Bug 2061188 - [abrt] epiphany: WTF::CrashOnOverflow::crash()(): epiphany killed by SIGABRT
Summary: [abrt] epiphany: WTF::CrashOnOverflow::crash()(): epiphany killed by SIGABRT
Keywords:
Status: CLOSED UPSTREAM
Alias: None
Product: Fedora
Classification: Fedora
Component: epiphany
Version: 35
Hardware: x86_64
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
Assignee: Michael Catanzaro
QA Contact: Fedora Extras Quality Assurance
URL: https://retrace.fedoraproject.org/faf...
Whiteboard: abrt_hash:556e66324bdad604f2054d0bfa1...
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2022-03-06 18:45 UTC by XoD
Modified: 2022-03-07 19:07 UTC (History)
10 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2022-03-07 19:07:54 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
File: backtrace (37.36 KB, text/plain)
2022-03-06 18:45 UTC, XoD
no flags Details
File: core_backtrace (37.80 KB, text/plain)
2022-03-06 18:45 UTC, XoD
no flags Details
File: cpuinfo (2.68 KB, text/plain)
2022-03-06 18:45 UTC, XoD
no flags Details
File: dso_list (160 bytes, text/plain)
2022-03-06 18:45 UTC, XoD
no flags Details
File: environ (1.59 KB, text/plain)
2022-03-06 18:45 UTC, XoD
no flags Details
File: maps (3.92 KB, text/plain)
2022-03-06 18:45 UTC, XoD
no flags Details
File: open_fds (2.89 KB, text/plain)
2022-03-06 18:45 UTC, XoD
no flags Details
File: proc_pid_status (1.38 KB, text/plain)
2022-03-06 18:45 UTC, XoD
no flags Details


Links
System ID Private Priority Status Summary Last Updated
WebKit Project 237535 0 None None None 2022-03-07 19:07:53 UTC

Description XoD 2022-03-06 18:45:24 UTC
Description of problem:
crash when start application with webkit gtk

Version-Release number of selected component:
epiphany-1:41.3-1.fc35

Additional info:
reporter:       libreport-2.15.2
backtrace_rating: 4
cgroup:         0::/user.slice/user-1000.slice/user/app.slice/app-gnome-org.gnome.Epiphany-4380.scope/4380
cmdline:        epiphany
crash_function: WTF::CrashOnOverflow::crash()
executable:     /usr/bin/epiphany
journald_cursor: s=ea8e0111ba2841ab81420fedc872d878;i=243e99;b=c48c84ea0b6645fe88b5555977466407;m=724dd56;t=5d990dc1bb698;x=4ab7b996d9619663
kernel:         5.16.12-200.fc35.x86_64
rootdir:        /
runlevel:       N 5
type:           CCpp
uid:            1000

Comment 1 XoD 2022-03-06 18:45:28 UTC
Created attachment 1864266 [details]
File: backtrace

Comment 2 XoD 2022-03-06 18:45:29 UTC
Created attachment 1864267 [details]
File: core_backtrace

Comment 3 XoD 2022-03-06 18:45:31 UTC
Created attachment 1864268 [details]
File: cpuinfo

Comment 4 XoD 2022-03-06 18:45:32 UTC
Created attachment 1864269 [details]
File: dso_list

Comment 5 XoD 2022-03-06 18:45:33 UTC
Created attachment 1864270 [details]
File: environ

Comment 6 XoD 2022-03-06 18:45:34 UTC
Created attachment 1864271 [details]
File: maps

Comment 7 XoD 2022-03-06 18:45:36 UTC
Created attachment 1864272 [details]
File: open_fds

Comment 8 XoD 2022-03-06 18:45:37 UTC
Created attachment 1864273 [details]
File: proc_pid_status

Comment 9 Michael Catanzaro 2022-03-06 20:08:39 UTC
Any chance you're able to reproduce this? It's crashing when processing your GL version string, which is *weird* to say the least.

Comment 10 XoD 2022-03-06 20:14:55 UTC
I reproduce it systematically seens some day, and have the problem with geary (witch also use webkit-gtk).

With gdb I have seen than : for an unknow reason, in GLContext::version :
 ::glGetString(GL_VERSION) return an empty string, so versionStringComponents is empty, and access to versionStringComponents[0] crash.

It's strange than glGetString(GL_VERSION) return an empty string, but it's seems bad than access to versionStringComponents wasn't protected.

Comment 11 Michael Catanzaro 2022-03-06 20:32:11 UTC
Are you using NVIDIA graphics driver, or have you otherwise done something else to replace mesa?

Comment 12 XoD 2022-03-06 20:40:26 UTC
No NVIDIA, I use mesa stock with a AMD card.
It's seems that it's appear after an update some day's ago.

And I'm able to play game (peraps more with vulkan), glxgears work and glxinfo say :
OpenGL vendor string: AMD
OpenGL renderer string: AMD Radeon RX 480 Graphics (POLARIS10, DRM 3.44.0, 5.16.12-200.fc35.x86_64, LLVM 13.0.0)
...
OpenGL version string: 4.6 (Compatibility Profile) Mesa 21.3.0 (git-4a94b5d22a)
OpenGL shading language version string: 4.60

Comment 14 Michael Catanzaro 2022-03-06 23:24:09 UTC
Looking at: https://stackoverflow.com/questions/12184506/why-does-glgetstringgl-version-return-null-zero-instead-of-the-opengl-versio

I wonder if some previous GL call has failed, or if OpenGL is somehow in a bad state. It might possibly be a variant of https://bugs.webkit.org/show_bug.cgi?id=201507, which we never figured out.

Comment 15 XoD 2022-03-07 16:20:19 UTC
I do additional debug (with gdb) :
in tryInitializeEGL GLContext::createOffscreenContext seems to initialize correctly the context.
eglContext->makeContextCurrent seems also work correctly.

But call to GLContext::current()->version() always fail.

I don't kow how to continue debug to have more informations. I have tried to step into ::glGetString(GL_VERSION), but gdb don't allow me to do this.

Comment 16 Michael Catanzaro 2022-03-07 16:23:15 UTC
(In reply to XoD from comment #15)
> I don't kow how to continue debug to have more informations. I have tried to
> step into ::glGetString(GL_VERSION), but gdb don't allow me to do this.

The most important next step is to report this on WebKit Bugzilla, WebKitGTK component, if you're willing to take the time to create an account there. I don't know what's wrong either, but you've got a great backtrace here and have made more progress with gdb than most people who report bugs, so there's at least some chance of getting to the bottom of it. On WebKit Bugzilla, you'll be able to engage directly with the right developers.

Comment 17 XoD 2022-03-07 17:52:44 UTC
(In reply to Michael Catanzaro from comment #16)

> The most important next step is to report this on WebKit Bugzilla, WebKitGTK
> component, if you're willing to take the time to create an account there. I
> don't know what's wrong either, but you've got a great backtrace here and
> have made more progress with gdb than most people who report bugs, so
> there's at least some chance of getting to the bottom of it. On WebKit
> Bugzilla, you'll be able to engage directly with the right developers.

Ok, thank-you.
I'm experienced programmer in C++, but with limited skill in linux tools and ecosystem ;)

I have created this bug : https://bugs.webkit.org/show_bug.cgi?id=237535


Note You need to log in before you can comment on or make changes to this bug.