fedora-workstation

#481 System hard freezes after upgrading to kernel 6.14.8 on Fedora 42 with AMD RDNA3 GPU.
Closed: Fixed by catanzaro. Opened by linuxkernel94.

Closed: Fixed
Reopen Issue

🔧 Issue Summary:
System freezes completely (requires hard reset) with both amdgpu.gfxoff=1 and gfxoff=0.

âť— Symptoms:
Total system lockup — SysRq and REISUB do not work.

Frequent when using hardware-accelerated apps (e.g., Brave, FreeTube).

Also occurs during normal desktop use (GNOME Wayland).

Journal logs show GPU-related errors, such as:

amdgpu: ring sdma timeout

amdgpu: SMU: I'm not done with your previous command

watchdog: hard LOCKUP on CPU

đź’» Hardware:
GPU: AMD Radeon™ RX 6600M

CPU: AMD Ryzen™ 7 5800H

OS: Fedora 42 Silverblue

Working Kernel: 6.14.6-300.fc42.x86_64

Problematic Kernel: 6.14.8-300.fc42.x86_64

đź“‹ Steps to Reproduce:
Boot into Fedora Silverblue 42 with kernel 6.14.8.

Launch any application using VA-API or hardware video decoding (e.g., Brave, FreeTube).

Wait a few minutes.

System hard-freezes with no way to recover.

🧪 Tests Performed:
Issue reproduced after a clean boot with no custom kernel modules.

Rolling back to kernel 6.14.6 fully resolves the issue.

Setting amdgpu.gfxoff=0 does not prevent the lockups.

Issue started immediately after updating to kernel 6.14.8.

âś… Temporary Workaround:
Roll back to kernel 6.14.6 using ostree rollback.

Pin this deployment and set it as the default with rpm-ostree set-default.

Hi, this is a policy issue tracker. It's not a good place to report software bugs. This really needs to be reported to the Linux kernel developers.

Metadata Update from @catanzaro:
- Issue close_status updated to: Won't fix
- Issue status updated to: Closed (was: Open)

Unfortunately the best upstream guidance I found for reporting kernel bugs is this document which is extremely long and complicated and notably advises users to report bugs to downstream rather than to upstream. You could theoretically report this against the kernel component on Red Hat Bugzilla, but I don't think the problem would be very likely to be noticed there. I would report to upstream if possible.

Metadata Update from @catanzaro:
- Issue status updated to: Open (was: Closed)

Thanks, Michael. It looks like a lot of users are affected—likely thousands—since AMD is the most widely used GPU on Linux, and this issue impacts all architectures from the 6000 series and up.

The system consistently freezes after just a few minutes of use, and I want to emphasize consistently.

This is a critical issue. Would it be possible to revert the update or take action to address it?

https://discussion.fedoraproject.org/t/amdgpu-driver-crash/154834/5

You could theoretically report this against the kernel component on Red Hat Bugzilla, but I don't think the problem would be very likely to be noticed there. I would report to upstream if possible.

Reporting to RHBZ is valuable, as the kernel maintainers do pay attention to it.

Hey, thanks Neal — you and Michael are absolute heroes. I just found out that the issue might actually be related to Mutter, though I’m not 100% certain yet.

Take a look here: https://bugzilla.redhat.com/show_bug.cgi?id=2369567

Yeah, looks like this turned out to be a problem in mutter.

Also looks like I reopened this by mistake.

Metadata Update from @catanzaro:
- Issue close_status updated to: Fixed
- Issue status updated to: Closed (was: Open)

This issue has been migrated to Fedora Forge:
https://forge.fedoraproject.org/workstation/tickets/issues/481

Please continue any further discussion there.
Metadata
None
None
None
None
None
Static archive. pagure.io was sunset in 2026; this is a read-only snapshot.
Red Hat Logo

Fedora is sponsored by Red Hat. Learn more about the relationship between Red Hat and Fedora »

© Red Hat, Inc. and others.