Crash at [FIRInstanceIDTokenManager checkTokenRefreshPolicyWithIID:] function

[phaninderkumar]

Environment:

• Xcode version: 11.0
• Firebase SDK version: 6.11.0
• Firebase Component: Core, Messaging
• Component version: 6.3.2, ~> 4.1.7

Problem:
My project keeps crashing at [FIRInstanceIDTokenManager checkTokenRefreshPolicyWithIID:] function.
My app collects location in suspended state if it exits a region. My assumption is this crashes happens at that time. Not always though.

Steps to reproduce:

Configure the firebase project and make sure the app is launched from suspended state multiple times

Attaching my crash log here
crashReport.txt

[phaninderkumar]

Here is the list of Firebase pods being used:

• Firebase/Core (6.11.0):
  • Firebase/CoreOnly
  • FirebaseAnalytics (= 6.1.3)
• Firebase/CoreOnly (6.11.0):
  • FirebaseCore (= 6.3.2)
• Firebase/Messaging (6.11.0):
  • Firebase/CoreOnly
  • FirebaseMessaging (~> 4.1.7)
• FirebaseAnalytics (6.1.3):
  • FirebaseCore (~> 6.3)
  • FirebaseInstanceID (~> 4.2)
  • GoogleAppMeasurement (= 6.1.3)
  • GoogleUtilities/AppDelegateSwizzler (~> 6.0)
  • GoogleUtilities/MethodSwizzler (~> 6.0)
  • GoogleUtilities/Network (~> 6.0)
  • "GoogleUtilities/NSData+zlib (~> 6.0)"
  • nanopb (~> 0.3.901)
• FirebaseAnalyticsInterop (1.4.0)
• FirebaseCore (6.3.2):
  • FirebaseCoreDiagnostics (~> 1.0)
  • FirebaseCoreDiagnosticsInterop (~> 1.0)
  • GoogleUtilities/Environment (~> 6.2)
  • GoogleUtilities/Logger (~> 6.2)
• FirebaseCoreDiagnostics (1.1.1):
  • FirebaseCoreDiagnosticsInterop (~> 1.0)
  • GoogleDataTransportCCTSupport (~> 1.0)
  • GoogleUtilities/Environment (~> 6.2)
  • GoogleUtilities/Logger (~> 6.2)
  • nanopb (~> 0.3.901)
• FirebaseCoreDiagnosticsInterop (1.0.0)
• FirebaseInstanceID (4.2.6):
  • FirebaseCore (~> 6.0)
  • GoogleUtilities/Environment (~> 6.0)
  • GoogleUtilities/UserDefaults (~> 6.0)
• FirebaseMessaging (4.1.7):
  • FirebaseAnalyticsInterop (~> 1.3)
  • FirebaseCore (~> 6.2)
  • FirebaseInstanceID (~> 4.1)
  • GoogleUtilities/AppDelegateSwizzler (~> 6.2)
  • GoogleUtilities/Environment (~> 6.2)
  • GoogleUtilities/Reachability (~> 6.2)
  • GoogleUtilities/UserDefaults (~> 6.2)
  • Protobuf (>= 3.9.2, ~> 3.9)
• GoogleAppMeasurement (6.1.3):
  • GoogleUtilities/AppDelegateSwizzler (~> 6.0)
  • GoogleUtilities/MethodSwizzler (~> 6.0)
  • GoogleUtilities/Network (~> 6.0)
  • "GoogleUtilities/NSData+zlib (~> 6.0)"
  • nanopb (~> 0.3.901)
• GoogleDataTransport (3.0.1)
• GoogleDataTransportCCTSupport (1.2.1):
  • GoogleDataTransport (~> 3.0)
  • nanopb (~> 0.3.901)
• GoogleUtilities/AppDelegateSwizzler (6.3.1):
  • GoogleUtilities/Environment
  • GoogleUtilities/Logger
  • GoogleUtilities/Network
• GoogleUtilities/Environment (6.3.1)
• GoogleUtilities/Logger (6.3.1):
  • GoogleUtilities/Environment
• GoogleUtilities/MethodSwizzler (6.3.1):
  • GoogleUtilities/Logger
• GoogleUtilities/Network (6.3.1):
  • GoogleUtilities/Logger
  • "GoogleUtilities/NSData+zlib"
  • GoogleUtilities/Reachability
• "GoogleUtilities/NSData+zlib (6.3.1)"
• GoogleUtilities/Reachability (6.3.1):
  • GoogleUtilities/Logger
• GoogleUtilities/UserDefaults (6.3.1):
  • GoogleUtilities/Logger
• KeychainSwift (11.0.0)
• MBProgressHUD (1.1.0)
• nanopb (0.3.904):
  • nanopb/decode (= 0.3.904)
  • nanopb/encode (= 0.3.904)
• nanopb/decode (0.3.904)
• nanopb/encode (0.3.904)
• Protobuf (3.10.0)

[maksymmalyhin]

@chliangGoogle Could you help with the issue investigation please. Here is what I investigated so far.

Looks like the crash happens because cachedIID is nil here. The particular crash itself is fixed in #4096.
As far as I can see from the code cachedIID should be nil only in case of the Keychain access failure. The Keychain access failure in the background may be a caused by using deprecated kSecAttrAccessibleAlwaysThisDeviceOnly value for kSecAttrAccessible on iOS 13 here. I assume it should be replaced by kSecAttrAccessibleAfterFirstUnlockThisDeviceOnly.

[charlotteliang]

@maksymmalyhin You are right, it is strange that IID becomes nil when a token exists. I haven't been able to reproduce on my side. I agree #4096 should fix this but we can also adjust the keychain option here.

[charlotteliang]

@phaninderkumar can you also send us the debugLog right before the crash if you can reproduce? This can help us figure out why the instanceID is not retrieved successfully

[phaninderkumar]

@chliangGoogle, I'm not able to reproduce this crash while the app is running. It happens when the app is launched from suspended state when the devices is locked and hence the only logs I have are from the Fabric crash logs.

[charlotteliang]

@phaninderkumar Thanks, this confirms that it is caused by the security framework on keychain access level, we've updated it at #4172. But your crash should be resolved in #4096 which will release soon.