Closed 0o220 closed 4 years ago
I don’t see a link to logs.i3wm.org. Did you follow https://i3wm.org/docs/debugging.html? (In case you actually provided a link to a logfile, please ignore me.)
@ohfill, @KubaJastrz, are you still experiencing this issue?
Happens for me as well, but I haven't been able to reproduce it either.
Yes, I see it almost everyday for the past few months. As I said here, I'd connect this issue with hardware acceleration in blink-based software. After appending --disable-gpu
flag to executables the issue doesn't occur.
For software like Figma, that requires hardware acceleration, you can temporarily solve the bug by killing gpu process:
alias find-gpu='ps aux | grep gpu-process | grep'
function kill-gpu() {
find-gpu $1 | egrep -o -e "[0-9]+" | head -n 1 | xargs -p kill
}
And just call kill-gpu <process-name>
everytime it happens.
It has happen to me with a bunch of electron apps! Not only with I3 but with Awesome also!
Signs point to this being a bug in chromium, someone should find the bug in their tracker or open a new issue there and reply with the link here
but never before then, so perhaps this issue began with the i3 upgrade to 4.16.1 on 27 Jan 2019
4.16.1 included only a handful of regression fixes so this is probably a coincidence.
I'm fairly certain that this issue started happening after this upgrade: https://pastebin.com/raw/urbwzv6p
May be related: https://bugs.chromium.org/p/chromium/issues/detail?id=370022
This has been plaguing me for months, as well. Example applications that "freeze":
Funny enough, they still accept keyboard-input but they no longer redraw. It happened once while typing this and was able to ^A, ^C
to save the comment :stuck_out_tongue:
One additional anecdote that I haven't seen mentioned is that it only happens when I'm using my laptop as a laptop. If I'm connected to a dock (eg: Lenovo Thunderbolt 3 Dock), then it never happens.
$ hwinfo --gfxcard
22: PCI 02.0: 0300 VGA compatible controller (VGA)
[Created at pci.386]
Unique ID: _Znp.Vg1X6XR+hGA
SysFS ID: /devices/pci0000:00/0000:00:02.0
SysFS BusID: 0000:00:02.0
Hardware Class: graphics card
Model: "Intel VGA compatible controller"
Vendor: pci 0x8086 "Intel Corporation"
Device: pci 0x5916
SubVendor: pci 0x17aa "Lenovo"
SubDevice: pci 0x224e
Revision: 0x02
Driver: "i915"
Driver Modules: "i915"
Memory Range: 0xeb000000-0xebffffff (rw,non-prefetchable)
Memory Range: 0x60000000-0x6fffffff (ro,non-prefetchable)
I/O Ports: 0xe000-0xe03f (rw)
Memory Range: 0x000c0000-0x000dffff (rw,non-prefetchable,disabled)
IRQ: 169 (83544423 events)
Module Alias: "pci:v00008086d00005916sv000017AAsd0000224Ebc03sc00i00"
Config Status: cfg=new, avail=yes, need=no, active=unknown
Primary display adapter: #22
The ArchLinux wiki for Intel Graphics refers to the previously mentioned chromium issue: https://wiki.archlinux.org/index.php/intel_graphics#DRI3_issues
The recommendation is to change from DRI 3
-> DRI 2
. Will see if this works but I don't like that the issue has been open for 5 years and this only started this year for me. 🍺(ツ)/¯
Also experiencing this with i3 4.16.1
on Fedora 30 (with Chromium, Riot and Keybase).
Same problem here with i3 4.17.1
on Arch Linux (kernel version 5.3.5
).
The problem appears with all chromium-based application (Google Chrome, Atom, Boostnote, Discord, Visual Studio).
The ArchLinux wiki for Intel Graphics refers to the previously mentioned chromium issue: https://wiki.archlinux.org/index.php/intel_graphics#DRI3_issues
The recommendation is to change from
DRI 3
->DRI 2
. Will see if this works but I don't like that the issue has been open for 5 years and this only started this year for me. 🍺(ツ)/¯
It seems to work for me! But with this solution after some time my touchpad starts freezing until it stops working (everything else continues working). Maybe it's a DRI2 related problem.
@mattiabiondi
But with this solution after some time my touchpad starts freezing until it stops working
As in your touchscreen or the touch/trackpad? Asking because my touch screen stops working periodically and I never associated with this issue. :P
@coxley
As in your touchscreen or the touch/trackpad? Asking because my touch screen stops working periodically and I never associated with this issue. :P
I was referring to the touch/trackpad, I have a touchscreen too but I didn't try it.
I have also been running into this lately, it triggers very reliably with both chrome and inkscape open for me. I have vertically oriented monitors using xrandr. Inkscape stops drawing/refreshing periodically and then comes back.
i3 version 4.16.1 (2019-01-27) Chrome 78.0.3904.108 (Official Build) (64-bit)
My colleague with exact same versions but horizontal displays don't seem to be suffering any problems and I didn't have problems until I turned both my 1440p monitors vertical.
Edit: Inkscape and chrome still work, but it appears as if they (especially inkscape) stop being redrawn after damage some times until it randomly starts again.
I am having this issue as well. I am able to reproduce it consistently - it occurs with Slack any time I receive a direct message. The app receives the message (I hear the ding sound) but the window freezes and won't redraw for about a minute or so. I tried changing my intel driver to use DRI 2
with no success.
nick@razer > i3 --moreversion 2>&- || i3 --version
Binary i3 version: 4.17.1 (2019-08-30) © 2009 Michael Stapelberg and contributors
Running i3 version: 4.17.1 (2019-08-30) (pid 2362)abort…)
Loaded i3 config: /home/nick/.config/i3/config (Last modified: Tue 10 Dec 2019 08:48:27 AM PST, 98914 seconds ago)
The i3 binary you just called: /usr/bin/i3
The i3 binary you are running: /usr/bin/i3
- Linux Distribution & Version: 5.4.2-arch1-1-ARCH x86_64
- Are you using a compositor (e.g., xcompmgr or compton): Yes, compton
I used a minimal setup to reproduce the issue and uploaded logs here: https://logs.i3wm.org/logs/5734685623713792.bz2
Hopefully that helps!
@ClydeDroid are you using a notification daemon? If not, you can try installing one like Dunst.
Also, since I've installed Picom (ex Compton) my problem has disappeared.
I am using i3 with KDE, so I kept the Plasma notifications enabled. Now that you mention it, though, the notifications aren't appearing, so maybe that has something to do with the issue.
I am using this fork of picom that allows drawing rounded windows: https://github.com/sdhand/compton
The issue is reproducible with and without compton running, though.
@mattiabiondi Installing Dunst fixed the issue!! Thanks a bunch!!
@mattiabiondi Installing Dunst fixed the issue!! Thanks a bunch!!
@ClydeDroid Perfect! I'm glad it helped :)
After troubleshooting this issue for the better part of a year, I opted to reinstall my OS in hopes that the problem won't reappear. Will report back here if I experience freezing again.
I am now running NixOS and it works fine. I will report back as well if it happens again! :)
I see that there are a lot of related issues regarding freezing in qutebrowser, electron etc:
https://github.com/qutebrowser/qutebrowser/issues/4809 https://github.com/qutebrowser/qutebrowser/issues/4641 https://github.com/electron/electron/issues/21415
Closing as not i3-specific
I'm submitting a…
Current Behavior
For the past 2–3 months (but never before then, so perhaps this issue began with the i3 upgrade to 4.16.1 on 27 Jan 2019), Chromium-based applications such as qutebrowser and Slack will periodically "freeze", requiring a
SIGKILL
and restart. I'll be able to move my cursor across the application window, but it won't respond to input from the keyboard or mouse. If I switch to another workspace and then back again, the frozen application's window won't be redrawn—that window will simply continue to display whatever was on the last workspace I just switched from. Affected applications tend to display this behavior at the same time. I've been monitoring application-specific logs without finding any clues to the root cause. This issue appears to be related to #1720.Expected Behavior
Applications should not freeze.
Reproduction Instructions
I've not been able to trigger this issue manually. It's quite unpredictable. I can go days without it happening—but sometimes, it'll occur several times back-to-back with no apparent cause.
Environment
Output of
i3 --moreversion 2>&-
: