New: [REGRESSION 6.12→6.18] ASUS N-KEY keyboard (0b05:19b6) USB resets ~1/sec under dGPU rendering load on Raphael xHCI
https://bugzilla.kernel.org/show_bug.cgi?id=221778
Bug ID: 221778
Summary: [REGRESSION 6.12→6.18] ASUS N-KEY keyboard (0b05:19b6)
USB resets ~1/sec under dGPU rendering load on Raphael
xHCI
Product: Drivers
Version: 2.5
Hardware: AMD
OS: Linux
Status: NEW
Severity: high
Priority: P3
Component: USB
Assignee: [email protected]
Reporter: [email protected]
Regression: NoHardware: ASUS ROG Strix G16 G614PM (AMD variant: Raphael iGPU + RTX 5060 Laptop dGPU), BIOS G614PM.316 (latest from ASUS).
Internal keyboard: ASUS "N-KEY Device", USB 0b05:19b6, hid-asus driver, attached to usb 3-1 on the AMD Raphael xHCI controller (1022:15b6/15b7).
Summary: since 6.18, rendering load on the Nvidia dGPU causes the internal keyboard's USB device to reset ~once per second ("usb 3-1: reset full-speed USB device" in dmesg — 4,557 resets in one 8h boot). The resets corrupt input below userspace: evtest on /dev/input/event3 shows missed key presses, releases arriving seconds late, and missing autorepeat. In games this manifests as stuck/dropped WASD keys.
Trigger isolation:
- iGPU rendering with dGPU powered but idle: CLEAN (tested via supergfxctl Integrated mode).
- dGPU rendering by application: resets begin immediately (70 resets in the first ~2 min after game launch).
Kernels tested:
- 6.12.75 (Arch linux-lts): CLEAN — 0 resets in a full game session.
- 6.18.38 (cachyos-lts): BROKEN.
- 7.1.4 (cachyos): BROKEN (retested 2026-07-22).
Ruled out: EC reset (no effect), BIOS (already latest), hardware (machine is clean on 6.12).
Distro: CachyOS (Arch-based, rolling). Nvidia driver 610.43.03 via nvidia-open-dkms.
Multiple users with G614/G615 models report identical symptoms on the same 0b05:19b6 device:
-
https://discuss.cachyos.org/t/keyboard-input-issues-on-asus-rog-strix-16-2025-with-cachyos-during-gaming/30823
- https://github.com/ublue-os/bazzite/issues/4590
I will be away from this machine until ~2026-07-24; I can respond to questions and run tests after that. Other affected users are active in the CachyOS thread linked above and may be able to test sooner.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
[Bug 221778] [REGRESSION 6.12→6.18] ASUS N-KEY keyboard (0b05:19b6) USB resets ~1/sec under dGPU rendering load on Raphael xHCI
https://bugzilla.kernel.org/show_bug.cgi?id=221778
--- Comment #1 from Miguel Cervantes ([email protected]) ---
A note from me, I did write this with AI I'm not sure about the flow of things yet in the linux space, I have to admit some of these processes are confusing for me, I just want the information at the very least out there in an useful place, I apologize in advance if there are obvious errors/things I'm omitting.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
[Bug 221778] [REGRESSION 6.12→6.18] ASUS N-KEY keyboard (0b05:19b6) USB resets ~1/sec under dGPU rendering load on Raphael xHCI
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=221778
Michał Pecio ([email protected]) changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |[email protected]
--- Comment #2 from Michał Pecio ([email protected]) --- 6.12 to 6.18 is a fairly large spread, and it isn't even clear if the bug is a regression or perhaps some CachyOS vs Arch difference. And the GPU uses an out of tree driver, hmm.
IDK about CachyOS, but Arch package archive has 'linux' (non -lts) packages for all major versions from 6.12 to 6.18. Could you start with testing these to find which release broke things? And try to use the same NVIDIA driver each time, or (preferably) see if the problem occurs with Nouveau.
I would wager that resets are deliberate (HID driver seems to do that in response to USB errors), but that still doesn't explain why USB errors occur due to GPU activity and why some kernels work but others don't.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
[Bug 221778] [REGRESSION 6.12→6.18] ASUS N-KEY keyboard (0b05:19b6) USB resets ~1/sec under dGPU rendering load on Raphael xHCI
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=221778
--- Comment #3 from Miguel Cervantes ([email protected]) --- (In reply to Michał Pecio from comment #2)
Quoted 4 lines
6.12 to 6.18 is a fairly large spread, and it isn't even clear if the bug is a regression or perhaps some CachyOS vs Arch difference. And the GPU uses an out of tree driver, hmm.
Quoted 8 lines
IDK about CachyOS, but Arch package archive has 'linux' (non -lts) packages for all major versions from 6.12 to 6.18. Could you start with testing these to find which release broke things? And try to use the same NVIDIA driver each time, or (preferably) see if the problem occurs with Nouveau.
I would wager that resets are deliberate (HID driver seems to do that in response to USB errors), but that still doesn't explain why USB errors occur due to GPU activity and why some kernels work but others don't.
So I installed Linux only about 2 weeks ago, I went from bazzite to cachy os and both experienced this issue, i couldn't recall bazzite history but cachy i went from 6.12 base download to 6.18 upgrade for a few days into the newest version to test the bug, this is why i posted it like that. I cannot otherwise confirm if its inbetween all versions at the moment.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
[Bug 221778] [REGRESSION 6.12→6.18] ASUS N-KEY keyboard (0b05:19b6) USB resets ~1/sec under dGPU rendering load on Raphael xHCI
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=221778
--- Comment #4 from Miguel Cervantes ([email protected]) --- Also I am in one of the packages you mentioned, hence me being in 6.12 at the moment and me posting the bug, I feel a bit stuck in here. The driver choice was me debugging for a while to figure out what was wrong and it's what i ended up with, apologies for the confusion.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.