In ag/20886210, the VHAL properties were separated into a new AIDL
interface but the names remained the same. In incremental builds with
pre-existing old versions of AIDL, these generated headers would remain
there and gets picked up due to include directory ordering.
Eliminate this possiblity by performing a cleanup of the generated
primary headers directory.
Change-Id: Ic5ef401b427b2772e8527b2dc9804781a51125be
hidl-gen -L c++-adapter-main doesn't create a proper depfile or have
other dependencies listed, and so the intermediates aren't getting
updated. This only causes compilation problems when an interface is
removed, but may break tests when interfaces are added.
Fixes: 146227852
Test: reproduce error and use cleanspec to fix it
Change-Id: Ia2c5c30355eb2da9abbf32a9a000f41c3b4461b8
The service supports multiple versions and its name should not be
version dependant.
For backward compatibility, keep the target name
android.hardware.audio@2.0-service (32bit only), existing
PRODUCT_PACKAGES do not have to be updated.
New products should depend on android.hardware.audio.service:32.
Additionally migrate to Android.bp, dropping support for
AUDIOSERVER_MULTILIB. Previously the HAL service architecture (32 vs 64)
would be the same as the audio server. This in not wanted as all android
audio HAL are 32 bits.
Bug: 121208203
Test: # Compare following commands before and after patch:
find $ANDROID_PRODUCT_OUT/vendor -name 'android.hardware.audio*'|xargs file
adb shell reboot # check device boots to UI and
# check the HAL restarts after the audioserver is killed
adb shell ps |grep audio; adb shell killall audioserver; adb shell ps
Change-Id: I25f41735175a6687a051ca1e5a7fee670eb8115e
Signed-off-by: Kevin Rocard <krocard@google.com>
the configstore functionality was added in Q. But The configstore will
be deleted. So additional function is deleted.
Bug: 124531214
Test: build & boot
Test: adb shell lshal | grep configstore & check configstore 1.2
Change-Id: Idbb00b972bc082bd75bba94dbff9cee19df48b89
Removing this file was inadvertently left out of commit 2b3f826.
Bug: 113292034
Test: build+flash from tree before 2b3f826, sync past 2b3f826,
apply this change, build+flash again
Change-Id: Ic12e57e2426d2e105e6b846fa8ca076c6100c2ba
Previously we couple wide color gamut display and color management together,
which is because we only do color management when we have wide color gamut
display. However, we would also want display that doesn't support wide color
gamut but is calibrated and is capable of managing sRGB gamut to have color
management. This means we will decouple wide color gamut display from color
management, a device can indicate that it's color managed without having wide
color gamut display and only manage color within sRGB gamut.
BUG: 111505327
Test: BUild, flash and check Natural/Boosted mode.
Change-Id: If1241ce040a6f691609d8f8d72a2d6f3141557cf
This change provides a reference implementation of the configstore HAL
v1.1.
Bug: 69691076
Test: tested on walleye-userdebug
Change-Id: I68ee224bcbda64f6fef91e8a0f95adb32d504aad
Configstore was up-revisioned from 1.0 to 1.1. However no new
APIs were added to 1.1. Therefore configstore in master will be down
revisioned to 1.0.
Bug: 71555815
Test: lshal | grep configsotre
Test: "cat proc/<configstore pid>/status | grep Seccomp " return:
Seccomp: 2
Change-Id: I65d2dc06fbe12d0c0ccc020bbd287e0b14320f2f
The interface lib has been in VNDK-SP because
android.hardware.graphics.mapper@1.0 was using it. However, since the
dependency has gone [1], there is no need keep it in VNDK-SP. The
VNDK-SP set should be kept as small as possible because libs in VNDK-SP
are subject to double-loading.
Unmark the 'support_system_process' property to exclude the lib from
VNDK-SP.
Bug: 69480083
Test: walleye boots to the UI
Change-Id: I8722c1ac15ddf56a627a12a0c649b4d734e5e5cd
OUT is only defined if some functions in envsetup.sh are run, which is
not the case on the build servers. I'm looking at removing the
environment variable in local builds to keep things consistent.
In this case, OUT_DIR was actually expected, not OUT, which is
equivalent to PRODUCT_OUT.
Test: none
Change-Id: I1e5e9f40727104716212d696927d1a32d7a74fab
The HAL interfaces for testing were copied to /system/lib[64]/vndk
directory since they are built as cc_library and thus falsely recognized
as members of VNDK, which isn't true.
Build them as cc_test_library so that they are filtered out. To do so,
.hidl_for_test files are added for the test HALs and update-makefiles.sh
were ran.
Bug: 64776708
Test: no android.hardware.tests.*.so file in /system/lib[64]/vndk on
2017 pixel devices
Change-Id: I5314db6cc08a43bccfc4521d4b80b614a91e7e48
Configstore was up-revisioned from 1.0 to 1.1. However no new
APIs were added to 1.1. Therefore configstore in mr1 will be down
revisioned to 1.0.
Bug: 64999074
Test: lshal | grep configsotre
Test: "cat proc/<configstore pid>/status | grep Seccomp " return:
Seccomp: 2
Change-Id: Ia2614439cd620fbe5fbd427571ab55a15089cc47
This change provides a reference implementation of the configstore HAL
v1.1.
Bug: 37727469
Test: Built sailfish-userdebug and configstore-1.1 works
Change-Id: I75e7fd1da8e90ae48d779a3ba28957c5a93a5529
Note cleanspec required because of b/38205169 (genrule
doesn't delete output files before building, so when
removing some, they get stale).
Test: power vts test compiles
Change-Id: Ie7ef0fc3d943a75fe898238cf861fb7f99f4146b
Fixes: 35813011
Fixes build breakage from interface changing after -Lc++-sources
and -Lc++-headers was added to hidl-gen.
Test: manual
Change-Id: I9ee0a0e7b72f8f2a3253d205943a696c07129a64
We need google shims on the vendor partition because they are providing
an implementation of a vendor defined interface. They were written by
google just as a courtesy/to make the transition easier. They're
basically a set for vendors to assemble their hal implementations
from.
Bug: 34135607
Test: marlin persist.hal.binderization on/off
Change-Id: I3f86c77c147514fe470f08088a66de312ae32d4e
