Crash in background on iOS 15 EXC_… | Apple Developer Forums

Developer Forums

Hello, Since the first iOS 15 beta and even with all the next public releases, we have a very weird crash each time our app enters background. The crash logs don't help at all unfortunately

Each time we see:

Exception Type:  EXC_CRASH (SIGKILL)
Exception Codes: 0x0000000000000000, 0x0000000000000000
Exception Note:  EXC_CORPSE_NOTIFY
Triggered by Thread:  0

and

Thread 0 name:
Thread 0 Crashed:
0   libsystem_kernel.dylib        	0x00000001b7a08564 mach_msg_trap + 8
1   libsystem_kernel.dylib        	0x00000001b7a08bfc mach_msg + 76 (mach_msg.c:119)
2   CoreFoundation                	0x0000000181089698 __CFRunLoopServiceMachPort + 372 (CFRunLoop.c:2646)
3   CoreFoundation                	0x000000018108d98c __CFRunLoopRun + 1212 (CFRunLoop.c:3000)
4   CoreFoundation                	0x00000001810a13c8 CFRunLoopRunSpecific + 600 (CFRunLoop.c:3268)
5   GraphicsServices              	0x000000019c8b238c GSEventRunModal + 164 (GSEvent.c:2200)
6   UIKitCore                     	0x0000000183a47060 -[UIApplication _run] + 1100 (UIApplication.m:3457)
7   UIKitCore                     	0x00000001837c4b8c UIApplicationMain + 2124 (UIApplication.m:5013)

And sometimes we have:

Kernel Triage:
VM - Compressor failed a blocking pager_get

or

Kernel Triage:
VM - pmap_enter failed with resource shortage
VM - Compressor failed a blocking pager_get

But no other info. We can't reproduce it on simulators or on real devices with non-production release (linked to Xcode). But the crashes are here each time we build a release version.

Any help would be appreciated!

NB: it seems similar to this one but we are not even sure

2021-11-01_00-13-03.9761_+0100-a5388bfdcea65316d79dc2c3a4ac95ad146c2cb3.crash

23k

Posted by

SebC99

Reply

I found a new one with the following reason: Termination Reason: RUNNINGBOARD 3735883980 (equivalent of 0xdead10cc)

—
SebC99

Apple Recommended

Many of you in this thread have opened bug reports in Feedback Assistant — thank you for filing those to help us address this! We have deployed a fix to the backend services that provide crashes for your app into the Xcode Organizer, so if you receive EXC_CRASH (SIGKILL) crash reports from now on through the Organizer, those reports should be attributed with an exception code to explain why the system terminated your app.

If you still see any reports without the exception code, please post the complete crash report to this thread using the attachment feature so we can verify that the code is still missing.

Posted by

edford

Replies

for @edford as requested above.

FB9837256 has 2 SIGKILL crash reports with no Termination Reason in them.

Kind Regards

Posted by

DCS17

Thanks for filing FB9837256! Your crashes appear to be due to the watchdog, so our watchdog article should be your starting point. Some other observations to aid you: One of the crashes happened at app launch, while the other happened 30 minutes after launch (compare the two timestamps at the top). Focus on ensuring your UI load and update paths are quick and also that you're not using Core Data in any unusual way, such as configuring it to load far larger amounts of data for every fetch than is the default. Note the differences in the Binary Images section between the 2 reports — the crash 30 minutes after launch has used some audio features, in case that helps you reproduce the scenario.

—
edford
@edford Thanks for advice. Much appreciated.

—
DCS17

Are there any news on this issue?

Posted by

meowpunk

Xcode Organizer reports a very high number of crashes in our app, iOS 15 only. (there is no such crashes in Xcode Organizer before September 2021).
We use Xcode 13.0.

All crashes have

Exception Type:  EXC_CRASH (SIGKILL)   
Exception Codes: 0x0000000000000000, 0x0000000000000000   
Exception Note:  EXC_CORPSE_NOTIFY   
Triggered by Thread:  0

and have no "Termination Reason" field.

Looks like it is a same issue as described in current thread It looks like both old devices (like iPhone 6s) and new devices (like iPhone 11, 12) are affected.

We have three main type of crashes, all with symptoms described above:

Related to audio with fmod::outputcoreaudio::reset in call stack
Related to shaders compiling
Related to writing to keychain.

Reported a bug FB9843106

Posted by

nikcher

Thank you for opening FB9843106. You are running into the standard watchdog terminations for not being responsive to user input. You can start with our documentation on watchdog terminations.

—
edford
We expected termination reason like 0x8badf00d Is it ok that crash logs have 0x0 in Exception Codes and have no "Termination Reason" field? Also is it ok that such reports come only from iOS 15 devices?

Are iOS 15 changed something in Watchdog timeouts (so Watchdog kills app faster than in iOS 14)?

—
nikcher

Hello, I've reported a similar issue here FB9857467, where there wasn't termination reason, but it could be related to the NotificatonCenter.

Exception Type:  EXC_CRASH (SIGKILL)
Exception Codes: 0x0000000000000000, 0x0000000000000000
Exception Note:  EXC_CORPSE_NOTIFY
Triggered by Thread:  0


Kernel Triage:
VM - Compressor failed a blocking pager_get
VM - Compressor failed a blocking pager_get


Thread 0 name:
Thread 0 Crashed:
0   libsystem_kernel.dylib        	0x00000001b8707ae4 __semwait_signal + 8
1   libsystem_c.dylib             	0x000000018bf19f00 nanosleep + 216 (nanosleep.c:104)
2   Foundation                    	0x0000000182875a18 +[NSThread sleepForTimeInterval:] + 156 (NSThread.m:504)
3   CoreFoundation                	0x0000000181002f88 __CFNOTIFICATIONCENTER_IS_CALLING_OUT_TO_AN_OBSERVER__ + 28 (CFNotificationCenter.c:652)
4   CoreFoundation                	0x000000018109ec3c ___CFXRegistrationPost_block_invoke + 52 (CFNotificationCenter.c:173)
5   CoreFoundation                	0x0000000181071ff4 _CFXRegistrationPost + 456 (CFNotificationCenter.c:199)
6   CoreFoundation                	0x0000000181018df4 _CFXNotificationPost + 716 (CFNotificationCenter.c:1147)
7   Foundation                    	0x0000000182815f68 -[NSNotificationCenter postNotificationName:object:userInfo:] + 96 (NSNotification.m:560)
8   UIKitCore                     	0x00000001843d3654 -[UIApplication _terminateWithStatus:] + 300 (UIApplication.m:6919)
9   UIKitCore                     	0x0000000183abbb38 -[_UISceneLifecycleMultiplexer _evalTransitionToSettings:fromSettings:forceExit:withTransitionStore:] + 128 (_UISceneLifecycleMultiplexer.m:808)
10  UIKitCore                     	0x0000000183cb0a80 -[_UISceneLifecycleMultiplexer forceExitWithTransitionContext:scene:] + 224 (_UISceneLifecycleMultiplexer.m:482)
11  UIKitCore                     	0x00000001843d029c -[UIApplication workspaceShouldExit:withTransitionContext:] + 212 (UIApplication.m:3828)
12  FrontBoardServices            	0x0000000192cef780 __63-[FBSWorkspaceScenesClient willTerminateWithTransitionContext:]_block_invoke_2 + 80 (FBSWorkspaceScenesClient.m:331)
13  FrontBoardServices            	0x0000000192caa06c -[FBSWorkspace _calloutQueue_executeCalloutFromSource:withBlock:] + 240 (FBSWorkspace.m:352)
14  FrontBoardServices            	0x0000000192cef718 __63-[FBSWorkspaceScenesClient willTerminateWithTransitionContext:]_block_invoke + 132 (FBSWorkspaceScenesClient.m:328)

Posted by

LM-Emanuel

@LM-Emanuel, this is probably due to a watchdog (0x8badf00d), but I'm not 100% sure in this case. Note your main thread is touching some threading primitives, which just happened to be during the app's termination time. Your background threads also show multiple frames that are unsymbolicated. but and also waiting on some threading primitives. I suggest you fully symbolicate the crash log](https://developer.apple.com/documentation/xcode/adding-identifiable-symbol-names-to-a-crash-report) and then open a Technical Support Incident with the fully symbolicated report so that we can easily discuss your situation further.

—
edford
Hi @edford, thanks for your answer. I've already opened a ticket with the crash log fully symbolicated here https://feedbackassistant.apple.com/feedback/9857467

—
LM-Emanuel

for @edford as requested above.

FB9858254 has 4 SIGKILL crash reports with no Termination Reason in them.

Kind Regards

Posted by

minion1989

@minion1989, these all look like 0x8badf00d for a blocked main thread. Note the time stamps, you've only been alive for a little bit of time, but long enough to be past the launch time watchdog, so this is likely due to a main thread that's being kept too busy to keep up with event processing crucial for a responsive app. Your main thread stacks are long and deep into a third-party framework, so I suggest you take this information and consult with your third-party for how to ensure your app remains responsive with the features the stack trace shows you are using from their framework.

—
edford

for @edford as requested above.

FB9860396 has SIGKILL without a Termination Reason in them.

Thanks.

Posted by

ramonvasc7

@ramonvasc7, are you a CallKit app? With the crucial information missing, I can't be 100% sure, but my first guess is that these represent the 0xbaadca11 exception code, indicating that you didn't respond to PushKit notifications. I'm not personally well-versed in the details of CallKit, so if you need additional help, please open a Technical Support Incident to discuss with my colleagues in more detail.

—
edford

Thankyou for this thread @edford both of the top occuring crashes suffer from this issue. Here is the feedback assistant link you wanted; https://feedbackassistant.apple.com/feedback/9869185.

Posted by

LKWD_JamesD

@LKWD_JamesD, thanks for opening your feedback. These all appear to be watchdog terminations, i.e. 0x8badf00d. Since you have a third-party game engine involved, your best course of action is to review some of your code with support for their product to ensure you don't stall the main thread in their environment. Note in one report, the app is trying to terminate, and code from that product is waiting on a lock to proceed. In another, there is use of audio frameworks, also driven through this game engine.

—
edford
Thank you for the response @edford. Any news on why this information isn't in the report? Is that an OS issue or a development/reporting issue? Anything I can track from Apples side about potential fixes?

—
LKWD_JamesD

Many of you in this thread have opened bug reports in Feedback Assistant — thank you for filing those to help us address this! We have deployed a fix to the backend services that provide crashes for your app into the Xcode Organizer, so if you receive EXC_CRASH (SIGKILL) crash reports from now on through the Organizer, those reports should be attributed with an exception code to explain why the system terminated your app.

If you still see any reports without the exception code, please post the complete crash report to this thread using the attachment feature so we can verify that the code is still missing.

Posted by

edford

@edford I'm seeing reports without exception code obtained through TestFlight as detailed here

Posted by

yj21496

@eskimo beat me to responding in that other thread, and it looks like you're set with the info you need now. Let us know if that's not the case!

—
edford

Hi edford,

Has several SIGKILL crash reports with no Termination Reason.

https://feedbackassistant.apple.com/feedback/9906125.

Thanks!

Posted by

MingXiao

@MingXiao All of the crash reports attached to FB9906125 have exception codes, but they look different than you're probably expecting. Please see this other thread for details.

—
edford
@edford Thank you. That's big!

—
MingXiao

@edford: I appreciate your contribution to this thread. I encountered a similar situation. When I downloaded the crash reports from TestFlight, there was no 'reason' field. Then I read all the replies and your comments and found that I needed to check the crash reports using the XCode crashers organizer.

I am highlighting this for future readers:

To find the 'reason' field, It is important that you analyze the crash using XCode crashes organizer. Find your crash in the organizer, right-click and then click 'show in finder. This will be an XCcrashPoint file. Open it and the reason field should be there. See attached screenshot

Hope it helps.

Awais

Posted by

awaisfayyaz17.apple

I am attaching my recent crash report as an attachment here. I have anlysed the crash report and i believe my app was stuck in some kind of deadlock. I am using locks to manager synchronization. Can you please spare some minutes and verify that my assumption is correct.

Bundle of thanks.

front-board.crash

Posted by

awaisfayyaz17.apple

You're on the right path. Notice a lot of the background threads are blocked on __ulock_wait coming from work in FHKit, and your main thread is thus also many frames deep with that framework and GCD also in the mix. This crash is different than all of the others in this thread (yours has the SIGKILL info). If you need a more in-depth discussion on how to proceed, please open a Technical Support Incident and include the crash report.

—
edford

We continue to have problems determine the root cause of Intermittant APP crash. We were told to look at the Third Party SDK that opens and closes the Lidar camera. We have confirmed we are opening and closing the camera properly. We built a test app to open and close the app and it ran 150 times without a problem.

Could a thread that saves the JPG pictures as a video cause the sensor_stream to hang and crash the app? Here is snippets of the crash logs. Any further help would be appreciated.

Crash Log1 Exception Type: EXC_BAD_ACCESS (SIGSEGV) Exception Subtype: KERN_INVALID_ADDRESS at 0x0000000134500000 Thread 46 name: Thread 46 Crashed: 0 libsystem_platform.dylib 0x00000001f110da84 _platform_memmove + 52 1 UnityFramework 0x0000000107c4eb60 Sensor_Stream_mFF850E96E26ADC438DFA62EFFABFDAEDE3B6DEEC + 632 (LightBuzz.BodyTracking.cpp:24277)

Crash Log2 Triggered by Thread: 56. Kernel Triage: VM - pmap_enter failed with resource shortage Thread 56 name: Thread 56 Crashed: 0 libsystem_platform.dylib 0x00000001f110da84 _platform_memmove + 52 1 UnityFramework 0x0000000109362b58 Sensor_Stream_mFF850E96E26ADC438DFA62EFFABFDAEDE3B6DEEC + 632 (LightBuzz.BodyTracking.cpp:24277)

Posted by

Daveinthehat

Hi 👋! He tried to solve this situation. Any clue would be greatly appreciated.

https://feedbackassistant.apple.com/feedback/11192902

Posted by

4cv1_

The issue discussed in this thread is resolved, and your report contains the expected info showing this is a a watchdog termination. To discuss further, please create a new forums post with the complete crash report provided in an attachment, or contact Developer Technical Support if you can't share the crash report publicly. Note the report attached to FB11192902 is not fully symbolicated, so you should first symbolicate your app's frames.

—
edford

@edford, I have also faced this issue and filed the report here: Feedback. Can you take a look at this one too, please?

Posted by

gharary