1670970 - Crash in [@ gdk_x11_keymap_get_group_for_state]

Status: RESOLVED WORKSFORME

(Core :: Widget: Gtk, defect)

Description

[Julien Cristau [:jcristau]]

Crash report: https://crash-stats.mozilla.org/report/index/ca8b89bf-cb08-4e13-8823-040ee0201013

Reason: SIGTRAP

Top 10 frames of crashing thread:

0 libglib-2.0.so.0 g_log_writer_default 
1 libglib-2.0.so.0 g_log_structured_array 
2 libglib-2.0.so.0 g_log_structured_standard 
3 libgdk-3.so.0 gdk_x11_keymap_get_group_for_state 
4 libX11.so.6 _XError 
5 libX11.so.6 _XFreeX11XCBStructure 
6 libX11.so.6 _XFreeX11XCBStructure 
7 libX11.so.6 _XReply 
8 libX11.so.6 XTranslateCoordinates 
9 libgdk-3.so.0 gdk_x11_screen_get_current_desktop 

Seeing these crashes starting October 3rd, from archlinux users mostly.

Comment 1

[Sean Feng [:sefeng211]]

Martin, this bug isn't moving in the right direction and still occurring recently. Can you take a look?

Comment 2

[Martin Stránský [:stransky] (ni? me)]

Let's add null pointer check to mGdkWindow and see if that helps.

Comment 3

[Martin Stránský [:stransky] (ni? me)]

Attachment: Bug 1670970 [Linux] Check GdkWindow before we pass it to gdk_window_get_origin(), r?jhorak

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/83e587a40f30
[Linux] Check GdkWindow before we pass it to gdk_window_get_origin(), r=jhorak

Comment 5

[Cosmin Sabou [:CosminS]]

https://hg.mozilla.org/mozilla-central/rev/83e587a40f30

Comment 6

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Crash volume is unaffected from the changes. Please take a look.

Comment 7

[Martin Stránský [:stransky] (ni? me)]

Hi Karl, do you think it's similar to Bug 1673767, i.e. we may need to make some X11 call non-fatal?
Thanks.

Comment 8

[Karl Tomlinson (:karlt)]

https://crash-stats.mozilla.org/report/index/ad05cc72-6259-4336-9272-af98f0210110 looks the same as bug 1673767 and I don't see any reports in versions where bug 1673767 either didn't exist or is fixed.

Comment 9

[Karl Tomlinson (:karlt)]

Sorry, I failed to check the crash report in comment 0, which is different, being in gdk_window_get_origin().

I can't think of a good reason why mGdkWindow would be non-null, but XTranslateCoords() would generate an X error.

It's possible that the error is from a previous X request.
Fixing bug 1682402 might provide more clues.

The reason for the majority of crashes with this signature should now be fixed by bug 1673767. All the recent crash reports that I've seen are in XGetWindowProperty() and so are different from the one in comment 0.

Comment 10

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

Closing because no crashes reported for 12 weeks.