Bug 515751

SUMMARY

Severe Bluetooth HID input jitter regression affecting game controllers in KDE Plasma 6.5.4 and newer.
The issue makes gaming unplayable, especially for competitive or precision gameplay, and affects real users across multiple Bluetooth adapters and controllers.

High polling rate controllers over Bluetooth exhibit significantly worse jitter, while lower polling rate controllers are less affected.
This behavior did not occur in Plasma 6.5.3, where Bluetooth controller input was stable and usable.

The problem is present on Wayland and X11, persists across updates (6.5.4 → 6.5.5), and is not related to GPU drivers, power management, or graphics stack.

USB connection does not show jitter, confirming the issue is specific to Bluetooth input handling.

STEPS TO REPRODUCE

Boot into KDE Plasma 6.5.4 or newer (Wayland or X11).

Connect a game controller via Bluetooth.

Open a game, gamepad tester, or any application requiring precise analog or button input.

Move analog sticks or perform rapid button inputs.

OBSERVED RESULT

Noticeable input jitter, stutter, and inconsistent timing on Bluetooth-connected controllers.

The issue is more severe on high polling rate controllers.

Gameplay feels unstable and unreliable.

Competitive gaming and precision input are effectively unplayable.

The problem occurs:

On Wayland and X11

With NVIDIA and safe graphics mode

With multiple Bluetooth adapters

USB connection works correctly with no jitter.

EXPECTED RESULT

Bluetooth controller input should be stable, smooth, and consistent, matching behavior seen in Plasma 6.5.3.

High polling rate controllers should not suffer excessive jitter.

Bluetooth and USB input latency/jitter should be comparable, as they were before the regression.

SOFTWARE/OS VERSIONS

Linux/KDE Plasma: KDE neon (Ubuntu 24.04 base)

KDE Plasma Version: 6.5.4 / 6.5.5

KDE Frameworks Version: 6.22.x

Qt Version: 6.10.1

Kernel: 6.x (issue reproduced across kernels)

BLUETOOTH ADAPTERS TESTED

Intel Dual Band Wireless-AC 7260
→ High jitter (severe)

Realtek RTL8771B Bluetooth Radio
→ Jitter present

TP-Link Bluetooth 5.3 UB5A
→ Jitter reduced but still present
→ Low/normal polling rate controllers mostly usable
→ High polling rate controllers still show high jitter

CONTROLLERS TESTED

Sony DualShock 4 (original)

DualShock 4 Data Frog P03

Wolf V-6 controller

All affected over Bluetooth.
All work correctly over USB.

ADDITIONAL INFORMATION

This is a regression: the issue was not present in Plasma 6.5.3.

Occurs regardless of compositor (Wayland/X11).

Not resolved by:

Power management tweaks

CPU governor changes

Disabling C-states

NVIDIA driver updates

Safe graphics mode

USB input path is unaffected, confirming Bluetooth-specific behavior.

The issue impacts real users, especially those playing competitive or latency-sensitive games.

TECHNICAL HYPOTHESIS (Bluetooth / HID Timing)

Based on testing across multiple controllers, adapters, and Plasma versions, the issue appears to be related to Bluetooth HID event timing and scheduling, rather than controller firmware, hardware, or graphics stack.

Key observations:

The jitter only occurs when controllers are connected via Bluetooth.

The same controllers exhibit no jitter over USB.

The issue is worse on high polling rate controllers, suggesting:

Input events may be batched, delayed, or unevenly scheduled

Event timestamps may be handled inconsistently

The problem affects multiple Bluetooth adapters and vendors, making adapter-specific firmware issues unlikely.

The regression coincides with Plasma 6.5.4, and was not present in 6.5.3.

Possible areas involved:

BlueZ HID input event timing or buffering

Bluetooth HID interrupt/event scheduling

KDE / KWin / libinput input processing changes affecting Bluetooth devices

Event coalescing, throttling, or resampling introduced in Plasma 6.5.4

Higher-frequency HID reports being mishandled compared to lower polling rate devices

Additional hints:

Low or normal polling rate controllers are less affected, while high polling rate controllers exhibit severe jitter.

This suggests that higher report frequency may exceed or trigger a timing edge case in the Bluetooth → input stack.

Both Wayland and X11 are affected, pointing away from compositor-specific rendering issues and toward shared input infrastructure.

While this report does not identify the exact faulty component, the behavior strongly suggests a regression in Bluetooth HID input timing introduced after Plasma 6.5.3.