/native/cmds/dumpstate/ && \
adb push ${OUT}/data/nativetest64/dumpstate_test* \
/data/nativetest64 && \
adb shell /data/nativetest64/dumpstate_test/dumpstate_test
Test: adb shell /data/nativetest64/dumpstate_smoke_test/dumpstate_smoke_test
Test: adb bugreport
Change-Id: Ic55c35c11d197e1214d3a3b336803b33579ae5fe
|
|
|
|
Do not specify test_per_src = true for input native tests. Therefore,
only a single binary will be generated for each test suite.
Input tests, for historical reasons, used a variable 'test_per_src' in
the makefile, meaning that a new make target will be generated for each
test C++ file. However, since the tests are using gtest, this is not
necessary. This was the case since the early days of transition from .mk
to .bp. At this point, input system is the only gtest suite that uses
this flag, and there's no merit to continue using it or maintaining.
The added benefit is that now, atest will automatically start supporting
the input native tests.
Bug: none
Test: atest libinput_tests inputflinger_tests
Merged-In: Ie762d31f0b5be5f608dd84bd86dffc566e6706bb
Change-Id: Ie762d31f0b5be5f608dd84bd86dffc566e6706bb
(cherry picked from commit b0d7481d767a86ea7969fb704d57dbf072681dce)
|