The 6.12.13 stable kernel update contains a number of important fixes across the tree.
The 6.12.12 stable kernel update contains a number of important fixes across the tree.
Updates may require up to 24 hours to propagate to mirrors. If the following command doesn't work, please retry later:
sudo dnf upgrade --refresh --advisory=FEDORA-2025-573a085a7c
Please login to add feedback.
0 | 12 | Test Case kernel regression |
This update has been submitted for testing by jforbes.
This update's test gating status has been changed to 'waiting'.
This update's test gating status has been changed to 'waiting'.
This update has obsoleted kernel-6.12.12-200.fc41, and has inherited its bugs and notes.
no kernel traces like 6.12.12 but security lockdown is still disabled
boot kernel with kernel args 'lockdown=integrity'
This update's test gating status has been changed to 'passed'.
default & performance test pass
Working fine here.
The lockdown-related warnings at boot that were present in 6.12.12 are gone, as someone mentioned earlier.
Some of the power-state changes have made it more difficult to disable the NMI watchdog via /proc/sys/kernel/nmi_watchdog just after booting, as some users do to save power. I had been relying on tlp to do this, but the setting does not persist now.
A related note on the upcoming 6.13.x series -- jforbes built kernel-6.13.2-200.fc41 in https://koji.fedoraproject.org/koji/buildinfo?buildID=2653859 at about the same time as this build. Despite some of the general power-saving AMD graphics-related bugfixes in the latest build, kernel-6.13.x, up to and including kernel-6.13.2-200.fc41, seems not to be able to properly configure my Bonaire (GFX7, VCN 2nd Gen., Sea Islands) dGPU when using the amdgpu driver -- I'm seeing the following error messages at boot, with radeon.cik_support=0 amdgpu.cik_support=1:
... [drm:amdgpu_device_ip_set_powergating_state [amdgpu]] ERROR set_powergating_state of IP block <vce_v2_0> failed -22 ... amdgpu 0000:01:00.0: amdgpu: failed to send message 146 ret is 0
And there are large regressions in performance:
env AMD_DEBUG=useaco glmark2 -s 2560x1440:
env __GLX_VENDOR_LIBRARY_NAME=mesa MESA_LOADER_DRIVER_OVERRIDE=zink GALLIUM_DRIVER=zink glmark2 -s 2560x1440:
glmark2 -s 2560x1440:
env AMD_DEBUG=useaco glmark2-es2 -s 2560x1440:
env __GLX_VENDOR_LIBRARY_NAME=mesa MESA_LOADER_DRIVER_OVERRIDE=zink GALLIUM_DRIVER=zink glmark2-es2 -s 2560x1440:
glmark2-es2 -s 2560x1440:
env AMD_DEBUG=useaco glmark2-es2-wayland -s 2560x1440:
env __GLX_VENDOR_LIBRARY_NAME=mesa MESA_LOADER_DRIVER_OVERRIDE=zink GALLIUM_DRIVER=zink glmark2-es2-wayland -s 2560x1440:
glmark2-es2-wayland -s 2560x1440:
env AMD_DEBUG=useaco glmark2-wayland -s 2560x1440:
env __GLX_VENDOR_LIBRARY_NAME=mesa MESA_LOADER_DRIVER_OVERRIDE=zink GALLIUM_DRIVER=zink glmark2-wayland -s 2560x1440:
env glmark2-wayland -s 2560x1440:
As the above numbers suggest, these regressions are not present in 6.12.x and earlier, up to and including kernel-6.12.13-200.fc41 (even though this shares some of the same recent power-saving and AMD graphics-related bugfixes as 6.13.2); nor are they present in the radeonsi driver. Unfortunately, the latter driver is less capable than amdgpu, so this is a considerable loss. I don't have other AMD GPUs to test if this is happening on other hardware. Please let me know if you'd like me to provide additional feedback elsewhere.
Default (libhugetlbfs, memfd & paxtest skipped) and performance tests pass; works in general.
This update has been pushed to testing.
This update can be pushed to stable now if the maintainer wishes
works really good, with nvidia(hdmi).
Works.
Works fine on Intel 12th gen, AMD 7900XTX. Passed regression tests. Not seeing the logging issues present in 6.12.12
This update's test gating status has been changed to 'waiting'.
This update's test gating status has been changed to 'passed'.
Framework 16 has a crash on the amdgpu driver:
amdgpu 0000:c5:00.0: [drm] REG_WAIT timeout 1us * 100 tries - dcn31_program_compbuf_size line:141
The 'good' news is that the system recovers from that crash and is working properly (afaik).
Ryzen 5600H, kernel test pass.
Working fine on baremetal
kernel tests PASSED
Suspension works
Works
AMD Ryzen 5. Both tests did PASS on a second test to confirm a fail I had on the the first run. On the first attempt ./default/cachedrop FAILED. Performance was a PASS.
Again for confirmation on the second attempt both tests did PASS.
no regressions noted
Looks good here; Framework 13 AMD 7840U, GNOME/Wayland, LUKS/btrfs.
Looks better than 6.12.12. Ryzen 3700x, Rog Strix B550-F, Nvidia from rpmfusion. No more tracer reports at the top of dmesg.
test suite run
no issues with Thinkpad P1 Gen4
No issues on Lenovo Legion Slim 5 Gen 9 AMD 8845HS and Nvidia 4070 proprietary drivers.
No regression or issues noted on my T470s. Passed kernel regression test.
This update has been submitted for stable by jforbes.
This update has been pushed to stable.
lenovo p16s amd gen2 : all working fine. kernel tests ran and uploaded. i am using amdgpu (i have a 7840U + 780M) but i am not having any issues with the 780M with this kernel.
i did run the same test as you on my 780M. but since this is the first time i do it, i dont really know what it means.
glmark2 Score: 4273
full paste here : https://pastebin.com/RAhgtYZV