Repository navigation
Temporary delays occur at regular intervals for all inputs on Linux. #38886
TypeOther Bug descriptionWhile osu!lazer is running, all input occasionally becomes delayed for a brief moment. This affects both keyboard and mouse input simultaneously. The issue occurs approximately once every few seconds and appears to happen at a very regular interval. It occurs consistently regardless of where in osu!lazer I am, including menus and gameplay, and is not limited to any specific game mode, beatmap, or screen. The issue started around the end of August 2026 after an update. My current environment is CachyOS with KDE Plasma on Wayland. I have only confirmed the issue in this environment. Screenshots or videos2026-09-21.19-22-45.mp4Version2026.921.0-lazer Logs |
Replies: 4 comments 1 reply
|
I can reproduce this on my system, and I have bisected this to ppy/osu-framework@2a44111 |
|
Following the above info: libsdl-org/SDL@18fc4d9 is the offending commit |
|
A final update - the offending commit in the SDL USB handling code slightly alters the checks performed when a device is connected. What this means for us: in order for the "bad" SDL code to run, a device must be getting plugged in. In my case, I checked I'd still consider this a regression in SDL since there's probably a blocking check in that new code which shouldn't be there, but the bad behavior requires a faulty system (unless there exists some legitimate USB device which constantly re-initializes itself). @MIYUKINNGU, check |
|
After some A/B testing on the libsdl code, I've opened an issue there: libsdl-org/SDL#16383 |
A final update - the offending commit in the SDL USB handling code slightly alters the checks performed when a device is connected.
What this means for us: in order for the "bad" SDL code to run, a device must be getting plugged in. In my case, I checked
dmesg --followand discovered that my motherboard's bluetooth hardware was constantly crashing and re-initializing, thus triggering the code. Once i disabled the bluetooth module the hitching stopped in Osu.I'd still consider this a regression in SDL since there's probably a blocking check in that new code which shouldn't be there, but the bad behavior requires a faulty system (unless there exists some legitimate USB device which constan…