Public bug reported:
## Summary
On an HP ProBook 445 14" G11 (AMD Ryzen, Rembrandt/YELLOW_CARP APU,
model 0x1002:0x1681), the system intermittently fails to fully power off
or reboot after upgrading from Kubuntu 25.10 (kernel 6.17.x) to Kubuntu
26.04 LTS (kernel 7.0.x). The OS-level shutdown/reboot sequence
completes cleanly according to systemd/journald logs, but the hardware
does not fully cut power afterward (fan keeps running, power LED stays
lit, screen goes black but the machine never truly turns off or
restarts). This did not occur before the 25.10 → 26.04 upgrade. BIOS/EC
firmware is already at the latest available version, ruling out a
firmware fix as the immediate solution.
## Environment
- **Hardware**: HP ProBook 445 14 inch G11 Notebook PC
- **CPU/APU**: AMD Ryzen (Rembrandt, codename YELLOW_CARP, PCI ID 1002:1681,
subsystem 103C:8C79)
- **BIOS/EC**: 01.08.03 Rev.A (confirmed latest available via HP support site)
- **OS**: Kubuntu 26.04 LTS "Resolute Raccoon" (upgraded from Kubuntu 25.10
"Questing Quokka")
- **Kernel**: 7.0.0-29-generic (also reproduced on 7.0.0-28-generic;
7.0.0-22-generic available for further testing)
- **Desktop**: KDE Plasma, Wayland session (SDDM)
- **Secure Boot**: Enabled
- **Disk encryption**: LUKS (root on NVMe, cryptsetup)
## Regression Context
- Kubuntu 25.10 shipped with kernel 6.17.x and shutdown/reboot worked reliably.
- Kubuntu 26.04 LTS ships with kernel 7.0.x (upstream renumbering from 6.x, not
a from-scratch rewrite, but a full kernel major bump with all associated driver
updates).
- The 6.17.x kernel packages are no longer installed on this system, so a
direct old-kernel bisect is not currently possible; only kernel points within
the 7.0.x series (7.0.0-22, -28, -29) remain available.
- The issue is **intermittent**, not deterministic: some shutdowns/reboots
complete correctly (confirmed via journal: `systemd-poweroff.service` finishes
and `poweroff.target` is reached cleanly), while others leave the hardware in a
half-powered state.
- The issue has also been observed on a plain `reboot` (not just `poweroff`),
ruling out an S5-specific ACPI quirk.
## Steps Already Taken (no effect)
1. Verified BIOS/EC firmware is already the latest version (01.08.03 Rev.A) via
`fwupdmgr` and HP's support site — no update available.
2. Added `acpi_osi=Linux` to `GRUB_CMDLINE_LINUX_DEFAULT` and ran `update-grub`
— the very next reboot after applying this still hung, so this had no
beneficial effect (parameter has since been reverted).
3. Confirmed via `journalctl -b -3 -e` that a "good" shutdown completes
normally end-to-end at the systemd/journald level:
```
systemd-logind[1744]: The system will power off now!
systemd-logind[1744]: System is powering down.
...
systemd[1]: Finished systemd-poweroff.service - System Power Off.
systemd[1]: Reached target poweroff.target - System Power Off.
```
No errors, no hung units, no failed unmounts in this particular boot's
shutdown sequence.
4. Ruled out a misconfigured shutdown-vs-reboot action (checked
`systemd-logind` logs directly; confirmed the correct target —
`poweroff.target` or `reboot.target` — was requested and reached in each case
examined).
5. Checked GPU runtime power management as a possible blocker:
```
cat /sys/bus/pci/devices/0000:04:00.0/power/runtime_status -> active
cat /sys/bus/pci/devices/0000:04:00.0/power/control -> on
```
This reflects that amdgpu runtime autosuspend is disabled (consistent with
kernel log message `amdgpu 0000:04:00.0: Runtime PM not available`), but this
should not affect the forced power-down sequence during full system shutdown,
so it is likely not the root cause (included here for completeness/future
reference).
6. Reviewed `/var/log/apt/history.log*` and confirmed the kernel series
transitioned 6.17.0-22 → 6.17.0-23 → 6.17.0-29 (Kubuntu 25.10) → 7.0.0-15 →
7.0.0-22 → 7.0.0-27 → 7.0.0-28 → 7.0.0-29 (Kubuntu 26.04), alongside routine
`linux-firmware-amd-graphics` package bumps in the same window.
## Relevant Kernel Log Excerpts
From `journalctl -k -b -1 | grep -iE "amd_pmc|amdgpu|pmc|acpi.*power"`
(a normal boot after a failed shutdown, for reference — not the failure
itself, since journald is terminated before the actual hang occurs
during shutdown):
```
ACPI: \_SB_.PCI0.GP31: Overriding _PRW sleep state (S4) by S0 from power
resources
ACPI: \_SB_.PCI0.GP41: Overriding _PRW sleep state (S4) by S0 from power
resources
ACPI: \_SB_.PCI0.BUSC.NHI0: Overriding _PRW sleep state (S4) by S0 from power
resources
ACPI: \_SB_.PCI0.BUSC.NHI1: Overriding _PRW sleep state (S4) by S0 from power
resources
x86/amd: Previous system reset reason [0x00200800]: ACPI power state transition
occurred
amdgpu 0000:04:00.0: initializing kernel modesetting (YELLOW_CARP 0x1002:0x1681
0x103C:0x8C79 0x0D).
amdgpu 0000:04:00.0: Runtime PM not available
amdgpu 0000:04:00.0: [drm] *ERROR* lttpr_caps phy_repeater_cnt is 0x0, forcing
it to 0x80.
amdgpu 0000:04:00.0: [drm] *ERROR* LTTPR count is nonzero but invalid lane
count reported. Assuming no LTTPR present.
```
Note: the `_PRW` overrides on the USB4/Thunderbolt controllers
(GP31/GP41/NHI0/NHI1) stand out, as these are exactly the kind of power
resources involved in a full S5/S4 ACPI transition, and USB4/Thunderbolt
controllers are a documented source of similar shutdown/resume hangs on
other Ryzen Rembrandt/Phoenix laptops.
The `x86/amd: Previous system reset reason [0x00200800]: ACPI power
state transition occurred` line confirms the EC did register an ACPI
power transition attempt from the prior (failed) shutdown — i.e., the
transition was initiated but did not fully complete in a way that cut
power to the whole platform.
## Suspected Cause
Given that:
- BIOS/EC is already current,
- the OS-level shutdown sequence completes without any logged error,
- the failure occurs strictly after journald terminates (i.e., during the
kernel's final ACPI S5/S4 handoff to firmware),
- the regression coincides with the 25.10→26.04 (kernel 6.17→7.0) upgrade,
the most likely root cause is a regression in kernel 7.0.x's handling of
the ACPI power-off/reboot path for this specific AMD Rembrandt platform
— possibly related to USB4/Thunderbolt (NHI0/NHI1) power resource
handling, `amd_pmc`, or general ACPI S5 transition changes introduced
between 6.17 and 7.0. This matches a broader, currently-being-worked-on
class of issues where AMD's Linux team is actively patching S5 power-
transition robustness on Ryzen platforms.
## Requested Investigation
- Confirm whether other HP ProBook 445 G11 (or other Rembrandt/Phoenix AMD
laptop) users on Kubuntu/Ubuntu 26.04 (kernel 7.0.x) see the same intermittent
power-off/reboot hang that was not present on 25.10 (kernel 6.17.x).
- Bisect kernel 7.0.x commits related to `amd_pmc`, ACPI S5/S4 transition, and
USB4/Thunderbolt (`thunderbolt`/`nhi`) power management between the 6.17 and
7.0 kernel branches.
- If a specific commit is identified, consider backporting a fix or reverting
the offending change for the 26.04 LTS kernel.
## Attachments to Include When Filing
- Output of `ubuntu-bug linux` (auto-collects `lspci`, `dmesg`,
`/proc/version_signature`, apport hardware info, etc.)
- `journalctl -b -3 -e --no-pager` (clean shutdown, for baseline comparison)
- `journalctl -k -b -1 | grep -iE "amd_pmc|amdgpu|pmc|acpi.*power"` (post-hang
boot kernel log)
- `sudo dmidecode -t bios` (confirms BIOS version 01.08.03 Rev.A)
- `cat /proc/cmdline`
ProblemType: Bug
DistroRelease: Ubuntu 26.04
Package: linux-image-7.0.0-29-generic 7.0.0-29.29
ProcVersionSignature: Ubuntu 7.0.0-29.29-generic 7.0.12
Uname: Linux 7.0.0-29-generic x86_64
ApportVersion: 2.34.1-0ubuntu0.1
Architecture: amd64
AudioDevicesInUse:
USER PID ACCESS COMMAND
/dev/snd/controlC1: steffen 2580 F.... wireplumber
/dev/snd/controlC0: steffen 2580 F.... wireplumber
/dev/snd/seq: steffen 2535 F.... pipewire
CasperMD5CheckResult: unknown
CurrentDesktop: KDE
Date: Fri Aug 7 09:19:33 2026
InstallationDate: Installed on 2025-11-03 (277 days ago)
InstallationMedia: Kubuntu 25.10 "Questing Quokka" - Release amd64 (20251007)
MachineType: HP HP ProBook 445 14 inch G11 Notebook PC
ProcFB: 0 amdgpudrmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-7.0.0-29-generic
root=UUID=0277168a-b564-41b9-a91d-79407262c34a ro
rd.luks.uuid=9d5c36a9-3dc3-4fdf-a1f9-b3f2aa284b34 acpi_osi=Linux
PulseList: Error: command ['pacmd', 'list'] failed with exit code 1: No
PulseAudio daemon running, or not running as session daemon.
SourcePackage: linux
UpgradeStatus: Upgraded to resolute on 2026-05-17 (82 days ago)
dmi.bios.date: 05/19/2026
dmi.bios.release: 8.3
dmi.bios.vendor: HP
dmi.bios.version: W78 Ver. 01.08.03
dmi.board.name: 8C79
dmi.board.vendor: HP
dmi.board.version: KBC Version 06.33.00
dmi.chassis.type: 10
dmi.chassis.vendor: HP
dmi.ec.firmware.release: 6.51
dmi.modalias:
dmi:bvnHP:bvrW78Ver.01.08.03:bd05/19/2026:br8.3:efr6.51:svnHP:pnHPProBook44514inchG11NotebookPC:pvrSBKPFV3:rvnHP:rn8C79:rvrKBCVersion06.33.00:cvnHP:ct10:cvr:sku9Y7J4ET#ABD:pfa103C_5336ANHPProBook:
dmi.product.family: 103C_5336AN HP ProBook
dmi.product.name: HP ProBook 445 14 inch G11 Notebook PC
dmi.product.sku: 9Y7J4ET#ABD
dmi.product.version: SBKPFV3
dmi.sys.vendor: HP
** Affects: linux (Ubuntu)
Importance: Undecided
Status: New
** Tags: amd64 apport-bug resolute wayland-session
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2163015
Title:
System fails to power off / hangs on shutdown-reboot after upgrade to
Kubuntu 26.04 (kernel 7.0)
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2163015/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
