On Tue May 19, 2026 at 09:58:33PM +0200, Vona Péter wrote:
> To: [email protected]
> Subject: inteldrm: GuC/HuC firmware crash and kernel hang during
> hardware video decoding on 7.9-current
> 
> Category: system
> System: OpenBSD 7.9-current (snapshot)
> Description:
>     While running GNOME desktop environment on 7.9-current (snapshot),
> severe 
>     kernel hangs occur during concurrent high network I/O (qBittorrent)
> and 
>     hardware-accelerated video playback (YouTube in Firefox). 
>     
>     The network stack completely stalls for a few minutes, after which
> the system 
>     recovers. The dmesg reveals a severe Intel DRM GuC/HuC firmware
> authentication 
>     failure and Command Transport (CT) timeout. This behavior was not
> present 
>     on OpenBSD 7.8.
> 
> How To Reproduce:
>     1. Boot into GNOME on 7.9-current snapshot.
>     2. Start a high-bandwidth download in qBittorrent.
>     3. Open Firefox and play a YouTube video using hardware
> acceleration.
>     4. After some time, the entire networking stack stalls, and the DRM
> errors 
>        appear in dmesg.
> 
> Relevant dmesg snippet:
>     drm:pid0:ct_send *ERROR* [drm] *ERROR* GT0: GUC: CT: No response
> for request 0x4000 (fence 1)
>     drm:pid0:intel_guc_ct_send *ERROR* [drm] *ERROR* GT0: GUC: CT:
> Sending action 0x4000 failed (0xffffffffffffffc4e) status=0
>     drm:pid0:intel_huc_auth *ERROR* [drm] *ERROR* GT0: HuC: all
> workloads authentication failed 0xffffffffffffffc4e
>     drm:pid73523:ct_handle_response *ERROR* [drm] *ERROR* GT0: GUC: CT:
> Unsolicited response message: len 1, data 0xf0000000 (fence 1, last 1)
>     drm:pid73523:ct_handle_hxg *ERROR* [drm] *ERROR* GT0: GUC: CT:
> Failed to handle HXG message (0xfffffffffffffffee) 0xffff800001b307d8h
>     drm:pid73523:ct_handle_msg *ERROR* [drm] *ERROR* GT0: GUC: CT:
> Failed to process CT message (0xfffffffffffffffee) 0xffff800001b307d4h
> 
> Fix/Workaround:
>     Disabling hardware acceleration in Firefox (gfx.webrender.software
> = true) 
>     prevents the firmware crash and the subsequent network/kernel
> stall.
> 
> Full dmesg and sysctl hw.uuid output below.
> 
> [email protected] ~ $ sysctl hw.uuid
> hw.uuid=9c50d392-8c98-4a9c-8915-efe1f25f5b45
> [email protected] ~ $ 
> 
> 

I've noticed the same thing on all my desktop systems (AMD/Intel). In my
case, it's not necessarily "concurrent high network I/O" but rather a
concurrent ports build, meaning a lot of disk activity.

I've had this for a very long time, so it has nothing to do with version
7.9.

The problem I'm having is that the system freezes without crashing or
displaying a kernel message. (That's why I've never reported anything
before)

>     Disabling hardware acceleration in Firefox (gfx.webrender.software
> = true) 

I'll try this.

Reply via email to