1479783 - Crash in java.lang.IllegalStateException: eglMakeCurrent failed EGL_BAD_MATCH at android.view.HardwareRenderer$GlRenderer.createSurface(HardwareRenderer.java)

Status: RESOLVED INCOMPLETE

(Firefox for Android Graveyard :: General, defect, P2)

Description

[Marcia Knous [:marcia]]

This bug was filed from the Socorro interface and is
report bp-baec8688-b77f-4f66-b16f-3886b0180730.
=============================================================

Not sure where this crash belongs, but is visible in b11 crash stats and has been rising throughout 62: https://bit.ly/2n1uIup. It appears as if only API 19 (Kit kat) is affected. 

This crash is present in other releases, but filing since it seems on the rise in 62 and currently sits at #12 browser crash.

Top 10 frames of crashing thread:

0 libxul.so GeckoAppShellSupport::ReportJavaCrash widget/android/nsAppShell.cpp:280
1 libdvm.so libdvm.so@0x83795 
2 libdvm.so libdvm.so@0xe0b5c 
3 data@[e-mail][e-mail] data@[e-mail][e-mail]@0x5caf82 
4 libxul.so libxul.so@0x19a27cd 
5 dalvik-heap (deleted) dalvik-heap @0x36f7f 
6 dalvik-LinearAlloc (deleted) dalvik-LinearAlloc @0x353b7f 
7 dalvik-heap (deleted) dalvik-heap @0x8c5f5f 
8 data@[e-mail][e-mail] data@[e-mail][e-mail]@0x386a1f 
9 libdvm.so libdvm.so@0x239d19 

=============================================================

Comment 1

[Pascal Chevrel:pascalc]

The spike last week may be related with the Samsung S8 Android update that was deployed last week (at least in Europe)

Comment 2

[Marcia Knous [:marcia]]

37 crashes/24 installs in b13 and sits at #11 in b13. We can continue to monitor this throughout the rest of the beta cycle. In 61.0 we have about 1500 crashes.

Comment 3

[Julien Cristau [:jcristau] (back April 22)]

This got worse around the end of July, and is now #10 in 62.0b15 (with many signatures in the top 10 being various kinds of OOM situations).

Comment 4

[:sdaswani]

Added to work queue.

Comment 5

[Marcia Knous [:marcia]]

This is #7 overall in 63.0b9, but the volume isn't that concerning and the 62.0.2 release volume is about 300 crashes.

Comment 6

[Marcia Knous [:marcia]]

This hasn't yet showed up in 63 release. Since the volume has been low, I think this is one we can remove the whiteboard tag.

Comment 7

[Kevin Brosnan [Ex-Mozilla]]

This signature changed to [java.lang.IllegalStateException: at android.view.HardwareRenderer$GlRenderer.createSurface(HardwareRenderer.java)]

https://crash-stats.mozilla.com/signature/?product=FennecAndroid&signature=java.lang.IllegalStateException%3A%20at%20android.view.HardwareRenderer%24GlRenderer.createSurface%28HardwareRenderer.java%29#summary

Comment 8

[Marcia Knous [:marcia]]

Adding another related signature which is showing up, affects 65 as well.

Comment 9

[Petru-Mugurel Lingurar [:petru]]

This crashes happen only on Android API <=19 and come from two EGL 1.0 methods that Android uses, called automatically when a layout needs to be displayed.
Possibly because they receive bad arguments, possibly a bug in that API. (latest spec being EGL 1.5), but with a reasonably low crash rate it's hard to identify a scenario that would allow us to pinpoint the issue.

- https://crash-stats.mozilla.com/signature/?product=FennecAndroid&signature=java.lang.IllegalStateException%3A%20at%20android.view.HardwareRenderer%24GlRenderer.createSurface%28HardwareRenderer.java%29
happens because of https://www.khronos.org/registry/EGL/sdk/docs/man/html/eglMakeCurrent.xhtml
Only on API 19 and API 18 from the same place
https://android.googlesource.com/platform/frameworks/base/+/kitkat-release/core/java/android/view/HardwareRenderer.java#1354

- https://crash-stats.mozilla.com/signature/?signature=java.lang.RuntimeException%3A%20at%20android.view.HardwareRenderer%24GlRenderer.createSurface(HardwareRenderer.java)
happens because of https://www.khronos.org/registry/EGL/sdk/docs/man/html/eglCreateWindowSurface.xhtml
Only on API 15 - 19 from the same place
https://android.googlesource.com/platform/frameworks/base/+/kitkat-release/core/java/android/view/HardwareRenderer.java#1349

- https://crash-stats.mozilla.com/signature/?signature=java.lang.IllegalStateException%3A%20eglMakeCurrent%20failed%20EGL_BAD_MATCH%20at%20android.view.HardwareRenderer%24GlRenderer.createSurface(HardwareRenderer.java)
Doesn't show anymore.


It would be helpful to get the detail message for that IllegalState/RuntimeExceptions but that doesn't seem to have been caught by our crash reporter.
There are no useful comments from the users which have experienced this as to what they were doing when the app crashed.
Hoping that the QA team will catch this in their normal app testing and report back with any STR.

Comment 10

[Liz Henry (:lizzard) (relman/hg->git project)]

I think from the relatively moderate volume of this crash signature on release 62, we call this P2. It also looks like we've already investigated and can call this stalled until we have steps to reproduce the crash.

Comment 11

[Chris Peterson [:cpeterson]]

[geckoview:p3] because this crash is not actionable at this time.

Comment 12

[BMO Automation]

We have completed our launch of our new Firefox on Android. The development of the new versions use GitHub for issue tracking. If the bug report still reproduces in a current version of [Firefox on Android nightly](https://play.google.com/store/apps/details?id=org.mozilla.fenix) an issue can be reported at the [Fenix GitHub project](https://github.com/mozilla-mobile/fenix/). If you want to discuss your report please use [Mozilla's chat](https://wiki.mozilla.org/Matrix#Connect_to_Matrix) server https://chat.mozilla.org and join the [#fenix](https://chat.mozilla.org/#/room/#fenix:mozilla.org) channel.

Comment 13

[BugBot [:suhaib / :marco/ :calixte]]

Since the bug is closed, the stalled keyword is now meaningless.
For more information, please visit auto_nag documentation.