mirror of
https://github.com/mumble-voip/mumble.git
synced 2026-03-03 00:46:56 -05:00
Mumble SIGKILL by rtkit on newer versions of pipewire #2982
Labels
No labels
GlobalShortcuts
Hacktoberfest
accessibility
acl
asio
audio
bonjour
bsd
bug
build
certificate
ci
client
code
documentation
external-bug
feature-request
gRPC
github
good first issue
help wanted
help-needed
ice
installer
linux
macOS
needs-ckeck-with-latest-version
needs-more-input
overlay
positional audio
priority/P0 - Blocker
priority/P1 - Critical
priority/P2 - Important
priority/P3 - Somewhat important
priority/P4 - Low
public-server-registration
qt
recording
release-management
server
stale-no-response
stale-support
support
task
test
theme
translation
triage
ui
windows
wontfix
x64
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
starred/mumble-mumble-voip#2982
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @eyJhb on GitHub (Apr 12, 2025).
Description
When starting Mumble when using PipeWire, it crashes immediately (shows window for a brief second, then crashes).
Steps to reproduce
Mumble version
1.5.735
Mumble component
Client
OS
Linux
Reproducible?
Yes
Additional information
I've narrowed it down to being a PipeWire update. It worked fine on PipeWire 1.2.7, when upgrading to 1.4.0 it stopped working. I've bisected PipeWire, and it seems like it's this commit here in client.conf.in
gitlab.freedesktop.org/pipewire/pipewire/-@24bcacc619 (ddf944f7b061f49f98c4de932a59d8238e538c2d), causing the issue. I'm unsure what needs to be changed in Mumble, even though I did try :)Including git bisect of pipewire just for completeness.
Relevant log output
Screenshots
No response
@davidebeatrici commented on GitHub (Apr 12, 2025):
You should ask your distribution package's maintainer to update PipeWire to 0.4.1, for reference: https://gitlab.freedesktop.org/pipewire/pipewire/-/issues/4595
@eyJhb commented on GitHub (Apr 13, 2025):
@davidebeatrici this also happens on 1.4.1.
@eyJhb commented on GitHub (Apr 13, 2025):
At least this is what I'm running.
Debug log with
PIPEWIRE_DEBUG=5: pipewire-debug.logIf I comment out these lines in Pipewire, I no longer get the crash https://gitlab.freedesktop.org/pipewire/pipewire/-/blob/master/src/daemon/client.conf.in?ref_type=heads#L45-53
Is Mumble confirmed to be working with 1.4.1 of PipeWire for both audio in/out? If so, then it's more likely a packaging issue, and I'll try to look elsewhere.
@eyJhb commented on GitHub (Apr 13, 2025):
Okay, so I did some more digging, and I found out the issue (I think). So, the issue linked here https://gitlab.freedesktop.org/pipewire/pipewire/-/issues/4595 has nothing to do with it since it's a segfault, whereas I'm getting SIGKILL.
The issue seems to be, that I'm running rtkit, and that specific commit I linked from PipeWire earlier enables
libpipewire-module-rtby default.The theory is then, that Mumble blocks for too long the rt thread, and is then SIGKILL by rtkit. Which is why I see mumble for a split second, before it's killed.
I'm not sure what the solution would be for Mumble, or if the solution is just to NOT run rtkit. But I don't think I'm the only one, that uses rtkit sadly, and more might experience this in the future.
Also, thanks for reopening the issues! Highly appreciated :)
EDIT: I should clarify, rtkit enforces/sets a RLIMIT_RTTIME, which is what ends up killing it AFAIK.
@davidebeatrici commented on GitHub (Apr 13, 2025):
Thank you for the detailed investigation and sorry for my misunderstanding.
Could you try libcrossaudio's TestLoopback application? It would help us rule out a few potential causes.
https://github.com/mumble-voip/libcrossaudio
Please note that ALSA is the default backend on Linux, you can switch to PipeWire by replacing
CROSSAUDIO_BACKEND_ALSAwithCROSSAUDIO_BACKEND_PIPEWIREinsrc/tests/Common.h.@eyJhb commented on GitHub (Apr 14, 2025):
I've tried running it now, and it doesn't seem to be killed/crash.
Has been running for a good 30s now. Should I observe anything else, besides it just hanging around as a application in pavucontrol?
@davidebeatrici commented on GitHub (Apr 14, 2025):
The test should actually play back the input audio, but in this case we just wanted to know whether it exhibited the same behavior as Mumble.
@eyJhb commented on GitHub (Apr 16, 2025):
Confirmed to playback input audio now, and still not killed by rt :) - Anything else I can do?
@davidebeatrici commented on GitHub (Apr 17, 2025):
No, that suffices, thank you! I'll take care of fixing the issue.
@eyJhb commented on GitHub (Apr 18, 2025):
Perfect! Thank you for looking into it :) - Feel free to aggressively ping me, if you need me to test anything. I don't mind at all :)
@nh2 commented on GitHub (Sep 16, 2025):
Still happening to me with pipewire
1.4.7and Mumble1.5.735on NixOS 25.05.@nh2 commented on GitHub (Sep 16, 2025):
Workaround
@ju1m commented on GitHub (Feb 19, 2026):
Mumble still crashes at startup for me with
libpipewire-1.4.9andmumble-1.5.857onnixos-25.11.nh2's
DISABLE_RTKIT=1workaround still works.