856613 - Sony Xperia Crashes - NoSuchMethodError: massageValue @ java.util.IdentityHashMap.put(IdentityHashMap.java:489)

Status: RESOLVED WORKSFORME

(Firefox for Android Graveyard :: General, defect)

Description

[Aaron Train [:aaronmt]]

Attachment: Raw log (Sony Xperia Neo L, Android 4.0.4)

E "04-01 04:25:21.288" 15992 GeckoAppShell ">>> REPORTING UNCAUGHT EXCEPTION FROM THREAD 1006 (""IntentService[AnnouncementsBroadcastServiceWorker]"")" 16024 787
E "04-01 04:25:21.288" 15992 GeckoAppShell "java.lang.NoSuchMethodError: massageValue" 16024 788
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at java.util.IdentityHashMap.put(IdentityHashMap.java:489)" 16024 789
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at org.mozilla.gecko.sync.log.writers.AndroidLevelCachingLogWriter.shouldLogDebug(AndroidLevelCachingLogWriter.java:78)" 16024 790
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at org.mozilla.gecko.sync.log.writers.AndroidLevelCachingLogWriter.debug(AndroidLevelCachingLogWriter.java:112)" 16024 791
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at org.mozilla.gecko.sync.log.writers.TagLogWriter.debug(TagLogWriter.java:38)" 16024 792
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at org.mozilla.gecko.sync.log.writers.TagLogWriter.debug(TagLogWriter.java:38)" 16024 793
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at org.mozilla.gecko.sync.Logger.debug(Logger.java:199)" 16024 794
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at org.mozilla.gecko.sync.Logger.debug(Logger.java:142)" 16024 795
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at org.mozilla.gecko.background.announcements.AnnouncementsBroadcastService.onHandleIntent(AnnouncementsBroadcastService.java:83)" 16024 796
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at android.app.IntentService$ServiceHandler.handleMessage(IntentService.java:65)" 16024 797
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at android.os.Handler.dispatchMessage(Handler.java:99)" 16024 798
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at android.os.Looper.loop(Looper.java:137)" 16024 799
E "04-01 04:25:21.288" 15992 GeckoAppShell "	at android.os.HandlerThread.run(HandlerThread.java:60)" 16024 800
F "04-01 04:25:21.288" 15992 "libc    " "Fatal signal 11 (SIGSEGV) at 0x00000000 (code=1)" 16024 801

https://appthwack.com/public/YJ343d9ck4

Reproducible on the device with ARMv7 builds. There might be other Xperia devices out there with similar issues.

Comment 1

[Richard Newman [:rnewman]]

This is an evil one. Apparently Sony is shipping a different IdentityHashMap implementation. No alternative other than "don't use java.util.IdentityHashMap".

Comment 2

[Mark Finkle (:mfinkle) (use needinfo?)]

(In reply to Richard Newman [:rnewman] from comment #1)
> This is an evil one. Apparently Sony is shipping a different IdentityHashMap
> implementation. No alternative other than "don't use
> java.util.IdentityHashMap".

Or ship our own

Comment 3

[Mark Finkle (:mfinkle) (use needinfo?)]

This crash is happening from inside the impl of Log.debug which seems like something that would affect every other Android app running on the phone. Something doesn't smell right here.

Comment 4

[Nick Alexander :nalexander [he/him]]

(In reply to Mark Finkle (:mfinkle) from comment #3)
> This crash is happening from inside the impl of Log.debug which seems like
> something that would affect every other Android app running on the phone.
> Something doesn't smell right here.

No, it's inside the implementation of Android Sync's debug logging framework.  This log writer caches Android's log level so we don't have to look them up each time we log.  This code runs (frequently!) on a huge range of devices and we've never seen this problem.  I agree with rnewman: this is a "who screwed up java.util.IdentityHashMap" issue and the initial finger points to Sony.  We can try to work around it by shipping our own hash map, or replacing IdentityHashMap with a different Java Collection, or even not caching log levels on some devices.

Comment 5

[Richard Newman [:rnewman]]

(In reply to Mark Finkle (:mfinkle) from comment #2)
> > No alternative other than "don't use
> > java.util.IdentityHashMap".
> 
> Or ship our own

That's why I included the package prefix ;)

Comment 6

[Chris Peterson [:cpeterson]]

(In reply to Nick Alexander :nalexander from comment #4)
> this is
> a "who screwed up java.util.IdentityHashMap" issue and the initial finger
> points to Sony.

Sony has shipped devices that broke binary compatibility with Stagefright video libraries, so I would not be surprised if they have also made "improvements" to Java libraries.

Comment 7

[Mattias Nilsson]

Hi,
Sony Mobile have not patched in IdentityHashMap.java, it is original android 4.0.4 and the massageValue method is in place.
We have tried to reproduce this issue with Firefox Sync, the same phone SW, Firefox version 20 with no luck. This crash is a strange one that we can see in low numbers in our crash statistics and all times related to the Firefox for android browser.

However with Firefox Sync we can 100% reproduce https://bugzilla.mozilla.org/show_bug.cgi?id=752772. After that the phone is unusable and loop crashes.

BR
Mattias
Sony Mobile Developer

Comment 8

[Richard Newman [:rnewman]]

Sounds like we need to start grubbing through bytecode! Thanks for checking on this, Mattias.

I'll see if we can get some traction on Bug 752772, also.

Comment 9

[Richard Newman [:rnewman]]

This isn't in crash-stats for the last 30 days. Is this still an issue?

Comment 10

[Kevin Brosnan [Ex-Mozilla]]

Lets just close it, reopening or filing a new bug is trivial. This has not seen any new data in over 5 months. No crashes in 30 days is good enough for me.