[BUG] Lunar Lake (Core Ultra) stuck at 400MHz after s2idle resume

Semyon

New Member
Joined
Jul 20, 2026
Messages
1
Reaction score
0
Credits
23
Hardware & Software:
1)Laptop: Lenovo ThinkPad X9-15 Gen 1
2)CPU: Intel Core Ultra 7 258 V
3)GPU: Intel Arc Graphics (using the xe driver)
4)Kernel: 7.1.3-arch1-3
5)Environment: Wayland (Hyprland also tried niri)
6)I have no thermald

After waking up from s2idle (PC10), the Wayland interface lags critically. The freezes last from 20 to 30 seconds, after which the system fully recovers and operates normally. During the lag period, the turbostat records a hard lock of all cores at the Lowest Frequency Mode (LFM) - 400 MHz, (even if write stresss -c 8).
Code:
sudo turbostat --show PkgWatt,CorWatt,Bzy_MHz,Avg_MHz,PkgTmp
...
Avg_MHz    Bzy_MHz    PkgTmp    PkgWatt    CorWatt
399    400    30    3.20    1.58
399    400    30    3.20    1.58
399    400
399    400
399    400
399    400
399    400
399    400

What I have already tried:
1) reinstall system
2) clear uefi and ec
3) install mainline kernel
4) change /proc/acpi/wakeup
5) set other flags ( nvme.noacpi=1 nvme_core.default_ps_max_latency_us=0 amd_iommu=off iommu=soft pcie_aspm=off xe.enable_psr=0 pcie_port_pm=off intel_pstate=no_hwp thermal.off=1 ) doesn't help
6) kernel, uefi, firmware is up to date
7) install throttled
8) Disable BD PROCHOT - unsuccessful
Code:
cat /sys/kernel/security/lockdown -pP
[none] integrity confidentiality
Code:
sudo cat /sys/module/msr/parameters/allow_writes
on
Add to grub msr.allow_writes=on, but
Code:
sudo rdmsr 0x1FC
ec005b
sudo wrmsr -a 0x1FC 0xec005a
sudo rdmsr 0x1FC
ec005b
First bit (BD PROCHOT) didn't change, it still active

What I have interesting found:
1) set intel_idle.max_cstate to 2 (and this helps, but battery leaks very fast)
2) Before suspend
Code:
sudo rdmsr -x 0x19C
88310000
During the lag
Code:
sudo rdmsr -x 0x19C
88463c00
After the lag
Code:
sudo rdmsr -x 0x19C
88352800
The 10-th bit is Power Limitation Status
The 11-th bit is Power Limitation log
Before sleep they were 00
During the lag were 11
After the lag became 10

What I want?
I've been trying to solve the problem for a long time. At some point, I was just wondering what the reason was and how to solve the problem as harmlessly as possible.

There is thread in arch forum "system lag after resuming from deep s2idle (PC10)"
 


Follow Linux.org


Top