- User Since
- Aug 25 2014, 5:54 AM (317 w, 6 d)
Feb 22 2019
Jun 11 2018
Some correction since I opened that ticket:
Feb 14 2018
So, I'm not sure I correctly qualified that ticket at all, or if the behavior changed with last E compiled version and/or new kernel drivers or whatever.
Feb 7 2018
So, the problem is really linked to the dock-embeded sound card. If I remove kernel module "snd-usb-audio", I'm able to click on F1 (restart E) when the crash occures and enlightenment actually restarts.
In that case, .e-crashdump.txt only contains:
Jan 23 2018
Thanks for the link, very well explained.
Jan 22 2018
@zmike I have no idea about how to do that, but if you have some link to a how-to or documentation, I can try
I recompilled, but nothing changed regarding the crash (as you thought)
Jan 19 2018
I just tested back, the trace is almost identical, so yes.
In fact, I only have alsa-lib for dependencies (chromimum, thinderbird etc). I was forced to switch to pulseaudio because of firefox which droped ALSA support.
So I have :
I have a dell XPS 9560 with arch linux. I'm able to connect to the dock without problem, it's when I restart enlightenment with [ctrl][alt][end] when the doc is connected that there is a crash. And then, all [F1] crash in turn. I need to disconnect the dock, and then hit F1, and then connect back the doc.
@bu5hm4n how can I help you on that? I can reproduce it even with the last Dell firmware. It may be higly computer / kernel / driver dependent, I don't know (but usb-c is extremelly complexe...)
Nov 27 2017
I can confirm it doesn't happen anymore in 0.22.1
Nov 24 2017
Sep 27 2017
I noticed something related to that bug. Sometime, not sure how/when, but if I try to open virtual box console several time, it works on the second or third try. By "works", I mean that the first time I get the "E segfault, F1/F12" pop-up and the console is closed, and on the second try it get the E segfault pop-up and after F1 (reload) the console stay open and I can use it.
Sep 13 2017
@zmike: it is recognizing if the laptop is but not if the plug was done during suspend. I will try to find the root cause of the problem, it used to works... But as you said, you can't do much if the problem is at that level :)
Very clear. It was what I thought given the log etc. Something is broken somewhere between udev, intel driver, or something else. I'm very good now at building an xorg stack with debug info turns on, so as soon as I have the motivation to do it, I will report.
Sep 12 2017
Also (so that I don't always tell bad things here :), since I have that behaviour, I also have a *much* better experience when connecting/disconnecting screens: all my windows positions on screens are restored, it is really nice :)
Sep 3 2017
Sep 1 2017
I tried with and without it, and it seems to behave exactly in the same way. Is there a way I could check that the option is correctly enabled (for ex an action to try or something?)
Aug 31 2017
I found the problem, it is pure firefox and linked to electrolysis in conjunction with ff55. This one can be closed, and I would like to thank you all for all the help, it provided the necessary push to get me solved that and other problem in efl/enlightenment and a segfault in intel video driver !
Aug 30 2017
Sorry, I forget to add that I also compiled enlightenment from git:
I used efl-git, my current version is:
Aug 29 2017
It's here for the time being, let me know if you want it elsewhere: https://gist.github.com/fanf/6325ff4ef709bd18834ce5b700e5e2c8
I had to recompile mesa to get the full debug logs (it was long and not pleasant :)
I used "-O -g -ffast-math -march=native -ggdb3" for the second case. Does I also need to compile efl?
@zmike: would you please tell me what I should use for CFLAGS, because I don't know C at all. I'm willing to help, thought.
I compiled enlightement from git (commit dece94b6ca558af603d0aab8644ad94f0f7c85bd from today) with CFLAGS="-O -g -ffast-math -march=native -ggdb3"
Mozilla dev were emitting that hypothesis, so I searched a way to confirm/infirm it. I reported it to because of your comment "some kind of driver/distro/firefox bug." to remove the "driver" part. Now, it only remains distro/firefox (and let be optimistic and add "hardware" to :))
I can reproduce it with the "nouveau" driver and my discrete nvidia video card (and with the intel one disabled on bios to be sure it does not interfere), on a freshly created user with an empty home (and so new firefox profil, new enlightenment config).
Aug 28 2017
Well, this is good news for E and FF. Not so for me, because it means that there is very little chance someone get to understand what is happening on my side :(
Sorry @zmike , (obviously) I wasn't clear. In the first post, I said that the problem happen if I either connect an external monitor or restart E. This is false. I can connect/disconnect monitors without the problem. The problem only happen when I restart (ctrl+alt+end) E and nothing else. The fact that I noticed the problem after connection/disconnection of external monitor is purely incidental (and most likelly happened because I almost also systemactically restart E after such an operation by habits because things in Fri, Aug 25, 11:43 PM comment (ie some focus problems, and lost of windows edge swithching).
I need to amend my first report: the switch from "every tabs are ok" to "spinning wheel on them" DOES NOT occures when I'm docking/undocking my laptop (so not related to external monitor) but happen when I restart enlightenment.
Aug 27 2017
For information, during more tests I found that segfault in i915/xorg driver: https://bugs.freedesktop.org/show_bug.cgi?id=102442
can you go through the trouble shooting guide on Intel drivers on the archwiki and see if SNA issues, DRI3 issues (or any other) is the case for you?
Aug 25 2017
can you try running firefox via a terminal and post that output?
- can you try if this happens with software rendering? save your work, changing this setting restarts e. it should be safe, but anyways.
- Settings -> Look -> Compositor -> Rendering -> Engine
Aug 14 2017
OK, so after double-checking, the previous stack trace isn't present anymore with efl 1.20.2-1 / enlightenment 0.21.8-6, but I have that one which seem (from my non-dev point of view) maybe related:
Removed the two last comments for inacurracies. Let met test that more properly and come back. Very sorry for the noise :(
Jun 28 2017
Jun 27 2017
I happen to have my .Xsession filled with that kind of errors:
May 30 2017
I can confirm that with kernel linux 4.11:
Feb 2 2016
One shelf, bottom, above everything.
I have only one shelf, which takes the total wideness of the bottom of the screen.
Nov 25 2015
If it helps, here is a screenshot with dejavu sans mono: https://www.enlightenment.org/ss/e-565429ca4e0b37.80422233.jpg
Nov 24 2015
I can confirm that:
- on current arch, terminology font rendering is broken
- this happened when upgrading to harfbuzz-1.1.0-1
- reverting to harfbuzz-1.0.6-2 correct the font rendering issue.