Home UI not fully shown followed by reset on AOSP14 QPR2 Android Automotive

Question

Home UI not fully shown followed by reset on AOSP14 QPR2 Android Automotive

Closed this issue 4 months ago · 13 comments

Hello,

with latest car build I observe the issue, that home UI comes up with black background and some submenus not working/shown.

After a while rpi5 runs a reset and reboots.

Anyone else having that issue?

aosp_log.txt

Answer 1 · 2024-05-07T13:23:22.000Z

I have just built and tested out android-14.0.0_r30 with aosp_rpi5_car-ap1a-userdebug (AAOS) on Raspberry Pi 5.
I am seeing similar behavior: boot animation completes fine, some System UI is shown, parts of the CarLauncher are seen, but incomplete. No apps can be started, but pi reacts fine to things like climate menu.
After a minute or so it reboots on its own.

@Samaelson the logs you got are not logcat. I believe we need logcat logs to see what keeps crashing on Java side and causes the reboot.

Answer 2 · 2024-05-07T14:22:11.000Z

I have tried to connect to Pi via USB-A to USB-A cable, but do not see it under adb devices even when it's "booted".
Looking at other issues, it seems like we need to use UART to get logcat. I only have standard UART adapters with dupont cables, however Pi5 has this weird 3-pin connector, which I have just ordered. So it will take me some time until I can look at it.

In the meantime, @KonstaT, maybe you have a guess on what is missing? Are we doing something wrong, or was aosp_rpi5_car-ap1a-userdebug version simply never tested with android-14.0.0_r30 and we're the first to try?

Answer 3 · 2024-05-07T15:09:35.000Z

New logs attached:
aosp_log+cat.txt

...
--------- beginning of crash
05-07 10:57:08.825 1037 1154 E AndroidRuntime: FATAL EXCEPTION: CarAudioService
05-07 10:57:08.825 1037 1154 E AndroidRuntime: Process: com.android.car, PID: 1037
05-07 10:57:08.825 1037 1154 E AndroidRuntime: java.lang.NullPointerException: Attempt to invoke virtual method 'int android.util.SparseArray.size()' on a null object reference
05-07 10:57:08.825 1037 1154 E AndroidRuntime: at com.android.car.audio.CarAudioService.audioDevicesAdded(CarAudioService.java:3409)
05-07 10:57:08.825 1037 1154 E AndroidRuntime: at com.android.car.audio.CarAudioDeviceCallback.onAudioDevicesAdded(CarAudioDeviceCallback.java:42)
05-07 10:57:08.825 1037 1154 E AndroidRuntime: at android.media.AudioManager$NativeEventHandlerDelegate$1.handleMessage(AudioManager.java:10114)
05-07 10:57:08.825 1037 1154 E AndroidRuntime: at android.os.Handler.dispatchMessage(Handler.java:107)
05-07 10:57:08.825 1037 1154 E AndroidRuntime: at android.os.Looper.loopOnce(Looper.java:232)
05-07 10:57:08.825 1037 1154 E AndroidRuntime: at android.os.Looper.loop(Looper.java:317)
05-07 10:57:08.825 1037 1154 E AndroidRuntime: at android.os.HandlerThread.run(HandlerThread.java:68)
05-07 10:57:09.195 0 0 I binder : undelivered transaction 13096, process died.
05-07 10:57:08.997 1203 1252 E AndroidRuntime: FATAL EXCEPTION: CarAudioService
05-07 10:57:08.997 1203 1252 E AndroidRuntime: Process: com.android.car, PID: 1203
05-07 10:57:08.997 1203 1252 E AndroidRuntime: java.lang.NullPointerException: Attempt to invoke virtual method 'int android.util.SparseArray.size()' on a null object reference
05-07 10:57:08.997 1203 1252 E AndroidRuntime: at com.android.car.audio.CarAudioService.audioDevicesAdded(CarAudioService.java:3409)
05-07 10:57:08.997 1203 1252 E AndroidRuntime: at com.android.car.audio.CarAudioDeviceCallback.onAudioDevicesAdded(CarAudioDeviceCallback.java:42)
05-07 10:57:08.997 1203 1252 E AndroidRuntime: at android.media.AudioManager$Na[ 80.248258] logd: Skipping entries from slow reader, pid 15152, fd 15, from LogBuffer::Prune()
tiveEventHandlerDelegate$1.handleMessage(AudioManager.java:10114)
05-07 10:57:08.997 1203 1252 E AndroidRuntime: at android.os.Handler.dispatchMessage(Handler.java:107)
05-07 10:57:08.997 1203 1252 E AndroidRuntime: at android.os.Looper.loopOnce(Looper.java:232)
05-07 10:57:08.997 1203 1252 E AndroidRuntime: at android.os.Looper.loop(Looper.java:317)
05-07 10:57:08.997 1203 1252 E AndroidRuntime: at android.os.HandlerThread.run(HandlerThread.java:68)
05-07 10:57:09.547 0 0 I binder : undelivered TRANSACTION_COMPLETE
05-07 10:57:09.552 0 0 I binder : undelivered transaction 14350, process died.
05-07 10:57:09.363 1287 1337 E AndroidRuntime: FATAL EXCEPTION: CarAudioService
05-07 10:57:09.363 1287 1337 E AndroidRuntime: Process: com.android.car, PID: 1287
05-07 10:57:09.363 1287 1337 E AndroidRuntime: java.lang.NullPointerException: Attempt to invoke virtual method 'int android.util.SparseArray.size()' on a null object reference
05-07 10:57:09.363 1287 1337 E AndroidRuntime: at com.android.car.audio.CarAudioService.audioDevicesAdded(CarAudioService.java:3409)
05-07 10:57:09.363 1287 1337 E AndroidRuntime: at com.android.car.audio.CarAudioDeviceCallback.onAudioDevicesAdded(CarAudioDeviceCallback.java:42)
05-07 10:57:09.363 1287 1337 E AndroidRuntime: at android.media.AudioManager$NativeEventHandlerDelegate$1.handleMessage(AudioManager.java:10114)
05-07 10:57:09.363 1287 1337 E AndroidRuntime: at android.os.Handler.dispatchMessage(Handler.java:107)
05-07 10:57:09.363 1287 1337 E AndroidRuntime: at android.os.Looper.loopOnce(Looper.java:232)
05-07 10:57:09.363 1287 1337 E AndroidRuntime: at android.os.Looper.loop(Looper.java:317)
05-07 10:57:09.363 1287 1337 E AndroidRuntime: at android.os.HandlerThread.run(HandlerThread.java:68)
05-07 10:57:09.727 0 0 I binder : undelivered TRANSACTION_COMPLETE
05-07 10:57:09.732 0 0 I binder : undelivered transaction 15336, process died.
05-07 10:57:09.738 0 0 I AidlLazyServ[ 80.434999] logd: Skipping entries from slow reader, pid 15152, fd 15, from LogBuffer::Prune()
iceRegistrar: Process has 0 (of 1 available) client(s) in use after notification apexservice has clients: 0
05-07 10:57:09.751 0 0 I binder : undelivered transaction 15364, process died.
05-07 10:57:09.757 0 0 I AidlLazyServiceRegistrar: Shutdown prevented by forcePersist override flag.
05-07 10:57:09.766 0 0 I binder : undelivered TRANSACTION_COMPLETE
05-07 10:57:09.771 10:57:10.355 1600 1641 E AndroidRuntime: at android.os.Looper.loop(Looper.java:317)
05-07 10:57:10.355 1600 1641 E AndroidRuntime: at android.os.HandlerThread.run(HandlerThread.java:68)
05-07 10:57:10.752 0 0 I servicemanager: Could not find android.hardware.automotive.vehicle.IVehicle/default in the VINTF manifest. No alternative in[ 81.076113] logd: Skipping entries from slow reader, pid 15152, fd 15, from LogBuffer::Prune()
stances declared in VINTF.
05
...

Answer 4 · 2024-05-08T09:18:06.000Z

Maybe try adding
<bool name="audioUseDynamicRouting">false</bool> to device/brcm/rpi5/overlay/CarServiceRpiOverlay/res/values/config.xml? Seems like audio zones are not defined/initialized properly, so my guess would be to disable them completely as a first step.

Answer 5 · 2024-05-08T14:50:18.000Z

I have tried to connect to Pi via USB-A to USB-A cable, but do not see it under adb devices even when it's "booted". Looking at other issues, it seems like we need to use UART to get logcat. I only have standard UART adapters with dupont cables, however Pi5 has this weird 3-pin connector, which I have just ordered. So it will take me some time until I can look at it.

In the meantime, @KonstaT, maybe you have a guess on what is missing? Are we doing something wrong, or was aosp_rpi5_car-ap1a-userdebug version simply never tested with android-14.0.0_r30 and we're the first to try?

You can revert raspberry-vanilla/android_device_brcm_rpi5@71b154a if you want to move the serial console to GPIO14/15. USB-A ports on Raspberry Pi are host mode only so it's not even technically possible to use ADB on those. You can use ADB on the USB-C port if you have outside power source (powered USB-C hub, 5V on the GPIO, etc).

No, I haven't tested all build variants on all tags and all devices. tv and car build targets are just side products of the tablet configuration as what AOSP is offering on these build targets doesn't provide much use outside development purposes.

I would assume Android Automotive is working on android-14.0.0_r22 because of #33. AOSP14 QPR2 caused all kinds of mayhem so android-14.0.0_r22 is branched out in a separate manifest.
https://github.com/raspberry-vanilla/android_local_manifest/tree/android-14.0.0_r22

As for the issue itself, it looks like it's required to implement automotive audio control HAL and occupant audio zones (if using legacy routing policy doesn't solve this).

Answer 6 · 2024-05-08T16:45:27.000Z

@khvMX : Tried your approach with setting audioUseDynamicRouting to false, but unfortunately it's not working. Is it working at yours?

Answer 7 · 2024-05-10T09:16:13.000Z

I have tried it and saw no difference in behavior. Unfortunately, I still could not get to logcat, so I did not proceed with debugging. I would expect that with this change it would trip elsewhere.
I'm building the android-14.0.0_r22 now. Note that you need to use instructions from this specific README.md

Answer 8 · 2024-05-10T13:40:30.000Z

I can confirm that android-14.0.0_r22 is booting fine:

Answer 9 · 2024-05-14T11:21:02.000Z

I can confirm that android-14.0.0_r22 is booting fine:

Thanks for update @khvMX . i tried with android-14.0.0_r34 branch but its not booting.
when i am trying to switch the branch to android-14.0.0_r2( repo init -u https://android.googlesource.com/platform/manifest -b android-14.0.0_r22 --depth=1) i am getting manifest default.xml not avaliable.

here is the complete log of swithching the branch:
ramesh@ramesh: repo init -u https://android.googlesource.com/platform/manifest -b android-14.0.0_r22 --depth=1
repo: reusing existing repo client checkout in /home/manyantan/aosp14
fatal: manifest 'default.xml' not available
fatal: remove-project element specifies non-existent project:
Repo command failed: UpdateManifestError
Unable to sync manifest default.xml

any help woule be appreciated

Thanks

Answer 10 · 2024-05-14T13:27:14.000Z

@ramesh-m99 to avoid all sort of weird unexpected issues, I usually start from scratch in another directory. For instance, I started with ~/aosp/rpi-android-14.0.0_r34/, and after that I created ~/aosp/rpi-android-14.0.0_r22/ where I'd execute repo init -u https://android.googlesource.com/platform/manifest -b android-14.0.0_r22 --depth=1 followed by instructions from this README.
Take care not to run out of space, at least remove ~/aosp/rpi-android-14.0.0_r34/out, or whole ~/aosp/rpi-android-14.0.0_r34/ directory.

Answer 11 · 2024-05-15T11:05:44.000Z

@khvMX : Based on the diff of CarAudioService.java (r22<->r34) may be flag name changed and your proposed

"audioUseDynamicRouting" set to "false" does not fit any more for r34?

Answer 12 · 2024-05-15T11:24:19.000Z

runInLegacyMode obviously depends on mUseDynamicRouting and mUseCoreAudioRouting

runInLegacyMode() was removed in >=r38.... obviously this change was turned back and runInLegacyMode only existed r29 to r37

Answer 13 · 2024-09-07T10:44:14.000Z

This was probably fixed AOSP upstream at some point. Android Automotive builds from both android-14.0.0_r67 and android-15.0.0_r1 branches boot to UI successfully.