Summary: |
Daemon can hang after [null-sink] asyncq.c: q overrun, queuing locally |
Product: |
PulseAudio
|
Reporter: |
Marcin Lewandowski <marcin> |
Component: |
core | Assignee: |
pulseaudio-bugs |
Status: |
RESOLVED
MOVED
|
QA Contact: |
pulseaudio-bugs |
Severity: |
major
|
|
|
Priority: |
medium
|
CC: |
lennart
|
Version: |
unspecified | |
|
Hardware: |
Other | |
|
OS: |
All | |
|
Whiteboard: |
|
i915 platform:
|
|
i915 features:
|
|
Attachments: |
gdb dump
|
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.
Created attachment 124804 [details] gdb dump Version 9.0. I have connected many (+/- 100) GStreamer based clients over native protocol with disabled SHM and realtime scheduling. Each client that has GStreamer's pulsesink loads module-null-sink and sends data to this sink. Each client that has GStreamer's pulsesrc loads module-null-sink and fetches data from this sink's monitor. Sporadically I load loopback module to "connect" these null sinks and their monitors. It works fine for a while and then loads of [null-sink] asyncq.c: q overrun, queuing locally appear in the syslog. All operations of PA are stopped. It's not desired behaviour, even if there's something wrong with some of the clients it should not cause whole daemon to hang. Stack trace is attached.