Kernel regression tests returned a cachedrop fail on the first run (I've seen this before... seems not to be that rare) but everything passed on subsequent runs.
Also timer-overhead is failing in VirtualBox only, but that is common for VirtualBox for some reason.
@charn@markec given the apparent involvement of wpa_supplicant, if you have a moment could you please reply (preferably in the bugzilla issue I opened above so we don't spam more here) with the type of wireless adapter and network mode you're having suspend success with? For reference mine is an AMD/Mediatek MT7922 on a 5GHz, 80MHz-wide ax network.
I have a framework laptop with AMD CPU https://frame.work/fr/fr/products/laptop-diy-13-gen-amd and my system is unstable since I boot on 6.6.5. I lose Wifi access after a while wpa_supplicant / NetworkManager goes to 100% CPU, and process are killed by OOM (on a 32GB RAM laptop)
Regression test ./runtests.sh PASS with kernel-6.6.5-200.fc39.x86_64 within QEMU/KVM with CPU-passthrough on an AMD Ryzen 7 PRO 6850U host: F38 KDE Spin with SELinux confined user account (user_u), VM up to date as of now with updates-testing enabled, uname -r = 6.6.5-200.fc39.x86_64, tainted = 0. So far it works fine with average activities: firefox browsing, dolphin file organization.
BZ#2253756 seems to not affect my AMD hardware. In my cockpit KVM VM, I cannot wake up (which is intended and not a bug), but I can initiate sleep, and this works fine without the problem described in BZ#2253756. When checking the log after resetting the VM (resetting because the VM cannot wake up), the journalctl -r --boot=-1 shows a normal going to sleep that ends with as intended with "Entering sleep state 'suspend'" (in the VM, cat /sys/power/mem_sleep = [s2idle]). So the bug seems to not affect all AMD hardware.
Similar problem to what was reported above - Lenovo Legion 5 with AMD CPU and GPU.
Both NetworkManager and wpa_supplicant using 100% CPU, and unable to stop restart services
Observing erratic behaviour and high CPU usage
Requires hard shutdown / restart (using magic SysRq keys)
This update has been submitted for testing by acaringi.
This update's test gating status has been changed to 'waiting'.
Kernel regression tests returned a cachedrop fail on the first run (I've seen this before... seems not to be that rare) but everything passed on subsequent runs.
Also timer-overhead is failing in VirtualBox only, but that is common for VirtualBox for some reason.
This update's test gating status has been changed to 'passed'.
This update has been pushed to testing.
Works.
This update can be pushed to stable now if the maintainer wishes
default and performance tests pass (KVM)
Fails to suspend: Framework AMD laptop: https://bugzilla.redhat.com/show_bug.cgi?id=
That's https://bugzilla.redhat.com/show_bug.cgi?id=2253756
Works
Works for me... The enabled Default and Performance tests pass OK.
Work Station > Asus Prime mobo, Ryzen5 5600g/AMDGPU 400 Series Chipset, (UEFI + Secure Boot) SSD's > RAID1 (ext4) running Plasma DE.
Thinkpad T14s, AMD Ryzen 7840U, Fedora 39, UEFI, Secure Boot, luks+btrfs. Default and performance tests pass OK. Also suspend seems to work.
Same here on T14s, AMD Ryzen 7840U, all good. also VM is okay.
Runs OK here. Ryzen 3700x, Rog Strix B550-F, Nvidia 2070 from rpmfusion, secure boot disabled, amd pstate enabled.
Seems to be working fine on my main machine (i7-7700HQ/iHD630 + GTX1050Ti, secure boot enabled)
Working fine
Default fedora btrfs filesystem.
Suspension working
Passed kernel regression tests
Works for me.
@charn @markec given the apparent involvement of wpa_supplicant, if you have a moment could you please reply (preferably in the bugzilla issue I opened above so we don't spam more here) with the type of wireless adapter and network mode you're having suspend success with? For reference mine is an AMD/Mediatek MT7922 on a 5GHz, 80MHz-wide ax network.
I have a framework laptop with AMD CPU https://frame.work/fr/fr/products/laptop-diy-13-gen-amd and my system is unstable since I boot on 6.6.5. I lose Wifi access after a while wpa_supplicant / NetworkManager goes to 100% CPU, and process are killed by OOM (on a 32GB RAM laptop)
Issue reported on https://bugzilla.redhat.com/show_bug.cgi?id=2253789
Hi @dimitrisk - I'm using Qualcomm's NFA725A here and seems to be fine.
Regression test ./runtests.sh PASS with kernel-6.6.5-200.fc39.x86_64 within QEMU/KVM with CPU-passthrough on an AMD Ryzen 7 PRO 6850U host: F38 KDE Spin with SELinux confined user account (user_u), VM up to date as of now with updates-testing enabled, uname -r = 6.6.5-200.fc39.x86_64, tainted = 0. So far it works fine with average activities: firefox browsing, dolphin file organization.
BZ#2253756 seems to not affect my AMD hardware. In my cockpit KVM VM, I cannot wake up (which is intended and not a bug), but I can initiate sleep, and this works fine without the problem described in BZ#2253756. When checking the log after resetting the VM (resetting because the VM cannot wake up), the
journalctl -r --boot=-1shows a normal going to sleep that ends with as intended with "Entering sleep state 'suspend'" (in the VM,cat /sys/power/mem_sleep=[s2idle]). So the bug seems to not affect all AMD hardware.Next one will be https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.6 :)
Similar problem to what was reported above - Lenovo Legion 5 with AMD CPU and GPU.
Both NetworkManager and wpa_supplicant using 100% CPU, and unable to stop restart services Observing erratic behaviour and high CPU usage Requires hard shutdown / restart (using magic SysRq keys)
Works fine with kernel 6.6.4 and older.
The 6.6.6 kernel is building, it is a single patch, reverting the problematic cfg80211 patch.
This update has been obsoleted by kernel-6.6.6-200.fc39.