Divested-Mobile/Mull-Fenix

Constant Crashes on GrapheneOS

Closed this issue · 11 comments

When visiting a website I get these soft crashes. Once it works normally, then freezes, after that I constantly get "Mull does not react" messages. They only give logs on the first opening.

type: ANR
osVersion: google/bluejay/bluejay:14/AP2A.240705.004/2024071600:user/release-keys
package: us.spotco.fennec_dos:21270220
process: us.spotco.fennec_dos
installer: dev.imranr.obtainium

activity: us.spotco.fennec_dos/org.mozilla.fenix.HomeActivity
cause: ANR Input dispatching timed out (5337a86 us.spotco.fennec_dos/org.mozilla.fenix.HomeActivity (server) is not responding. Waited 5001ms for MotionEvent). Buffer processing for the associated surface is stuck due to an unsignaled fence (window=SurfaceView[us.spotco.fennec_dos/org.mozilla.fenix.HomeActivity](BLAST)#35363, bufferId=0x00003CB60000002C, frameNumber=449). This potentially indicates a GPU hang.
info: ANR in us.spotco.fennec_dos (us.spotco.fennec_dos/org.mozilla.fenix.HomeActivity)
PID: 15542
Reason: Input dispatching timed out (5337a86 us.spotco.fennec_dos/org.mozilla.fenix.HomeActivity (server) is not responding. Waited 5001ms for MotionEvent). Buffer processing for the associated surface is stuck due to an unsignaled fence (window=SurfaceView[us.spotco.fennec_dos/org.mozilla.fenix.HomeActivity](BLAST)#35363, bufferId=0x00003CB60000002C, frameNumber=449). This potentially indicates a GPU hang.
Parent: us.spotco.fennec_dos/org.mozilla.fenix.HomeActivity
ErrorId: cb708b47-7c1f-49c7-9da8-2bfa4bdceffe
Frozen: false
Timestamp: 2024-07-22 00:20:01.272+0200
Load: 5.08 / 4.06 / 3.12
----- Output from /proc/pressure/memory -----
some avg10=1.73 avg60=2.72 avg300=1.56 total=520324622
full avg10=0.96 avg60=1.53 avg300=0.81 total=286331680
----- End output from /proc/pressure/memory -----
----- Output from /proc/pressure/cpu -----
some avg10=12.53 avg60=13.26 avg300=12.24 total=9937421335
----- End output from /proc/pressure/cpu -----
----- Output from /proc/pressure/io -----
some avg10=1.00 avg60=1.57 avg300=0.99 total=525763419
full avg10=0.54 avg60=0.98 avg300=0.65 total=380028862
----- End output from /proc/pressure/io -----

CPU usage from 1ms to 12993ms later (2024-07-22 00:20:01.336 to 2024-07-22 00:20:14.329) with 99% awake:
  43% 1437/system_server: 26% user + 16% kernel / faults: 60718 minor 4234 major
  40% 15542/us.spotco.fennec_dos: 25% user + 14% kernel / faults: 60958 minor 2006 major
  30% 17639/com.futo.platformplayer: 18% user + 11% kernel / faults: 62598 minor 3878 major
  28% 1927/com.android.systemui: 20% user + 7.9% kernel / faults: 47561 minor 6103 major
  24% 18832/im.molly.app: 15% user + 8.9% kernel / faults: 2986 minor 2589 major
  17% 542/surfaceflinger: 11% user + 5.5% kernel / faults: 318 minor 80 major
  15% 18629/us.spotco.fennec_dos:gpu: 9.8% user + 6.1% kernel / faults: 4115 minor 144 major
  12% 20815/us.spotco.fennec_dos:tab34: 8.7% user + 3.5% kernel / faults: 9881 minor 407 major
  11% 93/kswapd0: 0% user + 11% kernel
  5.5% 544/android.hardware.graphics.composer@2.4-service: 3.7% user + 1.7% kernel / faults: 58 minor 101 major
44% TOTAL: 20% user + 18% kernel + 2.7% iowait + 1.7% irq + 1.1% softirq
CPU usage from 48ms to 483ms later (2024-07-22 00:20:01.384 to 2024-07-22 00:20:01.818):
  101% 15542/us.spotco.fennec_dos: 55% user + 45% kernel / faults: 13571 minor 31 major
    97% 15562/Signal Catcher: 55% user + 41% kernel
  65% 18367/kworker/0:1-mali_mmu2: 0% user + 65% kernel
  39% 93/kswapd0: 0% user + 39% kernel
  40% 1437/system_server: 8.7% user + 32% kernel / faults: 621 minor 33 major
    37% 20995/AnrAuxiliaryTas: 8.7% user + 29% kernel
  21% 17639/com.futo.platformplayer: 14% user + 7.1% kernel / faults: 8 minor
    3.5% 17730/ExoPlayer:Playb: 0% user + 3.5% kernel
    3.5% 18134/MediaCodec_loop: 3.5% user + 0% kernel
    3.5% 18138/HwBinder:17639_: 3.5% user + 0% kernel
    3.5% 18145/MediaCodec_loop: 3.5% user + 0% kernel
  10% 882/samsung.hardware.media.c2@1.2-service: 8.1% user + 2.7% kernel / faults: 3 minor
    2.7% 18147/HwBinder:882_5: 2.7% user + 0% kernel
    2.7% 18153/HwBinder:882_5: 2.7% user + 0% kernel
    2.7% 18155/HwBinder:882_5: 2.7% user + 0% kernel
  7.5% 542/surfaceflinger: 5% user + 2.5% kernel
    5% 542/surfaceflinger: 5% user + 0% kernel
  4.6% 71/rcuop/7: 0% user + 4.6% kernel
  5.7% 1098/media.swcodec: 5.7% user + 0% kernel / faults: 11 minor 1 major
    5.7% 18139/oid.aac.decoder: 5.7% user + 0% kernel
  2.3% 49/rcuog/4: 0% user + 2.3% kernel
  2.5% 544/android.hardware.graphics.composer@2.4-service: 2.5% user + 0% kernel
    5% 592/HwBinder:544_1: 2.5% user + 2.5% kernel
  2.5% 707/loop15: 0% user + 2.5% kernel
  2.7% 898/audioserver: 2.7% user + 0% kernel
    2.7% 1417/AudioOut_15: 2.7% user + 0% kernel
  4.1% 20815/us.spotco.fennec_dos:tab34: 4.1% user + 0% kernel / faults: 3 minor
46% TOTAL: 13% user + 25% kernel + 2% iowait + 4.9% irq + 0.5% softirq

How often?
Just some websites? or many websites?
How many extensions do you have?

I have this issue every few days.

I have a looot of extensions and should probably try to reproduce without them :/

As it is happening so rarely, this is pretty annoying but I can try.

Extensions:

UBlock Origin
NoScript
Firefox Translate
Audiocontext Fingerprint Defender
Canvas Fingerprint Defender
Font Fingerprint Defender
WebGL fingeprint defender
Bypass Paywalls Clean (D)
Cookie Autodelete
Download all images
IPvFoo
"Enable Read mode"
LibRedirect
Offline QR Code
SingleFile
Tomato Clock
Wayback machine
Old Reddit Redirect
ETag Stoppa

I will try to remove the nonessential ones first.

I was able to reproduce without any addons.

Specifically using the website bike-discount.de

When adding things to the cart (writing cookies?) It constantly crashed.

I have 2 logs, can I send them to you personally? Not sure if they include sensitive details

Please try to reproduce on official Firefox for Android and report to Mozilla here if it happens: https://bugzilla.mozilla.org/home

True, I will do that.

I sent the crash logs every time, where do they end up?

I tried on latest Firefox, opened a bunch of tabs, used the same website, was not able to reproduce at all.

Even did the Web Basemark, it just doesnt crash.

I sent the crash logs every time, where do they end up?

In Mull? They don't go anywhere.

Lol thats not good.

Anyways here are some logs.

Mull-crash-no-addons.txt

An older log

Mull log e5211f169708.txt

These were generated by GrapheneOS/Android.

I am currently unable to reproduce on the same site. This may be a memory leak or oom issue. I closed all my previous apps before.

Maybe there is a memory management issue and firefox crashes. I have 91 open tabs but they should not be in RAM.

Maybe this happens with over 100 tabs?

Did you wrongly try to enable Fission via about:config?

yes I did! This could be the reason, thanks, that was my mistake

It is crazy that fission doesnt work to this day...

Update: I havent had crashes anymore since a long time.

Might have been an extension, maybe NoScript which I replaced with UBO.

Thanks for your work!