I have a location service based tracking and geofencing app which would run for days and weeks in the background on and iOS 12.2 ff device.
Now with iOS 13.2 the app gets terminated after a variable amount of time, but at least several hours, due to excessive cpu usage:
Date/Time: 2019-11-09 23:25:18 +0200
End time: 2019-11-09 23:26:06 +0200
OS Version: iPhone OS 13.2 (Build 17B84)
Architecture: arm64
Report Version: 29
Incident Identifier: 5B46660C-A347-477F-8AE2-B1401080892B
Data Source: Microstackshots
Shared Cache: 0x44db8000 94FD24C8-F407-3A82-8D27-367F5B6C7BEC
Command: Anchorwatch
Path: /private/var/containers/Bundle/Application/32CD03B2-449C-4E84-8E06-79FA9B50F3A9/Anchorwatch.app/Anchorwatch
Identifier: de.sioned.Anchorwatch
Version: 2.2.1 (14)
Beta Identifier: 68A95EFC-B341-476C-9277-D711242471EC
PID: 57424
Event: cpu usage
Action taken: Process killed
CPU: 48 seconds cpu time over 48 seconds (99% cpu average), exceeding limit of 80% cpu over 60 seconds
CPU limit: 48s
Limit duration: 60s
CPU used: 48s
CPU duration: 48s
Duration: 48.39s
Duration Sampled: 14.08s
Steps: 16
Hardware model: iPad6,12
Active cpus: 2
Heaviest stack for the target process:
16 ??? (libsystem_pthread.dylib + 49032) [0x1c4f64f88]
16 ??? (libdispatch.dylib + 74516) [0x1c4ecb314]
16 ??? (libdispatch.dylib + 36344) [0x1c4ec1df8]
16 ??? (libdispatch.dylib + 33488) [0x1c4ec12d0]
16 ??? (libdispatch.dylib + 104312) [0x1c4ed2778]
16 ??? (libdispatch.dylib + 33948) [0x1c4ec149c]
16 ??? (libdispatch.dylib + 124996) [0x1c4ed7844]
16 ??? (libdispatch.dylib + 125216) [0x1c4ed7920]
16 ??? (libsystem_kernel.dylib + 158196) [0x1c50419f4]
Powerstats for: Anchorwatch [57424]
Bundle ID: de.sioned.Anchorwatch
Adam ID: 0
Is first party: No
App version: 2.2.1
Build version: 14
Is Beta: No
Share with Devs: No
UUID: 1C943425-70F9-3670-98D0-45D3051B4BB7
Path: /private/var/containers/Bundle/Application/32CD03B2-449C-4E84-8E06-79FA9B50F3A9/Anchorwatch.app/Anchorwatch
Architecture: arm64
Footprint: 1046.66 MB
Start time: 2019-11-09 23:25:52 +0200
End time: 2019-11-09 23:26:06 +0200
Num samples: 16 (100%)
CPU Time: 13.971s
Primary state: 15 samples Non-Frontmost App, Non-Suppressed, Kernel mode, Effective Thread QoS Background, Requested Thread QoS Default, Override Thread QoS Unspecified
User Activity: 0 samples Idle, 0 samples Active, 16 samples Unknown
Power Source: 0 samples on Battery, 0 samples on AC, 16 samples Unknown
16 _pthread_wqthread + 275 (libsystem_pthread.dylib + 49032) [0x1c4f64f88]
16 _dispatch_workloop_worker_thread + 587 (libdispatch.dylib + 74516) [0x1c4ecb314]
16 _dispatch_lane_invoke$VARIANT$mp + 419 (libdispatch.dylib + 36344) [0x1c4ec1df8]
16 _dispatch_lane_serial_drain$VARIANT$mp + 299 (libdispatch.dylib + 33488) [0x1c4ec12d0]
16 _dispatch_mach_invoke$VARIANT$mp + 471 (libdispatch.dylib + 104312) [0x1c4ed2778]
16 _dispatch_lane_serial_drain$VARIANT$mp + 759 (libdispatch.dylib + 33948) [0x1c4ec149c]
16 _dispatch_event_loop_drain$VARIANT$mp + 315 (libdispatch.dylib + 124996) [0x1c4ed7844]
16 _dispatch_kq_drain + 123 (libdispatch.dylib + 125216) [0x1c4ed7920]
16 kevent_id + 8 (libsystem_kernel.dylib + 158196) [0x1c50419f4]
1 <User mode>
Binary Images:
0x102d80000 - ??? Anchorwatch <1C943425-70F9-3670-98D0-45D3051B4BB7> /private/var/containers/Bundle/Application/32CD03B2-449C-4E84-8E06-79FA9B50F3A9/Anchorwatch.app/Anchorwatch
0x1c4eb9000 - 0x1c4f2dfff libdispatch.dylib <B7EED4C7-560D-3DA6-9B50-ED52A150AAC6> /usr/lib/system/libdispatch.dylib
0x1c4f59000 - 0x1c4f69fff libsystem_pthread.dylib <F8B082D8-24D9-3B1E-B80B-645FC8A88E14> /usr/lib/system/libsystem_pthread.dylib
0x1c501b000 - 0x1c5048fff libsystem_kernel.dylib <AE4C3D7A-7D08-33E7-BCC6-11AC821B4E48> /usr/lib/system/libsystem_kernel.dylib
While in background the app does nothing more than to write each location update to a sqlite database and checking the current location against a safety perimeter. There is no reason to assume that the app suddenly after several hours would require such a high cpu usage.
I have not the slightest idea how to approach the problem and I am not even sure if I can do anything about it or if it is a bug in 13.2 and I have to wait for a fix.
My first assumption was, that the old iOS 12.0 bug which terminated background apps without reason and which was fixed in 12.2 was back. But that here seems to be something new. I was able to run the test app iOS 12 terminates apps in the background for no reason for several days.
Any hint how to interprete the log and take action ?
Edit:
It seems, that this is related to iOS 13.2 message: nehelper sent invalid result code [1] for Wi-Fi information request
Location service queries the WiFi interface in second intervals. Instrument shows that these query calls have a memory leak.
Turning off WiFi on the device seems to solve the problem, though that is no true solution. Now waiting for iOS 13.3 to leave beta.
It turned out, that a third party framework that I am using launches calls to CNCopyNetworkInfo in second intervals. These calls could not fullfilled as my App had no WiFi capabilities enabled and thus the calls to CNCopyNetworkInfo caused small memory leaks that added up over time.
After enabling WiFi access capabilities the memory leaks vanished.