Open Bug 1060009 Opened 5 years ago Updated 3 years ago
Investigate native crash at /dev/ashmem/libmozalloc
.so reported on Google Play
This bug is for investigating a huge jump in native crash at /dev/ashmem/libmozalloc.so as reported on Google Play affecting mainly Samsung based devices on all channels in Android 4.3 and Android 4.4. User Comments never work after the new update. was trying to use clip board to add comment to social feed and crashed Does not reliably paste; shuts down arbitrarily. I was trying to comment on an article and Firefox stopped when I hit "reply". Top Devices Galaxy S4 (jflte) Galaxy Note3 (hlte) Galaxy S3 (m0)
Attachment #8480822 - Attachment is obsolete: true
I saw some near hangs using Gmail desktop and Galaxy S3 I9300 mainly in the compose window editing the recipients. Switching to Google keyboard was a bit smoother.
(In reply to Aaron Train [:aaronmt] from comment #0) > > Top Devices > > Galaxy S4 (jflte) I have this device (I9505XXUGNG8, 4.4.2 KOT49H)
Hmm no luck running into any crashes on my S4. Looking at the dashboard now, this has now increased by 143% since last week with an increase in 5,723 reports.
Mark: can we send out a Twitter blast, see if anyone can reliably hit this, and what they're doing to do so?
(In reply to Richard Newman [:rnewman] from comment #6) > Mark: can we send out a Twitter blast, see if anyone can reliably hit this, > and what they're doing to do so? What would the tweet say? "Does Firefox crash?" I think we need a more focused STR to get useful responses.
Using a Samsung phone or tablet? Did Firefox suddenly start crashing in the last week or two? What are you doing when it crashes? What keyboard do you use?
Maybe some of Samsung's updates might have caused this? http://samsung-updates.com/
I haven't seen any crashes, Verizon Galaxy S4. I use Swype.
I also use nightly, not the market version.
My Nexus5, Android 4.4.4, KTU84P crashed repeatedly. I was reading tweets using the twitter app, then reading gmail and opened a link to the Cordova issues server e.g. https://issues.apache.org/jira/browse/CB-7310 from one email. Nightly crashed each time I restarted it because the offending page was reloaded at startup. The crash reporter sent the report and I allowed it to include the page url. I uninstalled Nightly and reinstalled it the next day. Today I can not reproduce the crashes.
(In reply to Axel Nennker from comment #12) > My Nexus5, Android 4.4.4, KTU84P crashed repeatedly. > I was reading tweets using the twitter app, then reading gmail and opened a > link to the Cordova issues server e.g. > https://issues.apache.org/jira/browse/CB-7310 from one email. > Nightly crashed each time I restarted it because the offending page was > reloaded at startup. > The crash reporter sent the report and I allowed it to include the page url. > I uninstalled Nightly and reinstalled it the next day. > Today I can not reproduce the crashes. I think in the case here, we're not seeing the crash-reporter for this signature. Regardless, thanks for reporting.
I think we might actually be getting crash reports for these, at least for some of them. This one looks similar: https://crash-stats.mozilla.com/report/list?product=FennecAndroid&query_type=contains&range_unit=days&process_type=any&hang_type=any&signature=libmozalloc.so%400xe52&date=2014-09-03+15%3A00%3A00&range_value=28#tab-sigsummary
Some of the reports above dip down to API 16 and the devices are scattered with everything
Finkle sent me some of the attached logcats on crash-stats, and I found that there were some assertion failures that caused these. I'm thinking this one is too.
Anyone have a Moto G? "Hi, i am using moto g and if i open any link from any note then firefox crashed. Work on it. - Umang"
I spent a couple hours yesterday trying to get Firefox to crash on my I9505 (S4) and I couldn't. I did hit many input related hangs though.
The crashes here seem similar for 32. They are all assertion failures in IPC (thread_local_posix.cc:33) https://crash-stats.mozilla.com/report/list?product=FennecAndroid&query_type=contains&range_unit=days&process_type=any&hang_type=any&signature=libmozalloc.so%400xe5a&date=2014-09-09+15%3A00%3A00&range_value=28#tab-reports
(In reply to James Willcox (:snorp) (firstname.lastname@example.org) from comment #19) > The crashes here seem similar for 32. They are all assertion failures in IPC > (thread_local_posix.cc:33) > > https://crash-stats.mozilla.com/report/ > list?product=FennecAndroid&query_type=contains&range_unit=days&process_type=a > ny&hang_type=any&signature=libmozalloc.so%400xe5a&date=2014-09- > 09+15%3A00%3A00&range_value=28#tab-reports A couple are xpcom_runtime_abort( ###!!! ABORT: Could not read the system fonts directory: file /home/moz/fennec/gecko/gfx/thebes/gfxFT2FontList.cpp, line 1189), custom roms likely
A couple comments yesterday mention opening links from Facebook. Trying both via native & web app.
Rob and I looked into this. Distinct spike in crashes was on Aug 22nd on release Firefox but also on Aug 22nd on Beta Firefox per Google play. I'm fairly confident in saying that it's not caused by anything in 32. Android releases take time to filter out, so if it were caused by that, I'd expect a ramp and not a sudden spike like we're seeing. I would guess that a website updated, an ad is circulating that crashes us, or a software update in a 3rd party program (including possible malware).
From Sept 20, we can see a huge drop in term of crash (going back to the previous level). cf screenshot Closing as wontfix?
That's really awkward and concerning.
So weird. I think we're still going to try to get to the bottom of this.
Anyway, I am going to untrack it. Please resubmit if it appears again.
You need to log in before you can comment on or make changes to this bug.