`usb-service` and `thermal-service` crashes

[0x1a8510f2]

I woke up today to tens or maybe hundreds of notifications about android.hardware.usb-service and android.hardware.thermal-service crashing.

Their respective crash logs look like this:

type: crash
osVersion: google/husky/husky:14/AP2A.240905.003/2024090400:user/release-keys
uid: 1000 (u:r:hal_usb_impl:s0)
cmdline: /vendor/bin/hw/android.hardware.usb-service
processUptime: 0s

signal: 6 (SIGABRT), code -1 (SI_QUEUE)
threadName: android.hardwar
MTE: enabled

backtrace:
    /apex/com.android.runtime/lib64/bionic/libc.so (abort+168, pc 66ac8)
    /vendor/bin/hw/android.hardware.usb-service (aidl::android::hardware::usb::UsbDataSessionMonitor::UsbDataSessionMonitor(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::function<void ()>)+1416, pc 2af58)
    /vendor/bin/hw/android.hardware.usb-service (aidl::android::hardware::usb::Usb::Usb()+564, pc 174d4)
    /vendor/bin/hw/android.hardware.usb-service (main+60, pc 1339c)
    /apex/com.android.runtime/lib64/bionic/libc.so (__libc_init+120, pc 5f0a8)

type: crash
osVersion: google/husky/husky:14/AP2A.240905.003/2024090400:user/release-keys
uid: 1000 (u:r:hal_thermal_default:s0)
cmdline: /vendor/bin/hw/android.hardware.thermal-service.pixel
processUptime: 1s

abortMessage: ThermalHAL could not be initialized properly.

signal: 6 (SIGABRT), code -1 (SI_QUEUE)
threadName: android.hardwar
MTE: enabled

backtrace:
    /apex/com.android.runtime/lib64/bionic/libc.so (abort+168, pc 66ac8)
    /system/lib64/liblog.so (__android_log_default_aborter+16, pc 6360)
    /vendor/lib64/libbase.so (android::base::LogMessage::~LogMessage()+356, pc 17414)
    /vendor/bin/hw/android.hardware.thermal-service.pixel (aidl::android::hardware::thermal::implementation::ThermalHelperImpl::ThermalHelperImpl(std::__1::function<void (aidl::android::hardware::thermal::Temperature const&)> const&)+3596, pc 224ec)
    /vendor/bin/hw/android.hardware.thermal-service.pixel (aidl::android::hardware::thermal::implementation::Thermal::Thermal()+164, pc 14f64)
    /vendor/bin/hw/android.hardware.thermal-service.pixel (main+112, pc 14510)
    /apex/com.android.runtime/lib64/bionic/libc.so (__libc_init+120, pc 5f0a8)

My phone is typically plugged in overnight but last night it wasn't, so those crashes seem a little out of place.

The phone is a Pixel 8 Pro.

[muhomorr]

Is this a reproducible issue? Capture a full system log from Settings > System > View logs if it is.

[0x1a8510f2]

If it is then I wouldn't know. It only happened once and I don't recall doing anything to cause it. I'll try to get a full system log if I see it again.

[0x1a8510f2]

@muhomorr I've been able to repro the issue. In fact, it seems to occur fairly frequently on boot with the latest release (can't confirm about previous releases). I've got a log but I'm not entirely comfortable attaching it publicly. Is there a way to send the logs privately to you / the devs?

[muhomorr]

@0x1a8510f2 Send the log to one of the following places:

Matrix: @muhomorr:grapheneos.org Discord: muhomorr_ (with underscore at the end) Email: muhomorr@grapheneos.org

[muhomorr]

@0x1a8510f2 The log that you've sent doesn't contain these crashes. Was the log captured immediately after the crashes occurred?

[0x1a8510f2]

I'm pretty sure, yeah. The crashes happened after a reboot and I went to get the log immediately. I'll see if I can repro and get another logcat cause this seems to happen after reboots usually.

[0x1a8510f2]

Sending you another log. This time just the thermal service crashing after reboot. Got the log immediately after unlocking my phone so I hope the needed information is in there.

[muhomorr]

There's no crash in this log either. These crash reports are due to a separate issue: historical crash reports are re-processed by the OS at boot in some cases, which triggers the GrapheneOS crash report UI.

[0x1a8510f2]

I see. Is there any way to get notified as soon as the crash occurs so I can get you the correct log?

[muhomorr]

Capture a log if there are multiple crashes in a row of android.hardware.usb-service and android.hardware.thermal-service.pixel again, as you've stated in the original post.

[0x1a8510f2]

The issue is, I'm not getting any indication of the crashes, except those notifications on reboot. So if that's not when the crash happens then I don't know when it does.

[muhomorr]

I'm not getting any indication of the crashes, except those notifications on reboot

This means that there's no new crashes. One-time crash reports right after boot are due to a crash reporting bug.

[muhomorr]

@0x1a8510f2 Are you still getting these crashes and/or crash reports? 2024100800 release contains changes that should help with these issues.

[0x1a8510f2]

I have not seen this notification for some time now, but I have noticed that my USB OTG isn't working at all at the moment. Specifically, nothing happens when I plug a USB mass storage device or crypto wallet into my phone. The latter at least, used to work fine maybe a month or two ago. I'm wondering if it has anything to do with those USB service crashes.

For the record, my USB exploit protection is set to the following, and I'm plugging in the device with the phone unlocked:

If this is likely to be an unrelated issue though, I'm happy with this issue being closed. I'll re-open if I see the error again.

[muhomorr]

Have you checked Settings > Connected devices > USB?

[0x1a8510f2]

I don't even see any "USB" section under "Connected devices".

That's whether a device is currently connected or not.

Should I open a separate issue about this?

[muhomorr]

@0x1a8510f2 Need to see a full system log captured right after USB storage device is connected.

[0x1a8510f2]

I'm actually going through the RMA process with Google at the moment because I'm also having some hardware issues with my phone. I'll see if the issue persists once I have my phone back and try to get the log for you then if so.