]> Gitweb @ Texas Instruments - Open Source Git Repositories - git.TI.com/gitweb - android/platform-hardware-interfaces.git/log
android/platform-hardware-interfaces.git
5 years agoMerge "Updated HALs to test structs/safe_unions containing interfaces"
Treehugger Robot [Mon, 16 Jul 2018 23:23:55 +0000 (23:23 +0000)]
Merge "Updated HALs to test structs/safe_unions containing interfaces"

5 years agoAllow NOT_SUPPORTED for setLinkCapacityReportingCriteria on GERAN.
Amit Mahajan [Mon, 25 Jun 2018 21:16:37 +0000 (14:16 -0700)]
Allow NOT_SUPPORTED for setLinkCapacityReportingCriteria on GERAN.

Test: run vts -m VtsHalRadioV1_2Target
Bug: 110716988
Change-Id: I9a5f014d498db00d818870cbd73e119f66562ef1
(cherry picked from commit c7896501db413c9a0fa44d2697da4079610075d8)

5 years agoUpdated HALs to test structs/safe_unions containing interfaces
Nirav Atre [Fri, 13 Jul 2018 23:03:31 +0000 (16:03 -0700)]
Updated HALs to test structs/safe_unions containing interfaces

Bug: 111019943
Test: Ran the hidl_test suite (C++ and Java)
Change-Id: I93e8fb9611327e263a02af8fccef1fc878746e02

5 years agoMerge "Fix unnecessary copy initialization warnings" android-p-preview-5
Treehugger Robot [Fri, 13 Jul 2018 20:32:41 +0000 (20:32 +0000)]
Merge "Fix unnecessary copy initialization warnings"

5 years agoMerge "Make safeunion@1.0 test HAL Java-compatible"
Treehugger Robot [Fri, 13 Jul 2018 19:41:21 +0000 (19:41 +0000)]
Merge "Make safeunion@1.0 test HAL Java-compatible"

5 years agoFix unnecessary copy initialization warnings
Chih-Hung Hsieh [Fri, 13 Jul 2018 18:48:16 +0000 (11:48 -0700)]
Fix unnecessary copy initialization warnings

Bug: 30413862
Test: build with WITH_TIDY=1
Change-Id: I5de56c163d26e7522141ca06020321c481eea233

5 years agoMerge "Fix IUsb version number in vts test" android-o-mr1-iot-release-1.0.2
Hsin-Yi Chen [Tue, 10 Jul 2018 02:35:41 +0000 (02:35 +0000)]
Merge "Fix IUsb version number in vts test"

5 years agoMake safeunion@1.0 test HAL Java-compatible
Nirav Atre [Sun, 8 Jul 2018 02:31:21 +0000 (19:31 -0700)]
Make safeunion@1.0 test HAL Java-compatible

This change removes Java-incompatible types (memory, handle) from the
safeunion@1.0 test HAL. Moving forward, the idea is to keep this HAL
Java & C++-friendly, and create another version (likely safeunion@2.0)
to regress on C++-only types; tracked by b/110269925.

Bug: n/a
Test: Ran the existing hidl_test suite
Change-Id: Ic7458c3f5b571e77e5f894a9854a2b201f53fe18

5 years agoMerge "Adding HIDL interface type to safeunion test HAL"
Treehugger Robot [Fri, 6 Jul 2018 20:03:36 +0000 (20:03 +0000)]
Merge "Adding HIDL interface type to safeunion test HAL"

5 years agoFix IUsb version number in vts test
Hsin-Yi Chen [Wed, 4 Jul 2018 02:35:01 +0000 (10:35 +0800)]
Fix IUsb version number in vts test

Bug: 110913709
Test: vts-tradefed run vts -m VtsHalUsbV1_1Target
Change-Id: I9ae6add2fe8ff7138e07797c3a191b6dfe611948

5 years agoMerge "Secure_element: vts: Modify openBasicChannel as per OMAPI standard"
Treehugger Robot [Tue, 3 Jul 2018 14:38:41 +0000 (14:38 +0000)]
Merge "Secure_element: vts: Modify openBasicChannel as per OMAPI standard"

5 years agoAdding HIDL interface type to safeunion test HAL
Nirav Atre [Sun, 1 Jul 2018 20:48:44 +0000 (13:48 -0700)]
Adding HIDL interface type to safeunion test HAL

This change adds a new HIDL safe union, InterfaceTypeSafeUnion
(containing an interface+array) to the safeunion test HAL.

Bug: 79878527
Bug: 110269925
Test: Existing hidl_test suite passes successfully. Added more test-
cases to hidl_test_client.cpp to exercise InterfaceTypeSafeUnion
(included in a separate CL).

Change-Id: I04ffeeeabccc1dc2651aac9371569c36d37f4f87

5 years agoSecure_element: vts: Modify openBasicChannel as per OMAPI standard
Ruchi Kandoi [Mon, 2 Jul 2018 20:20:34 +0000 (13:20 -0700)]
Secure_element: vts: Modify openBasicChannel as per OMAPI standard

CHANNEL_NOT_AVAILABLE should be returned when opening basic channel
is not permitted on the secure element

Test: Test with SIM1 HAL
Bug: 110945797
Change-Id: I2dd4eafaf4c534b49a37690a9baadcc5ac93190c

5 years agoFix missed refactor of serial
sqian [Wed, 27 Jun 2018 19:12:26 +0000 (12:12 -0700)]
Fix missed refactor of serial

Bug: 109822806
Test: run vts
Change-Id: If88384699dbb4c2ddd1c580d2b35be38afd0f8e8
Merged-In: If88384699dbb4c2ddd1c580d2b35be38afd0f8e8
(cherry picked from commit c0e2e2b071df5cdff10193556163e2975f0b682c)

5 years agoFixed VTS failure in no SIM case
Jack Yu [Fri, 22 Jun 2018 20:31:09 +0000 (13:31 -0700)]
Fixed VTS failure in no SIM case

Added RadioError::NONE as a possible return value for data call
setup API. The data call could be setup for emergency purposes
when no SIM inserted.

Test: VTS
Bug: 109767888
Merged-In: I4469c371f999b99d35f4078df000f05ee1f3c84d
Change-Id: I4469c371f999b99d35f4078df000f05ee1f3c84d
(cherry picked from commit 431eb118f84de14c94ec18f887197944ec17e712)

5 years agoMerge "Test HAL for HIDL safe_union construct"
Treehugger Robot [Wed, 27 Jun 2018 00:57:51 +0000 (00:57 +0000)]
Merge "Test HAL for HIDL safe_union construct"

5 years agoTest HAL for HIDL safe_union construct
Nirav Atre [Fri, 8 Jun 2018 23:51:29 +0000 (16:51 -0700)]
Test HAL for HIDL safe_union construct

This change implements a basic HAL and server-side functionality to test
the HIDL implementation of safe unions.

Bug: 79878527
Test: Ran make, new tests in hidl_test/ pass successfully (included in a
separate CL).

Change-Id: Ia420137bc1dc0a188e04176081c7f5418e74449c

5 years agoCall notify for voiceReg and dataReg VTS tests
Eric Schwarzenbach [Wed, 6 Jun 2018 23:09:58 +0000 (16:09 -0700)]
Call notify for voiceReg and dataReg VTS tests

Bug: 80554443
Test: vts
Change-Id: I19f7131f7b65323065f3c283a89668ff761815e3
Merged-In: I19f7131f7b65323065f3c283a89668ff761815e3
(cherry picked from commit 2bb626a2975ef3398adbcb12fe50804cf71efc39)

5 years agoMerge "Add setIndicationFilter response notification in 1.2 VTS" android-p-preview-4
Treehugger Robot [Mon, 25 Jun 2018 22:46:51 +0000 (22:46 +0000)]
Merge "Add setIndicationFilter response notification in 1.2 VTS"

5 years agoAdd setIndicationFilter response notification in 1.2 VTS
sqian [Thu, 7 Jun 2018 05:35:20 +0000 (22:35 -0700)]
Add setIndicationFilter response notification in 1.2 VTS

Test: run vts
Bug: 109822806
Change-Id: I95102a148a26cab80235cc68ad83143a9bb6eb39
Merged-In: I95102a148a26cab80235cc68ad83143a9bb6eb39
(cherry picked from commit 0ff41b94b30a0db0ce4a6a5b29600ced5853a762)

5 years agoAllow REQUEST_NOT_SUPPORTED as an error for startNetworkScan
Amit Mahajan [Fri, 15 Jun 2018 18:16:28 +0000 (11:16 -0700)]
Allow REQUEST_NOT_SUPPORTED as an error for startNetworkScan

We are trying to tighten the APIs. However for this case since the
documentation was not updated, we are allowing NOT_SUPPORTED for
now and will be cleaned up in a later release.

Test: run vts -m VtsHalRadioV1_2Target
Bug: 110118713
Change-Id: Id9dd3d7bac99bed36ceb9c906189f1fea78d5a2c
(cherry picked from commit a7587b5a7fdace0d193ad333e6f5cda8b995570b)

5 years agoAdd REQUEST_NOT_SUPPORTED for changeIccPinForApp and changeIccPin2ForApp
sqian [Fri, 15 Jun 2018 21:24:31 +0000 (14:24 -0700)]
Add REQUEST_NOT_SUPPORTED for changeIccPinForApp and changeIccPin2ForApp

Bug: 110037298
Test: run vts -m VtsHalRadioV1_0Target
Change-Id: Iaba800814c5a6950d086a2f105714d01fdaa14a4
Merged-In: Iaba800814c5a6950d086a2f105714d01fdaa14a4
(cherry picked from commit 7470cbd04dbb90ec4a6481d08db54c6a8c5f2efd)

5 years agoSomething fishy in 1.1
Nathan Harold [Thu, 21 Jun 2018 18:53:55 +0000 (11:53 -0700)]
Something fishy in 1.1

Change-Id: I39efa4355cf37b45f3d1068821ec25bc387c37d7

5 years agoDocumentation improvements.
David Gross [Thu, 7 Jun 2018 23:09:59 +0000 (16:09 -0700)]
Documentation improvements.

Document that every model must have at least one input and one output.

Document graph constraints.

Elaborate a bit on "execution order".

Indicate how operands of different lifetimes are defined (or not).

Fix ANEURALNETWORKS_L2_NORMALIZATION documentation for output shape.

ANEURALNETWORKS_SPACE_TO_DEPTH has inconsistent names for dimensions.

ANEURALNETWORKS_PAD documentation doesn't clearly describe rank or
dimensions of output tensor.

Bug: 66478689
Bug: 67601776
Bug: 69631404
Bug: 79482779
Bug: 79539157
Bug: 79539928

Test: mma at tree root

Change-Id: I6227bea2a486d4b45022d590915e10a124d323d4
Merged-In: I6227bea2a486d4b45022d590915e10a124d323d4

5 years agoDocumentation improvements.
David Gross [Thu, 7 Jun 2018 02:57:27 +0000 (19:57 -0700)]
Documentation improvements.

Fix Model::pools documentation.

Specify that an output cannot overlap with an output/input/constant.

Document circumstances under which a prepared model should be expected to execute correctly.

Bug: 87805381
Bug: 79888238
Bug: 77237715

Test: mma at tree root

Change-Id: Ibf3daacad79f1d7e6127acf3887bd98b7c6bdf4e
Merged-In: Ibf3daacad79f1d7e6127acf3887bd98b7c6bdf4e

5 years agoMerge "Document unspecified dimensions."
David Gross [Fri, 15 Jun 2018 22:46:17 +0000 (22:46 +0000)]
Merge "Document unspecified dimensions."

5 years agoDocument unspecified dimensions.
David Gross [Tue, 5 Jun 2018 19:51:08 +0000 (12:51 -0700)]
Document unspecified dimensions.

Test: mm

Bug: 77234888
Bug: 109687058
Change-Id: Ia8928e494d8c712bc9cfb592d67476fde9753f45
(cherry picked from commit 8939b5df129777b6597f5f973552ae99ab956b35)
Merged-In: Ia8928e494d8c712bc9cfb592d67476fde9753f45

5 years agoMerge "Revert "configstore: include crash_dump seccomp policy""
Treehugger Robot [Fri, 15 Jun 2018 16:55:17 +0000 (16:55 +0000)]
Merge "Revert "configstore: include crash_dump seccomp policy""

5 years agoRevert "configstore: include crash_dump seccomp policy"
Jeffrey Vander Stoep [Fri, 15 Jun 2018 14:21:11 +0000 (14:21 +0000)]
Revert "configstore: include crash_dump seccomp policy"

This reverts commit 1ca0e947a49a8c9624d83dc4947c101ba7fc3eb6.

Reason for revert: This change caused a configstore crash on Hikey

Change-Id: Ied327b331ac084349bf4d10c34f5702b4da7aab0

5 years agoAllow general errors for getImsiForApp().
Amit Mahajan [Tue, 12 Jun 2018 22:23:39 +0000 (15:23 -0700)]
Allow general errors for getImsiForApp().

This is to match it with other 1.0 tests where general errors are
allowed. For newer tests we have decided to not allow these errors,
but a failure for this old test is reported when run with SIM
present.

Test: run vts -m VtsHalRadioV1_0Target
Bug: 109889468
Change-Id: If36083b7832706a50805932e8ba08e4eb397f3fe

5 years agoMerge "configstore: include crash_dump seccomp policy"
Treehugger Robot [Wed, 13 Jun 2018 18:06:36 +0000 (18:06 +0000)]
Merge "configstore: include crash_dump seccomp policy"

5 years agoconfigstore: include crash_dump seccomp policy
Jeff Vander Stoep [Wed, 13 Jun 2018 14:01:58 +0000 (07:01 -0700)]
configstore: include crash_dump seccomp policy

Change-Id: I37ebd0da5747c7c475e9d2d63dc5f9fdc511f8e7
Fixes: 78776142
Test: Crash configstore, verify useful crash info in the logs.

5 years agoMerge "Include libhidl-gen-utils to HAL tests"
Zhuoyao Zhang [Wed, 13 Jun 2018 00:05:50 +0000 (00:05 +0000)]
Merge "Include libhidl-gen-utils to HAL tests"

5 years agoMerge "Do not install assembled framework matrix to device."
Treehugger Robot [Tue, 12 Jun 2018 19:26:14 +0000 (19:26 +0000)]
Merge "Do not install assembled framework matrix to device."

5 years agoDo not install assembled framework matrix to device.
Yifan Hong [Tue, 12 Jun 2018 01:07:53 +0000 (18:07 -0700)]
Do not install assembled framework matrix to device.

Test: `m` dist
Test: make system matrix incompatible, `m` dist fails
Test: `m` and /system/etc/vintf files are installed
Bug: 110056937
Change-Id: If0bf2023f48ef31fffaf8636b3893cbfb40270ee

5 years agoInclude libhidl-gen-utils to HAL tests
Zhuoyao Zhang [Fri, 8 Jun 2018 17:57:30 +0000 (10:57 -0700)]
Include libhidl-gen-utils to HAL tests

Bug: 109880878
Test: make vts
Change-Id: Ic5a639559602688b40296a3de28abdd58b8b7e65

5 years agoMerge "Fix bug in VTS attestation cert verification."
Treehugger Robot [Fri, 8 Jun 2018 13:00:19 +0000 (13:00 +0000)]
Merge "Fix bug in VTS attestation cert verification."

5 years agoMerge "secure_element: vts: Fix openBasicChannel and transmit test case"
Treehugger Robot [Thu, 7 Jun 2018 21:52:27 +0000 (21:52 +0000)]
Merge "secure_element: vts: Fix openBasicChannel and transmit test case"

5 years agosecure_element: vts: Fix openBasicChannel and transmit test case
Ruchi Kandoi [Wed, 6 Jun 2018 22:21:41 +0000 (15:21 -0700)]
secure_element: vts: Fix openBasicChannel and transmit test case

Status Response for basic channel can be only 2 bytes as well.
Update the test to test for 2 or more bytes.
Basic Channel if opened successfully should be closed.

Transmit test case should specify the channel number in the command.

Test: atest VtsHalSecureElementV1_0TargetTest
Bug: 109833790
Change-Id: I91c943b4297067f6ae6872efb038f1295f92f5c8
Merged-In: I91c943b4297067f6ae6872efb038f1295f92f5c8

5 years agoAdd explicit README for configstore future goals.
Steven Moreland [Wed, 6 Jun 2018 20:03:13 +0000 (13:03 -0700)]
Add explicit README for configstore future goals.

Previously, supposedly, each individual client of configstore
was supposed to add the add_hwservice attribute itself to get
ahold of the specific sub-interface of configstore relevant to
it. However, there is only one configstore interface,
ISurfaceFlingerConfigs.

From this point onward, the configstore hal is to be thought of
as specifically relating to surface flinger. Other properties
may be added as other attributes/packages.

For instance, if we want a configstore entry for 'IFooConfig',
then we would add the configuration to one of the following
packages:
- android.hardware.foo@X.Y (to the interface itself)
- android.hardware.foo.config@X.Y (to a configuration of the interface)
- android.hardware.configstore.foo@X.Y (as a sub-interface of configstore)

and then it could be associated with the sepolicy attributes
(respectively):
- hal_foo
- hal_foo_config (or just hal_foo if they are 1-1)
- hal_configstore_foo

The specific pattern to be followed irrelevant to this CL
and subject to future discussion, the point being that we're going
to have a separate sepolicy attribute (and package, although this
isn't strictly necessary) for each separate domain's configuration.

Bug: 109806245
Test: N/A
Change-Id: I837e0bdd11e028265756a905c234e087924c1d6b

5 years agoUpdate comments for doc comments. android-o-mr1-iot-release-1.0.1
Steven Moreland [Thu, 17 May 2018 21:32:17 +0000 (14:32 -0700)]
Update comments for doc comments.

Doc comments look like "/** ... */" and they
can only be in certain places.

Bug: 79865343
Test: m
Change-Id: Ic15c08ff7dc6e4f9827c1dbe7f7236c11a572ec1
Merged-In: Ic15c08ff7dc6e4f9827c1dbe7f7236c11a572ec1

5 years agoMerge "Audio HAL HIDL wrapper: Fix incorrect conversion of TTY Mode"
Treehugger Robot [Thu, 17 May 2018 19:51:35 +0000 (19:51 +0000)]
Merge "Audio HAL HIDL wrapper: Fix incorrect conversion of TTY Mode"

5 years agoUpdate namespace to match framework/ml/nn
Mika Raento [Tue, 17 Apr 2018 15:49:50 +0000 (16:49 +0100)]
Update namespace to match framework/ml/nn

Bug: 72448000
Test: make vts
Change-Id: I3543913561b33e4d1e9d960ec1f81624fb0a1e6b
Merged-In: I3543913561b33e4d1e9d960ec1f81624fb0a1e6b
(cherry picked from commit de16694a6079355867287e4613ae89123c224171)

5 years agoNNAPI: sync NDK and HAL documentation
Michael Butler [Fri, 27 Apr 2018 01:41:24 +0000 (18:41 -0700)]
NNAPI: sync NDK and HAL documentation

Bug: 72650109
Test: mma
Merged-In: I3e0a3680b89d80be500d8975f587f2d7c33fea10
Change-Id: I3e0a3680b89d80be500d8975f587f2d7c33fea10
(cherry picked from commit e98352806b55b5811635d40a06157e5a52e2d89c)

5 years agoUpdate neuralnetworks/*/types.hal to match impl
Mika Raento [Wed, 18 Apr 2018 15:28:40 +0000 (16:28 +0100)]
Update neuralnetworks/*/types.hal to match impl

Updates hardware/interfaces/neuralnetworks/1.(0|1)/types.hal to match
the NeuralNetworks.h header in framework/ml/nn. Only comments have
changed.

Updated using framework/ml/nn/tools/sync_enums_to_hal.py.

Change-Id: I0754868ad8acf6e2e0c5b83661d04682febec9b0
Merged-In: I0754868ad8acf6e2e0c5b83661d04682febec9b0
Bug: 77604249
Test: checked changes with git diff
Test: mm in $ANDROID_BUILD_TOP
(cherry picked from commit 7e64e7f92447d74f53191be0aeeb36089fdebe61)

5 years agoconfigstore: remove references to board platforms
Steven Moreland [Mon, 7 May 2018 22:34:23 +0000 (15:34 -0700)]
configstore: remove references to board platforms

These platforms should set TARGET_USE_CONTEXT_PRIORITY.

Change-Id: I9cd1984437996405c057246f383d6cf5327667e6
Fixes: 74082983
Test: m only (sanity, these boards aren't in tree for me)

5 years agoMerge changes from topic "lazy-dumpstate" android-p-preview-2 android-p-preview-3
android-build-prod (mdb) [Mon, 7 May 2018 18:07:21 +0000 (18:07 +0000)]
Merge changes from topic "lazy-dumpstate"

* changes:
  Make dumpstate HAL lazy.
  dumpstate: add interface into rc file

5 years agoMake dumpstate HAL lazy.
Steven Moreland [Tue, 1 May 2018 23:26:21 +0000 (16:26 -0700)]
Make dumpstate HAL lazy.

Bug: 78509314
Test: HAL isn't running
Test: adb bugreport HAL runs
Test: HAL isn't running after bugreport finishes
Test: manually inspect state using `lshal` and logs
    using `adb logcat | grep android.hardware.dump`

(cherry picked from commit 711b806727c85d91cf805db4f78b4a9609e2bf43)

Change-Id: Ic97516a4297e7933430fb4fa3fd0109f8c1b1d0a

5 years agodumpstate: add interface into rc file
Wei Wang [Thu, 5 Apr 2018 20:32:14 +0000 (13:32 -0700)]
dumpstate: add interface into rc file

Bug: 7748994
Test: Build
Change-Id: I5e86b0e63928f42ea7b247bd0de0719daf80a10d
Merged-In: I5e86b0e63928f42ea7b247bd0de0719daf80a10d

5 years agoAudio HAL HIDL wrapper: Fix incorrect conversion of TTY Mode
Kevin Rocard [Thu, 1 Feb 2018 06:20:16 +0000 (22:20 -0800)]
Audio HAL HIDL wrapper: Fix incorrect conversion of TTY Mode

The TTYMode enum numeric value was converted to a char* instead of its
literal value and vice versa.

Instead convert it to/from its literal value.
Eg: TTYMode::FULL -> "FULL" instead of "3"
    "FULL" -> 0 as atoi conversion would fail

Test: atest VtsHalAudioV4_0TargetTest
Bug: 71486871
Change-Id: I29bbf6bf3b5532269afcc5d39ea10eff2871bdea
Merged-In: I29bbf6bf3b5532269afcc5d39ea10eff2871bdea
Signed-off-by: Kevin Rocard <krocard@google.com>
Signed-off-by: Jungyee Yoo <jungyee.yoo@lge.corp-partner.google.com>
5 years agoFix bug in VTS attestation cert verification.
Shawn Willden [Fri, 4 May 2018 16:08:36 +0000 (10:08 -0600)]
Fix bug in VTS attestation cert verification.

Keymaster VTS is failing to verify that the last certificate in the
chain is self-signed.  CTS and GTS tests verify this, but it should be
validated at this level as well.

Bug: 79123157
Test: VtsHalKeymasterV3_0TargetTest
Change-Id: I5ff33fc8186182c2cf8d43d90cd59f89ce45d416

5 years agorename BUILT_SYSTEM_{COMPATIBILITY_,}MATRIX android-o-mr1-iot-release-1.0.0 android-wear-p-preview-2
Steven Moreland [Tue, 1 May 2018 16:59:32 +0000 (09:59 -0700)]
rename BUILT_SYSTEM_{COMPATIBILITY_,}MATRIX

For consistency with BUILT_VENDOR_MATRIX.

Bug: 66917623
Test: manual
Change-Id: I666eceeb9fc20e5f5754f605c2fbb6530e51b326

5 years agoRemove early checks on manifest.
Steven Moreland [Mon, 30 Apr 2018 23:23:44 +0000 (16:23 -0700)]
Remove early checks on manifest.

These are to be checked in verify_vintf_matrices at
the end of the build when all of the manifest fragments
are available.

Bug: 66917623
Test: check that verify_vintf_matrices hits these checks
Change-Id: I3f184a05a6a95751384da077257dddcf73d3a6f2

5 years agoRestore kernel requirements for frozen matrices.
Yifan Hong [Wed, 25 Apr 2018 21:36:00 +0000 (14:36 -0700)]
Restore kernel requirements for frozen matrices.

Requirements for legacy/1/2.xml has changed because
android-base.cfg files are changed. These files should
build from an older snapshot of the requirements.

Fixes: 78576469
Test: builds
Test: vts_treble_vintf_test

Change-Id: I908ebbe495ae06f983ef983a1be7937cf6be1c12
Merged-In: I908ebbe495ae06f983ef983a1be7937cf6be1c12

5 years agoAdd 4.14.0 kernel to all matrices.
Yifan Hong [Thu, 12 Apr 2018 21:50:09 +0000 (14:50 -0700)]
Add 4.14.0 kernel to all matrices.

P framework can work with 4.14 kernel. 0 should
be updated to the correct value in a follow-up
patch.

This creates the following tags in all matrices:
Kernel 4.14 common requirements
Kernel 4.14 requirements on CONFIG_ARM is set
Kernel 4.14 requirements on CONFIG_ARM64 is set

Bug: 72409164
Bug: 77886167
Test: manually inspect built matrix
Change-Id: I82168df314ad61b21440ec5c703db8358468e083
Merged-In: I82168df314ad61b21440ec5c703db8358468e083

5 years agoMerge "HWC2On1Adapter: Fix fd leak in Layer::dump"
android-build-prod (mdb) [Fri, 27 Apr 2018 18:27:53 +0000 (18:27 +0000)]
Merge "HWC2On1Adapter: Fix fd leak in Layer::dump"

5 years agoMerge "Replace hidl_enum_{iterator -> range}."
android-build-prod (mdb) [Fri, 27 Apr 2018 18:08:06 +0000 (18:08 +0000)]
Merge "Replace hidl_enum_{iterator -> range}."

5 years agoHWC2On1Adapter: Fix fd leak in Layer::dump
Dan Stoza [Fri, 30 Mar 2018 17:31:54 +0000 (10:31 -0700)]
HWC2On1Adapter: Fix fd leak in Layer::dump

Removes the fence dump from Layer::dump, since:
  a) It was leaking (a dup() without a close())
  b) It's not that useful anyway since it wasn't displaying the actual
     fence fd

Test: Manual
Bug: 73979009
Change-Id: I8f7446a05a1bab8c3ca781610ebeb98d17fa483b

5 years agoMerge "graphics: add libhwc2on{1,fb}adapter"
Dan Stoza [Thu, 26 Apr 2018 18:26:31 +0000 (18:26 +0000)]
Merge "graphics: add libhwc2on{1,fb}adapter"

5 years agoReplace hidl_enum_{iterator -> range}.
Steven Moreland [Wed, 25 Apr 2018 19:49:05 +0000 (12:49 -0700)]
Replace hidl_enum_{iterator -> range}.

This class is technically an iterable object, not a range.

Bug: 78573628
Test: N/A
Change-Id: Ibe307cb2e55e030ecd48e7cab09e5863f6a5e2f7

5 years agoMerge "Audio: setParam improve status_t to Result consistency"
Kevin Rocard [Mon, 23 Apr 2018 21:41:14 +0000 (21:41 +0000)]
Merge "Audio: setParam improve status_t to Result consistency"

5 years agoMerge "Mark configstore-utils as double_loadable"
Treehugger Robot [Sat, 21 Apr 2018 02:14:51 +0000 (02:14 +0000)]
Merge "Mark configstore-utils as double_loadable"

5 years agoAudio: setParam improve status_t to Result consistency
Kevin Rocard [Tue, 17 Apr 2018 18:00:58 +0000 (11:00 -0700)]
Audio: setParam improve status_t to Result consistency

The rest of the API (*::analyseStatus) returns NOT_SUPPORTED
when the legacy API returns -ENOSYS.

setParameter legacy -> treble shim did not follow this
conversion due to the legacy API stating that for get_paramers,
-ENOSYS should be returned if
"the implementation does not accept a parameter change while the
 output is active but the parameter is acceptable otherwise",
aka INVALID_STATE.

Thus setParameter shim used to return
 - OK for OK
 - INVALID_STATE for -ENOSYS
 - INVALID_ARGUMENTS for everything else

This leads to several problems:
 - an implementation of the legacy API can not report NOT_SUPPORTED
 - is inconsistent with the rest of the status_t conversion methods
 - shim methods implemented over getParameter can not distinguish
   failures error required by the .hal documentation

Most importantly, on the system side, the Result is transformed to a
status_t again but without any special logic for methods wrapping
getParameter in the shim.
See: analyzeResult in
     frameworks/av/media/libaudiohal/2.0/ConversionHelperHidl.cpp
This can not be changed as the system can not know which methods
are implemented with a legacy wrapper under the Treble API boundary.

This mean that if:
 - hal return -ENOSYS ()
 - shim converts it to INVALID_STATE
 - libaudiohal converts it to NOT_ENOUGH_DATA ()
Thus the checkForNewParameter_l's "status == INVALID_OPERATION"
test in frameworks/av/services/audioflinger/Threads.cpp
are now always false and broken.
This has been broken since the introduction of the Treble shim for O.

Thus setParam now converts status_t to Result in the same way
as all the other shim methods.

Bug: 72873273
Bug: 69811500
Bug: 69010523
Test: playback and record for media and voice call
Change-Id: I41204c0807d2bd4675e941771cbc9a43d7d14855
Merged-In: I41328afce56ce31d4a26159ca2d4b16d14cce05b
Signed-off-by: Kevin Rocard <krocard@google.com>
5 years agoSim Status Present VTS
sqian [Fri, 20 Apr 2018 21:54:12 +0000 (14:54 -0700)]
Sim Status Present VTS

Fix and enhance sim-present tests,
Save VTS running time,
Fix serial number inconsistancy issue

Fix and enhance sim-present tests:
In 1.0:
- setupDataCall timeout, need more waiting time
- requestIccSimAuthentication returns REQUEST_NOT_SUPPORTED,
need to check it
- sendSms timeout, would need more waiting time
- sendSMSExpectMore timeout, would need more waiting time
- getAllowedCarriers, getting CardState::RESTRICTED, the previous test
of setAllowedCarriers is doing resetting back to no carrier restriction,
but that needs some time to populate.
In 1.1:
- setSimCardPower_1_1 set sim card power down that makes other tests
fail, reset it back with sim card power on.

Save VTS running time,
- Use waiting loop and prevent unnecessary waiting to save the whole
running time.

Fix serial number inconsistancy issue
- During the enforcement running, it is liked the serial number is
not consistent. And it happens in b/78249227. Suspect that when sim
card is inserted, during the testing running time, the radio may
request some response that is not triggered from the test, but the
test may receive it and think it is what is triggered by the test. The
fix is to check serial number before notifying of unlock the test
lock.

Bug: 76125134
Bug: 78248071
Bug: 78139665
Bug: 78249227
Test: run vts -m VtsHalRadioV1_0Target; run vts -m VtsHalRadioV1_1Target
Change-Id: I08fe9d7ac16283c7ce1a5aeb6b3b372786a8d5c3
Merged-In: I08fe9d7ac16283c7ce1a5aeb6b3b372786a8d5c3
(cherry picked from commit 788eb8083059ff0dca686734ceebf942042c4893)

5 years agohealth 2.0: fix typo in README
Phong Tran [Thu, 19 Apr 2018 00:20:52 +0000 (07:20 +0700)]
health 2.0: fix typo in README

Test: none

Change-Id: I6c4a1b08ac49d84a77976a488cad53ce14bb9d91

5 years agoMerge changes from topic "nnapi-aosp-sync"
Michael Butler [Wed, 18 Apr 2018 22:00:10 +0000 (22:00 +0000)]
Merge changes from topic "nnapi-aosp-sync"

* changes:
  NNAPI: Add execution preference to prepareModel (HAL)
  Sync NNAPI Operand and Operation documentation fixes
  Fix the spec for TENSOR_QUANT8_ASYMM to match our validation.
  Fix the NNAPI HAL documentation about ADD and MUL
  NN validation tests
  Add validation tests for consistency of model inputs and outputs.
  Disable arm32 asan for VtsHalNeuralnetworksV1_1TargetTest

5 years agoNNAPI: Add execution preference to prepareModel (HAL)
Michael Butler [Wed, 11 Apr 2018 23:30:09 +0000 (16:30 -0700)]
NNAPI: Add execution preference to prepareModel (HAL)

A model can be prepared in different ways to optimize for different
use-cases. This CL propagates the execution preference across the HAL so
that the NN service can better fit the users needs.

Bug: 77864669
Test: mma
Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_1TargetTest
Merged-In: Ib928d510d462f36b6a87d5e81010513db7829fa8
Change-Id: Ib928d510d462f36b6a87d5e81010513db7829fa8
(cherry picked from commit 2504c2fe4f3dc6975165d35a8ea1610d8bc4c477)

5 years agoSync NNAPI Operand and Operation documentation fixes
Michael Butler [Tue, 3 Apr 2018 20:03:50 +0000 (13:03 -0700)]
Sync NNAPI Operand and Operation documentation fixes

This CL adds the typo fixes, bug fixes, and missing descriptions added
to the NNAPI but not to the NN HAL. This CL also adds description of
implicit padding, which was present but not documented in 1.0.

Bug: 77541934
Bug: 75459529
Test: mma
Merged-In: I7baa9e515057c43e759849284b3bb0d420a71f1b
Change-Id: I7baa9e515057c43e759849284b3bb0d420a71f1b
(cherry picked from commit 25f3ad328b1298b8abd3fe1b9006f92e550cf6db)

5 years agoFix the spec for TENSOR_QUANT8_ASYMM to match our validation.
Miao Wang [Thu, 5 Apr 2018 23:27:00 +0000 (16:27 -0700)]
Fix the spec for TENSOR_QUANT8_ASYMM to match our validation.

  - Scale of 0.0 is invalid for quantized tensor.

Bug: 77236592
Test: mma
Merged-In: I3a53d6303d8c964d451e17a3b1671de82d0ff335
Change-Id: I3a53d6303d8c964d451e17a3b1671de82d0ff335
(cherry picked from commit a82d39102af38725f72a415a69330f9304637f96)

5 years agoFix the NNAPI HAL documentation about ADD and MUL
Miao Wang [Fri, 23 Mar 2018 21:50:59 +0000 (14:50 -0700)]
Fix the NNAPI HAL documentation about ADD and MUL

  - ADD and MUL supports QUANT8_ASYMM since OMR1. It was a bug missing
  them in the HAL documentation.
  - Added the updated hash to current.txt for this ABI preserving
  change.

Bug: 75459529
Test: mm
Merged-In: I492a7431c0dbb3dc5967c586d080eb134e380bf2
Change-Id: I492a7431c0dbb3dc5967c586d080eb134e380bf2
(cherry picked from commit f62984027c2cd9041f1b7a9867a142331cb705a9)

5 years agoNN validation tests
Michael Butler [Thu, 22 Mar 2018 23:37:57 +0000 (16:37 -0700)]
NN validation tests

This CL adds validation tests for all of the existing generated models.
The strategy of this CL is this: given a valid model or request, make a
single change to invalidate the model or request, then verify that the
vendor service driver catches the inconsistency and returns
INVALID_ARGUMENT.

Bug: 67828197
Test: mma
Test: VtsHalNeuralnetworksV1_0TargetTest
Test: VtsHalNeuralnetworksV1_1TargetTest
Merged-In: I8efcdbdccc77aaf78992e52c1eac5c940fc81a03
Change-Id: I8efcdbdccc77aaf78992e52c1eac5c940fc81a03
(cherry picked from commit f76acd0312f7d47bd2e371f027a54bca581d8f8f)

5 years agoAdd validation tests for consistency of model inputs and outputs.
David Gross [Fri, 23 Mar 2018 16:51:02 +0000 (09:51 -0700)]
Add validation tests for consistency of model inputs and outputs.

Test: VtsHalNeuralnetworksV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all --gtest_filter=Flavor/NeuralnetworksInputsOutputsTest.*
Test: VtsHalNeuralnetworksV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-float-fast --gtest_filter=Flavor/NeuralnetworksInputsOutputsTest.*
Test: VtsHalNeuralnetworksV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-quant --gtest_filter=Flavor/NeuralnetworksInputsOutputsTest.*

Bug: 67828197
Merged-In: I245227dce095b9cbbb9b527ad99aa71d11f77c4f
Change-Id: I245227dce095b9cbbb9b527ad99aa71d11f77c4f
(cherry picked from commit 7a76d8a7a84d7980ecb37011d109a5a0f7287a70)

6 years agoMark configstore-utils as double_loadable
Jiyong Park [Mon, 9 Apr 2018 03:19:43 +0000 (12:19 +0900)]
Mark configstore-utils as double_loadable

configstore-utils is explicitly marked as double_loadable since it is one of the
(indirect) dependencies of the LLNDK library libvulkan
and at the same time the lib itself is marked as VNDK. Such lib can be
double loaded inside a vendor process.

Note: even without this change, the library is already capable of being
double loaded due to the dependency graph around it. This change is to
make it explicit so that double loading of a library is carefully
tracked and signed-off by the owner of the lib.

Bug: 77155589
Test: m -j
Merged-In: Ibece0476b819e3387836305ee3b1c9dbf614920b
Change-Id: Ibece0476b819e3387836305ee3b1c9dbf614920b
(cherry picked from commit 9fb3c20bc62a3b2977cf2028cfa1c8c4938ee5b2)

6 years agoDisable arm32 asan for VtsHalNeuralnetworksV1_1TargetTest
Pirama Arumuga Nainar [Tue, 6 Mar 2018 19:23:53 +0000 (11:23 -0800)]
Disable arm32 asan for VtsHalNeuralnetworksV1_1TargetTest

Bug: http://b/74200014

Building this module with asan triggers an internal error in gold.

Test: mma SANITIZE_TARGET=address in
hardware/interfaces/neuralnetworks/1.1/vts/functional

Merged-In: I3ce5352a6c86c8dca257d35af06989e57f871749
Change-Id: I3ce5352a6c86c8dca257d35af06989e57f871749
(cherry picked from commit cca38c56f4fcd051df6311ea3ffd56a5e572f910)

6 years agoRadio Sap 1.2 default implementation
sqian [Sat, 7 Apr 2018 03:50:58 +0000 (20:50 -0700)]
Radio Sap 1.2 default implementation

Checked points:
- service is on
- vts can run on it
- provided a dummy implementation that a VTS test can pass it
- applied with recent update in radio 1.2 hal
- format repaired
- pass on a 1.0 VTS test, a 1.1 VTS test, and a 1.2 VTS test

Bug: 74114758
Test: run vts

Change-Id: I8a052e3cedb41db9028552ab88f1e26492718497
Merged-In: I8a052e3cedb41db9028552ab88f1e26492718497
(cherry picked from commit 0de4d31569444d3d4b819bbf3e783ee9f0f17c9d)

6 years agoFix Radio response/indication 1.2 android-o-mr1-iot-preview-8
sqian [Wed, 11 Apr 2018 02:06:48 +0000 (19:06 -0700)]
Fix Radio response/indication 1.2

Radio VTS client 1.2 radio response cannot be cast from radio response
in the service. To fix it, the client radio response and indication
should extend 1.2 IRadioResponse and IRadioIndication

Bug: 77815815
Test: ran it on the default service
Change-Id: Ib014af98d52d9f208d2139f4a239e9d61ea4d569
Merged-In: Ib014af98d52d9f208d2139f4a239e9d61ea4d569
(cherry picked from commit 79bafb943bb53e48351cd159724164c7526976c4)

6 years agoMerge "Add a Test for RadioBandMode to Ensure Auto Mode"
nharold [Fri, 13 Apr 2018 00:21:48 +0000 (00:21 +0000)]
Merge "Add a Test for RadioBandMode to Ensure Auto Mode"

6 years agoFix documentation.
David Gross [Tue, 10 Apr 2018 21:27:33 +0000 (14:27 -0700)]
Fix documentation.

1. Fix documentation for Operand::numberOfConsumers.
2. Fix typos.

Bug: 69055733
Bug: 76016141
Bug: 74111771

Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.0::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all

Merged-In: Ib3d5bbfa701070597731992c3867f72d7b0280bd

Change-Id: I9572b51fa5d3fef01554a5386bd4137215e52599
(cherry picked from commit 541e24311de0561a27bcba7f197f9bd46be67359)

6 years agoAdd a Test for RadioBandMode to Ensure Auto Mode
Nathan Harold [Fri, 6 Apr 2018 23:03:11 +0000 (16:03 -0700)]
Add a Test for RadioBandMode to Ensure Auto Mode

The radio band mode is perilous because depending
on the setting and the detected locale, other
settings may be disabled. That can leave the modem
in a soft-brick state. Thus, BandMode = 0 = AUTOMATIC
should always be supported so that the modem can
update and select a band mode of its choosing.

Bug: 28124606
Test: vts radio - getAvailableBandModes
Merged-In: I1f25d9125ef06d290e3a89b5d2162c6bfe939eba
Change-Id: I1f25d9125ef06d290e3a89b5d2162c6bfe939eba
(cherry picked from commit c9e391801ba5ccd24b41481b26c2861b71ef03d0)

6 years agoMerge "Update RegState for new CellIdentity types"
Treehugger Robot [Tue, 10 Apr 2018 19:32:23 +0000 (19:32 +0000)]
Merge "Update RegState for new CellIdentity types"

6 years agoAdd defaultSysCodePowerState in types.hal in NFC 1.1
Ruchi Kandoi [Tue, 3 Apr 2018 00:52:50 +0000 (17:52 -0700)]
Add defaultSysCodePowerState in types.hal in NFC 1.1

Test: Nfc Enable/Disable; test if config is read in logcat
Bug: 77496460
Merged-In: I13e21ba18c03c3c7b9eff3aa67e843e7f063b233
Change-Id: I13e21ba18c03c3c7b9eff3aa67e843e7f063b233
(cherry picked from commit c20034218208f157564b5450fa981b8caf5efc19)

6 years agoMerge "secure_element: VTS: Remove MANAGE_CHANNEL commands in transmit"
Treehugger Robot [Fri, 30 Mar 2018 23:23:04 +0000 (23:23 +0000)]
Merge "secure_element: VTS: Remove MANAGE_CHANNEL commands in transmit"

6 years agoUpdate RegState for new CellIdentity types
Eric Schwarzenbach [Wed, 14 Mar 2018 23:22:38 +0000 (16:22 -0700)]
Update RegState for new CellIdentity types

Bug: 74826109
Test: make, make vts
Change-Id: I6350e5da5bc064e665e54d161ee589c3eadf901a
Merged-In: I6350e5da5bc064e665e54d161ee589c3eadf901a
(cherry picked from commit 74befb69b9c532e5be77a1a64eb38713b8d51118)

6 years agographics: add libhwc2on{1,fb}adapter
Chia-I Wu [Wed, 27 Dec 2017 21:46:06 +0000 (13:46 -0800)]
graphics: add libhwc2on{1,fb}adapter

Moved from frameworks/native.

Test: build
Change-Id: I03bf1d073306aaa2ce155aee50fb9ac056906323
Merged-In: I03bf1d073306aaa2ce155aee50fb9ac056906323

6 years agoMerge changes from topic "check_unused_hals"
Yifan Hong [Fri, 30 Mar 2018 17:26:14 +0000 (17:26 +0000)]
Merge changes from topic "check_unused_hals"

* changes:
  matrix: ENFORCE_NO_UNUSED_HALS
  matrix: generalize case for LOCAL_WARN_REQUIRED_HALS
  matrix: IGNORE_TARGET_FCM_VERSION -> VINTF_*
  matrix: define LOCAL_MODULE for all modules.

6 years agosecure_element: VTS: Remove MANAGE_CHANNEL commands in transmit
Ruchi Kandoi [Fri, 30 Mar 2018 01:13:05 +0000 (18:13 -0700)]
secure_element: VTS: Remove MANAGE_CHANNEL commands in transmit

Use Dummy APDU and Android CTS AID for the test cases.

Test: Run VtsHalSecureElementV1_0TargetTest
Bug: 77296379
Change-Id: Id5397c9efb000a1737176ebacb6ded0837dc2d61

6 years agoMerge "Add BLOCK_SUSPEND capability for Bluetooth HAL service"
Tom Cherry [Thu, 29 Mar 2018 19:51:20 +0000 (19:51 +0000)]
Merge "Add BLOCK_SUSPEND capability for Bluetooth HAL service"

6 years agomatrix: ENFORCE_NO_UNUSED_HALS
Yifan Hong [Wed, 28 Mar 2018 17:27:10 +0000 (10:27 -0700)]
matrix: ENFORCE_NO_UNUSED_HALS

Enforce that all hals in the device manifest are
in framework compatibility matrix -- at build time.

Bug: 65028233
Test: build a phone without DEVICE_FRAMEWORK_COMPATIBILITY_MATRIX.
Change-Id: Ibf220f80ca3621ddbf8288d372d5f40336e45248

6 years agomatrix: generalize case for LOCAL_WARN_REQUIRED_HALS
Yifan Hong [Thu, 29 Mar 2018 17:06:10 +0000 (10:06 -0700)]
matrix: generalize case for LOCAL_WARN_REQUIRED_HALS

Add LOCAL_ASSEMBLE_VINTF_ENV_VARS_OVERRIDE and
LOCAL_ASSEMBLE_VINTF_ERROR_MESSAGE for a more generic case.

Test: m framework_compatibility_matrix.xml -j
Bug: 65028233
Change-Id: Iec142d161daa18122e7a6daebb82c7230cb98ba6

6 years agomatrix: IGNORE_TARGET_FCM_VERSION -> VINTF_*
Yifan Hong [Thu, 29 Mar 2018 17:57:12 +0000 (10:57 -0700)]
matrix: IGNORE_TARGET_FCM_VERSION -> VINTF_*

All "private" environment variables to assemble_vintf
(inserted by the build rules, not build system variables)
should have this prefix (to distinguish against,
e.g. PRODUCT_ENFORCE_VINTF_MANIFEST)

Test: m framework_compatibility_matrix -j

Bug: 65028233
Change-Id: Iaefa88cefc0918fb86235ea6fcd8e367cbd82788

6 years agomatrix: define LOCAL_MODULE for all modules.
Yifan Hong [Thu, 29 Mar 2018 17:51:48 +0000 (10:51 -0700)]
matrix: define LOCAL_MODULE for all modules.

Test: m framework_compatibility_matrix -j
Bug: 65028233
Change-Id: Iab61771b8ebe0e1b771404bb25a0a04caf399585

6 years agoAdd BLOCK_SUSPEND capability for Bluetooth HAL service
Tom Cherry [Thu, 29 Mar 2018 17:17:07 +0000 (10:17 -0700)]
Add BLOCK_SUSPEND capability for Bluetooth HAL service

This is needed for some platforms

Bug: 74554159
Bug: 77213213
Test: walleye BT works
Change-Id: If0033c5a7fc09057415b27196fa4541e9aadfb4b

6 years agoMerge "Build DEVICE_FRAMEWORK_COMPATIBILITY_MATRIX_FILE"
Yifan Hong [Thu, 29 Mar 2018 00:37:56 +0000 (00:37 +0000)]
Merge "Build DEVICE_FRAMEWORK_COMPATIBILITY_MATRIX_FILE"

6 years agoBuild DEVICE_FRAMEWORK_COMPATIBILITY_MATRIX_FILE
Yifan Hong [Fri, 23 Mar 2018 19:03:33 +0000 (12:03 -0700)]
Build DEVICE_FRAMEWORK_COMPATIBILITY_MATRIX_FILE

... along with compatibility_matrix.device.xml.
The new module cm.device.xml contains the original content
of empty.xml (kernel configs, sepolicy, etc.) as well
as HALs from DEVICE_FRAMEWORK_COMPATIBILITY_MATRIX_FILE.
This variable points to a framework compatibility matrix
that contains framework HALs as a vendor extension; all
HALs in this file must be optional.

Bug: 65028233
Test: m framework_compatibility_matrix.xml -j and manual inspection
Test: vts_treble_vintf_test

Change-Id: I94949f62f7f5b332d845f2e7fa4714df7c49ed3e

6 years agoMerge "Update vts test module bp files to use VtsHalTargetTestDefaults"
Yuexi Ma [Mon, 26 Mar 2018 21:57:41 +0000 (21:57 +0000)]
Merge "Update vts test module bp files to use VtsHalTargetTestDefaults"

6 years agoMerge "Add capabilities to wifi HAL service"
Tom Cherry [Mon, 26 Mar 2018 17:30:28 +0000 (17:30 +0000)]
Merge "Add capabilities to wifi HAL service"

6 years agoAdd capabilities to wifi HAL service
Tom Cherry [Fri, 23 Mar 2018 22:36:09 +0000 (15:36 -0700)]
Add capabilities to wifi HAL service

Bug: 74554159
Test: wifi works on walleye
Merged-In: Ia6784e78a7faec914cbbecbe84800c4fbe46c00b
Change-Id: I8deb730bbd9cb473d68ffc696b31031abf617eec

6 years agoMerge "Fix crash on hotplug disconnect"
Treehugger Robot [Sat, 24 Mar 2018 01:12:08 +0000 (01:12 +0000)]
Merge "Fix crash on hotplug disconnect"

6 years agoFix crash on hotplug disconnect
Lloyd Pique [Fri, 15 Dec 2017 01:59:48 +0000 (17:59 -0800)]
Fix crash on hotplug disconnect

ComposerClient destroys its internal model of the display while handling
the onHotPlug event from the Hwc. Subsequently SurfaceFlinger destroys
its model of the display, and destroys all Hwc layers associated with
the display.

This fixes the code for destroying layers to not dereference an invalid
iterator if the display does not exist, allowing destruction to
continue.

It also fixes a similar issue which could occur if a HWC layer is being
created for a display at around the same time as the disconnect event.

Test: hotplug disconnect no longer crashes
Bug: 38464421
Change-Id: I0f2d28fe89fccf997b4bbb9fa6b5c0e6a6e49b93
Merged-In: I0f2d28fe89fccf997b4bbb9fa6b5c0e6a6e49b93
(cherry picked from commit 2765f9d406aae93adceaec67e8053f2ac1129312)

6 years agoMerge "Add capabilities to bluetooth HAL service"
Tom Cherry [Fri, 23 Mar 2018 19:57:40 +0000 (19:57 +0000)]
Merge "Add capabilities to bluetooth HAL service"