search

ValveSoftware / steam-for-linux

Issue tracking for the Steam for Linux beta client
4.23k stars 174 forks source link

Opening steam with running google chrome randomly hang/crash whole system #9793

Open Pharmasolin opened 1 year ago

Pharmasolin commented 1 year ago

Your system information

Please describe your issue in as much detail as possible:

I noticed that from time to time when I have opened google chrome and I open steam in few moments I will see popup that steam/google chrome is not responding, and if I will kill any of the crashed apps - whole system will hang and crash. I will need to power off and power on my laptop.

Steps for reproducing this issue:

  1. No clear steps, I have opened google chrome, then open steam and randomly this steam app startup will crash my whole system. This started when steam switched to a new UI.

-- I will re-generate logs again after my OS will crash again and I will reboot PC, may be those logs will be more helpful. https://gist.github.com/Pharmasolin/43290fc327cf54b7b64a04c19dff8c52

kisak-valve commented 1 year ago

Hello @Pharmasolin, in general, no OpenGL application including Steam should be capable of taking down the whole system and symptoms like this usually hint towards a video driver issue. It might be worthwhile to also mention this issue to your video driver vendor.

Pharmasolin commented 1 year ago

@kisak-valve I'm afraid that's the issue with the gnome/mutter version on LTS Ubuntu. Because the desktop hangs I can't do anything but music from Spotify is playing.

And there was a bug with chromium apps in gnome. Please let this issue be open for a while, maybe someone will add more context.

Pharmasolin commented 1 year ago

This is the output from journalctl -b -1 | grep -v 'slack.desktop\[1712\]' (log where I turned on steam, I hid mentions of the slack in the log, because there are a lot of info/warnings from it there). Basically I had turned on google chrome, slack and steam (all chromium based apps?) Not sure if this bug related to steam itself and where to report. If someone will read this issue, please let me know where I should report this one.

лип 04 23:00:45 denys-Legion-5-15ACH6A steam.desktop[72282]: ExecuteSteamURL: "steam://open/steam://settings/"
лип 04 23:00:56 denys-Legion-5-15ACH6A systemd[1092]: vte-spawn-2fe99b25-4477-41e6-b715-6a96aaf3296d.scope: Consumed 14.748s CPU time.
лип 04 23:02:15 denys-Legion-5-15ACH6A dbus-daemon[1135]: [session uid=1000 pid=1135] Activating service name='org.gnome.Nautilus' requested by ':1.34' (uid=1000 pid=1292 comm="/usr/bin/gnome-shell ")
лип 04 23:02:15 denys-Legion-5-15ACH6A dbus-daemon[1135]: [session uid=1000 pid=1135] Successfully activated service 'org.gnome.Nautilus'
лип 04 23:02:15 denys-Legion-5-15ACH6A org.gnome.Nautilus[80663]: Initializing nautilus-dropbox 2020.03.04
лип 04 23:02:15 denys-Legion-5-15ACH6A dbus-daemon[688]: [system] Activating via systemd: service name='org.freedesktop.hostname1' unit='dbus-org.freedesktop.hostname1.service' requested by ':1.259' (uid=1000 pid=80663 comm="/usr/bin/nautilus --gapplication-service ")
лип 04 23:02:15 denys-Legion-5-15ACH6A systemd[1]: Starting Hostname Service...
лип 04 23:02:15 denys-Legion-5-15ACH6A nautilus[80663]: Called "net usershare info" but it failed: Failed to execute child process “net” (No such file or directory)
лип 04 23:02:15 denys-Legion-5-15ACH6A dbus-daemon[688]: [system] Successfully activated service 'org.freedesktop.hostname1'
лип 04 23:02:15 denys-Legion-5-15ACH6A systemd[1]: Started Hostname Service.
лип 04 23:02:38 denys-Legion-5-15ACH6A nautilus[80663]: Called "net usershare info" but it failed: Failed to execute child process “net” (No such file or directory)
лип 04 23:02:45 denys-Legion-5-15ACH6A systemd[1]: systemd-hostnamed.service: Deactivated successfully.
лип 04 23:02:54 denys-Legion-5-15ACH6A gnome-shell[1292]: meta_window_set_stack_position_no_sync: assertion 'window->stack_position >= 0' failed
лип 04 23:04:34 denys-Legion-5-15ACH6A systemd[1092]: Started Application launched by gnome-shell.
лип 04 23:04:34 denys-Legion-5-15ACH6A gnome-keyring-daemon[1131]: asked to register item /org/freedesktop/secrets/collection/login/2, but it's already registered
лип 04 23:04:34 denys-Legion-5-15ACH6A google-chrome.desktop[80892]: libva error: vaGetDriverNameByIndex() failed with unknown libva error, driver_name = (null)
лип 04 23:04:50 denys-Legion-5-15ACH6A systemd[1092]: Started Application launched by gnome-shell.
лип 04 23:04:51 denys-Legion-5-15ACH6A gnome-system-monitor.desktop[81374]: glibtop(c=81374): [WARNING] statvfs '/run/user/1000/doc' failed: Operation not permitted
лип 04 23:04:51 denys-Legion-5-15ACH6A gnome-system-monitor.desktop[81374]: glibtop(c=81374): [WARNING] statvfs '/run/user/1000/doc' failed: Operation not permitted
лип 04 23:04:51 denys-Legion-5-15ACH6A gnome-system-monitor.desktop[81374]: glibtop(c=81374): [WARNING] statvfs '/run/user/1000/doc' failed: Operation not permitted
лип 04 23:04:51 denys-Legion-5-15ACH6A gnome-system-monitor.desktop[81374]: glibtop(c=81374): [WARNING] statvfs '/run/user/1000/doc' failed: Operation not permitted
лип 04 23:04:51 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Ping serial 36632114 was reused for window W452, previous use was for window 0x2000010.
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Overwriting existing binding of keysym 37 with keysym 37 (keycode 10).
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Overwriting existing binding of keysym 38 with keysym 38 (keycode 11).
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Overwriting existing binding of keysym 31 with keysym 31 (keycode a).
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Overwriting existing binding of keysym 32 with keysym 32 (keycode b).
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Overwriting existing binding of keysym 34 with keysym 34 (keycode d).
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Overwriting existing binding of keysym 35 with keysym 35 (keycode e).
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Overwriting existing binding of keysym 36 with keysym 36 (keycode f).
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Overwriting existing binding of keysym 33 with keysym 33 (keycode c).
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Overwriting existing binding of keysym 39 with keysym 39 (keycode 12).
лип 04 23:04:54 denys-Legion-5-15ACH6A gnome-system-mo[81374]: gtk_widget_event: assertion 'WIDGET_REALIZED_FOR_EVENT (widget, event)' failed
лип 04 23:05:00 denys-Legion-5-15ACH6A google-chrome.desktop[80892]: [80932:80939:0704/230500.951511:ERROR:ssl_client_socket_impl.cc(978)] handshake failed; returned -1, SSL error code 1, net_error -3
лип 04 23:05:00 denys-Legion-5-15ACH6A systemd[1092]: app-gnome-google\x2dchrome-80886.scope: Consumed 17.072s CPU time.
лип 04 23:05:02 denys-Legion-5-15ACH6A kernel: GpuWatchdog[72412]: segfault at 0 ip 00007fd3f5d929a6 sp 00007fd3ea9fd370 error 6 in libcef.so[7fd3f18ef000+7770000] likely on CPU 14 (core 7, socket 0)
лип 04 23:05:02 denys-Legion-5-15ACH6A kernel: Code: 89 de e8 0d ef 6e ff 80 7d cf 00 79 09 48 8b 7d b8 e8 4e 66 2c 03 41 8b 84 24 e0 00 00 00 89 45 b8 48 8d 7d b8 e8 5a d3 b5 fb <c7> 04 25 00 00 00 00 37 13 00 00 48 83 c4 38 5b 41 5c 41 5d 41 5e
лип 04 23:05:15 denys-Legion-5-15ACH6A gnome-shell[1292]: Window manager warning: Ping serial 36655684 was reused for window 0x4000003, previous use was for window W452.
лип 04 23:05:16 denys-Legion-5-15ACH6A gnome-system-monitor.desktop[81374]: glibtop(c=81374): [WARNING] statvfs '/run/user/1000/doc' failed: Operation not permitted
Pharmasolin commented 1 year ago

ChatGPT:

The line GpuWatchdog[72412]: segfault at 0 ip 00007fd3f5d929a6 sp 00007fd3ea9fd370 error 6 in libcef.so[7fd3f18ef000+7770000] indicates that there was a segmentation fault (segfault) within the libcef.so library. This library is part of the Chromium Embedded Framework (CEF), which is used by many applications to embed web content.

In this case, the log mentions "GpuWatchdog", which is a process that watches for GPU (Graphics Processing Unit) tasks that are taking too long. If it detects a long-running GPU task, it terminates it to prevent the entire system from becoming unresponsive.

The segfault means that the GpuWatchdog process tried to access memory that it wasn't allowed to, which is often a sign of a software bug. It could be a bug in the GPU drivers, in the Chromium Embedded Framework, or in the way the application is using them.

If you're seeing this error repeatedly, it might be worth trying a few things:

Updating your system: Make sure your system is fully updated. This includes the operating system, the application, and any drivers. If it's a known bug, there's a chance it's been fixed in a newer version.

Switching to a different GPU driver: If you're using proprietary GPU drivers, you could try switching to the open-source ones, or vice versa.

Disabling hardware acceleration: Some applications offer an option to disable hardware acceleration, which can sometimes work around issues like this.

Reporting the bug: If none of the above helps, consider reporting the bug to the developers of the application or the GPU driver. Make sure to provide them with as much information as possible, including the full error message and the circumstances under which it occurs.

melynx commented 1 year ago

I'm actually having the same issue here. Running Ubuntu 23.04 with a thunderbolt 3 eGPU enclosure with a Radeon Vega 64. Running Chrome using "integrated graphics" with steam running seem to cause the random crash + lockup as described. However, it seems like by running Chrome using the "dedicated gpu" (DRI PRIME offloading) seems to not result in the random crash, at least for the limited time I'm doing it. (There is a bug that causes Steam to crash loop if it is run using PRIME)

calindan2013 commented 1 year ago

same issue here, Ryzen 5900X + Radeon 6900XT, Debian 12.1, GNOME 43.6, running Chrome + Steam in background and other apps, I suddenly get a hard freeze of the desktop, keyboard becomes unresponsive and mouse lights go off, no way of switching to a terminal or doing anything except keeping power button pressed to shut down. Tried all suggested option in the BIOS for hard freezes involving AMD processors, but keeps happening every week:

root@localhost:~# tac /var/log/syslog | grep "libcef" 2023-07-25T01:17:45.406198+01:00 localhost kernel: [14382.710339] GpuWatchdog[14867]: segfault at 0 ip 00007f87bf1929a6 sp 00007f87b62a68f0 error 6 in libcef.so[7f87bacef000+7770000] likely on CPU 8 (core 10, socket 0) 2023-07-24T21:22:45.254209+01:00 localhost kernel: [ 282.915405] GpuWatchdog[11950]: segfault at 0 ip 00007f87bf1929a6 sp 00007f87b62a68f0 error 6 in libcef.so[7f87bacef000+7770000] likely on CPU 13 (core 1, socket 0) 2023-07-24T11:16:43.093311+01:00 localhost kernel: [ 1088.729831] GpuWatchdog[15198]: segfault at 0 ip 00007f7330f929a6 sp 00007f732804e8f0 error 6 in libcef.so[7f732caef000+7770000] likely on CPU 15 (core 3, socket 0) 2023-07-23T10:25:17.987090+01:00 localhost kernel: [ 523.786541] GpuWatchdog[11285]: segfault at 0 ip 00007fae835929a6 sp 00007fae7a5508f0 error 6 in libcef.so[7fae7f0ef000+7770000] likely on CPU 1 (core 1, socket 0) 2023-07-23T01:32:17.746643+01:00 localhost kernel: [ 9338.927904] GpuWatchdog[351214]: segfault at 0 ip 00007f24a11929a6 sp 00007f24982248f0 error 6 in libcef.so[7f249ccef000+7770000] likely on CPU 21 (core 11, socket 0)

Pharmasolin commented 1 year ago

@calindan2013 fyi, disabling hardware acceleration in steam settings fixes this issue.

calindan2013 commented 1 year ago

yes that worked thanks. still, this situation should not be allowed to degenerate into a complete system freeze by the kernel and/or its video hardware modules.

Pharmasolin commented 1 year ago

yes that worked thanks. still, this situation should not be allowed to degenerate into a complete system freeze by the kernel and/or its video hardware modules.

I don't know how to report this bug up in a chain and whose bug it is. CEF, mesa (radeonsi(?)), or gnome (mutter?)

user1-github commented 1 year ago

After 3 days of headache and lots of googling, trying to figure out why does this happen to me on Debian 12.1 while Steam Flatpak is downloading games, I've come to the conclusion that this is caused by yet another rare bug in Chromium (which is unfortunately full of those). Either that, or the way the specific app utilizes Chromium (For those who don't know, Steam client embeds Chromium for webpages and web uis). I mean I've seen people had this exact same GpuWatchdog segfault with electron and other Chromium apps as well and in those cases it also caused a full gui freeze. This also seem to happen on any gpu vendor / driver on Linux, both open source and proprietary and on any DE, with both X and Wayland. So the solution to this is either Chromium devs should fix this, or maybe Steam can have some workaround to prevent this issue.

Since I didn't experience this before the big Steam ui update, I think it might have something to do with the fact that the Steam client now makes heavier use of web uis.

user1-github commented 1 year ago

Hello @Pharmasolin, in general, no OpenGL application including Steam should be capable of taking down the whole system and symptoms like this usually hint towards a video driver issue. It might be worthwhile to also mention this issue to your video driver vendor.

@kisak-valve But they can. See my comment above. The root cause of the issue is most likely a Chromium bug. The fact that it takes the whole gui down is a flaw in the Linux gui stack (this is one of the few areas where Windows is ahead of Linux, because just like this issue, incorrect graphics api usage on Windows also doesn't take the Windows gui down ((unless it's a bug in the kernel driver))).

zerocrates commented 1 year ago

I've been having this issue since the new UI also (Ubuntu 23.04, on Intel integrated graphics).

I'm using Wayland, so there's a wrinkle: this doesn't take down the whole system UI, just X programs. This means that the problem is recoverable: killing Xwayland (it has to be a kill -9, the regular TERM signal won't do anything) will kill Steam and anything else using X that was running at the time, but they can then just be restarted.

My main workaround has been to have the actual Steam UI open less. The hang will happen quite often if Steam is just sitting in the background all the time, but more or less will never happen if Steam's actual windows are closed and it's just a notification icon. But the risk remains that the moment you do go to briefly interact with it will cause the hang.

dsalt commented 11 months ago

The same crash occasionally happens here. There's no desktop hang, though. (Linux 6.5.9, Mesa 23.2.1, RX 6600 XT; Devuan daedalus.)

lologarithm commented 10 months ago

I am having the same issue, I am unsure if its from steam, discord, spotify, or any number of other apps that might use chromium.

hardware: amd 5950x, radeon 6900XT, versions: Linux 6.3.13, kisak-mesa 23.2.1

example log:

[30705.371465] GpuWatchdog[646929]: segfault at 0 ip 00007f88ba792bc6 sp 00007f88af5fd430 error 6 in libcef.so[7f88b62ef000+7770000] likely on CPU 13 (core 13, socket 0)

Willing to do any debugging or grabbing of logs if it helps someone.

KenSharp commented 10 months ago

This affects more than just Steam. Users of the Electron framework could help by working together.

https://bugs.launchpad.net/ubuntu/+source/linux-signed-hwe-6.2/+bug/2045951