Issue metadata
Sign in to add a comment
|
Native crash at /system/lib/libc.so
Reported by
gary.lis...@sasets.com,
Jun 3 2016
|
||||||||||||||||||||||
Issue description
Steps to reproduce the problem:
1. Open our Cordova App
2. Perform a save action,
3. Crash after ajax completes
Android 5.0 on a Galaxy Tab S 8.4 (klimtlte)
I appreciate the repro steps aren't that helpful as that is all I have from my user that is experiencing this issue. The webview is crashing repeatedly on him after he performs a save. I'm just hoping there's something in the stack trace below that can help uncover the issue
*** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
Build fingerprint: 'samsung/klimtltexx/klimtlte:5.0.2/LRX22G/T705XXU1BOJ8:user/release-keys'
Revision: '10'
ABI: 'arm'
pid: 21848, tid: 21893, name: Compositor >>> com.XXX.XXX <<<
signal 6 (SIGABRT), code -6 (SI_TKILL), fault addr --------
Abort message: '[FATAL:synchronous_compositor_output_surface.cc(97)] Renderer compositor context loss
'
r0 00000000 r1 00005585 r2 00000006 r3 00000000
r4 9c588db8 r5 00000006 r6 00000016 r7 0000010c
r8 b6effdf4 r9 9c58895c sl 00000000 fp a39b89ac
ip 00005585 sp 9c5884b0 lr b6ea958d pc b6ecd4c0 cpsr 600f0010
backtrace:
#00 pc 0003b4c0 /system/lib/libc.so (tgkill+12)
#01 pc 00017589 /system/lib/libc.so (pthread_kill+52)
#02 pc 0001819b /system/lib/libc.so (raise+10)
#03 pc 00014ae5 /system/lib/libc.so (__libc_android_abort+36)
#04 pc 00012aac /system/lib/libc.so (abort+4)
#05 pc 01879469 /data/app/com.google.android.webview-2/lib/arm/libwebviewchromium.so
What is the expected behavior?
What went wrong?
App crash (Cordova)
Crashed report ID:
How much crashed? Just one tab
Is it a problem with a plugin? N/A
Did this work before? N/A
Chrome version: Channel: n/a
OS Version: 5.0
Flash Version:
,
Jun 3 2016
Reliably every single time? If yes, can you share the apk and exact repro steps? If not, then this should not happen anymore starting in m51
,
Jun 3 2016
,
Jun 3 2016
Hi, I'm waiting to hear back from the user but I'm 99% sure it is not every single time, it's intermittent. Apologies for not being up to speed on terminology here, m51 refers to the Android System WebView version 51?
,
Jun 3 2016
> Apologies for not being up to speed on terminology here, m51 refers to the Android System WebView version 51? yep
,
Jun 3 2016
Ok thanks for your help, is there any common cause to the bug fix in m51 that I could potentially work around to mitigate the issue until the fix arrives?
,
Jun 3 2016
That crash indicates that there is a EGL context loss on the device. Before M49, the webview will just limp along, but rendering is probably broken until the process is restarted. In M49, I made all context loss crash immediately. And starting in M51, webview *should* be able to recover from context loss by recreating it (chrome has done this for a looooong time). What causes context loss? Who knows, out of memory, hit a corner case in a driver, or just generally bad drivers. We'll only look at those if they are reliably reproducible. Might be helpful to look at logs before the crash
,
Jun 3 2016
Ok, thanks for your help with this, will do my best to get logs but as it's not reliably reproducible I think the best course of action is to just wait for M51.
,
Jun 3 2016
|
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by rsgav...@chromium.org
, Jun 3 2016