Public bug reported:
NVMe controllers lost after s2idle resume (D3cold -> D0 "device
inaccessible") on HP HyperX OMEN 15-gb0xxx (AMD Ryzen 7 8745HX),
reproducible on 6.12 and 7.0
1) Ubuntu 26.04.1 LTS (Release 26.04)
2) linux-image-7.0.0-38-generic 7.0.0-38.38 (also reproduced with
linux-image-7.0.0-1015-oem 7.0.0-1015.15 and with Debian 13.7 live, kernel 6.12)
3) After suspend (systemctl suspend or closing the lid) and wake-up, the system
resumes normally and both NVMe drives are available
MORE DETAILS:
After `systemctl suspend` (s2idle, the only sleep state exposed) and wake-up,
one or both NVMe controllers fail to leave D3cold:
nvme 0000:02:00.0: Unable to change power state from D3cold to D0, device
inaccessible
nvme nvme0: Disabling device after reset failure: -19
When the system disk is affected, the machine dies with `Buffer I/O
error`, `EXT4-fs ... shutdown requested`, `JBD2: I/O error` and needs a
forced power-off. The same error occurs on Debian 13.7 live (kernel
6.12), so it is not a regression in the 7.0 series, and it is
independent of the distribution.
Hardware (DMI)
- sys_vendor: HP
- product_name: HyperX OMEN Gaming Laptop 15-gb0xxx
- product_family: 103C_5335M7 HyperX OMEN
- board_name: 8EEF, board_version: 04.21
- BIOS: Insyde F.16, date 09/07/2026 (updated from F.10, no change in behaviour)
- CPU: AMD Ryzen 7 8745HX; GPUs: AMD Radeon 610M + NVIDIA RTX 5060 Laptop
- Sleep states: `/sys/power/mem_sleep` = `[s2idle]` (no S3)
- NVMe drives (different vendors, different root ports):
- SK hynix PVC10 HFS512GEM9X173N, fw HPS2, 0000:02:00.0 behind root port
0000:00:01.2 (the dGPU is behind 0000:00:01.1)
- WD PC SN7100S SDFPMSL-512G-1101, fw 7611M001, 0000:04:00.0 behind root port
0000:00:03.1
Reproduction
1. Boot (Ubuntu 26.04 with 7.0.0-38-generic or 7.0.0-1015-oem, or Debian 13.7
live with 6.12).
2. `systemctl suspend`, wake with the power button or lid.
3. Observe the NVMe errors above (on Debian live, where nothing from the
internal drives is in use, the session survives and the errors are visible in
`dmesg`; on an installed system the root file system is lost).
Boot log shows, for both drives: `nvme 0000:02:00.0: platform quirk:
setting simple suspend` and `nvme 0000:04:00.0: platform quirk: setting
simple suspend`, i.e. ACPI StorageD3Enable is advertised; `amd_pmc` is
loaded.
What was tried (no improvement unless noted)
- `nvme.noacpi=1`, `pcie_port_pm=off`, both together (verified in /proc/cmdline)
- `d3cold_allowed=0` for both drives and their parent root ports
- BIOS F.10 -> F.16
- Kernel 7.0.0-1015-oem instead of 7.0.0-38-generic; Debian 6.12 (same error,
on the SK hynix drive; the WD drive recovered in that single run)
- Not yet tried: `nvme_core.default_ps_max_latency_us=0`, `pcie_aspm=off`
- Hibernation via systemd: `HibernateMode=test_resume` works (image written and
read back, drives not power-cycled); with real power-off
(`platform`/`shutdown`) the machine usually hangs after `PM: hibernation:
hibernation entry` and needs a forced power-off.
Attachments
- `dmesg` after boot and after the failed resume, `lspci -vvnn` (with
serial-number capabilities removed), `acpidump`, output of `sudo cat
/sys/kernel/debug/amd_pmc/s0ix_stats` if available, and the amd-s2idle report
from amd-debug-tools if possible.
- A raw capture of the failure on Debian live is more useful than the
hand-transcribed excerpt currently in
`inbox/hp-omen-attachments/debian-live-6.12-nvme-excerpt.txt`.
Request
Is a DMI-based quirk (for example disabling D3cold/StorageD3Enable usage for
this board) acceptable? I can test patches on this machine and on a Debian live
USB without touching the internal drives.
ProblemType: Bug
DistroRelease: Ubuntu 26.04
Package: linux-image-7.0.0-1015-oem 7.0.0-1015.15
ProcVersionSignature: Ubuntu 7.0.0-1015.15-oem 7.0.14
Uname: Linux 7.0.0-1015-oem x86_64
ApportVersion: 2.34.1-0ubuntu0.1
Architecture: amd64
CasperMD5CheckResult: pass
CurrentDesktop: ubuntu:GNOME
Date: Fri Oct 16 12:38:46 2026
InstallationDate: Installed on 2026-10-13 (3 days ago)
InstallationMedia: Ubuntu 26.04.1 LTS "Resolute Raccoon" - Release amd64
(20260826)
ProcEnviron:
LANG=pl_PL.UTF-8
PATH=(custom, no user)
SHELL=/bin/bash
TERM=xterm-256color
XDG_RUNTIME_DIR=<set>
SourcePackage: linux-signed-oem-7.0
UpgradeStatus: No upgrade log present (probably fresh install)
** Affects: linux-signed-oem-7.0 (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/2171066
Title:
NVMe drives lost after s2idle resume (D3cold -> D0, device
inaccessible) on HP HyperX OMEN 15-gb0xxx (board 8EEF, BIOS F.16)
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux-signed-oem-7.0/+bug/2171066/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
