Hibernation fails due to resource busy during swap file check

[bsdz]

Hopefully, the automatically collected information is useful. The issue manifests when I try to hibernate the device. I see the following in the logs:

~$ journalctl --no-hostname -b -1 | tail -6
Feb 25 08:23:54 systemd[1]: nvidia-hibernate.service: Deactivated successfully.
Feb 25 08:23:54 systemd[1]: Finished NVIDIA system hibernate actions.
Feb 25 08:23:54 systemd[1]: Starting Hibernate...
Feb 25 08:23:54 kernel: PM: Image not found (code -16)
Feb 25 08:23:54 systemd-sleep[50360]: Entering sleep state 'hibernate'...
Feb 25 08:23:54 kernel: PM: hibernation: hibernation entry

I looked up error code 16 as EBUSY. This happens when using pm-hibernate or "systemctl hibernate". I played qaround with the nvidia systemd hibernate units but it didn't help.

Hibernate used to work before I upgraded to jammy. I was on focal and bionic before that.

ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: linux-image-5.15.0-18-generic 5.15.0-18.18
ProcVersionSignature: Ubuntu 5.15.0-18.18-generic 5.15.12
Uname: Linux 5.15.0-18-generic x86_64
NonfreeKernelModules: nvidia_modeset nvidia
ApportVersion: 2.20.11-0ubuntu78
Architecture: amd64
CasperMD5CheckResult: unknown
CurrentDesktop: ubuntu:GNOME
Date: Sat Feb 26 09:27:29 2022
InstallationDate: Installed on 2019-08-07 (933 days ago)
InstallationMedia: Ubuntu 18.04.2 LTS "Bionic Beaver" - Release amd64 (20190210)
MachineType: Gigabyte Technology Co., Ltd. X399 AORUS XTREME
ProcFB: 0 EFI VGA
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-5.15.0-18-generic root=UUID=08f6326a-30f9-4f23-bab7-72a48afad354 ro quiet splash resume=UUID=08f6326a-30f9-4f23-bab7-72a48afad354 resume_offset=34816 vt.handoff=7
PulseList: Error: command ['pacmd', 'list'] failed with exit code 1: No PulseAudio daemon running, or not running as session daemon.
RelatedPackageVersions:
 linux-restricted-modules-5.15.0-18-generic N/A
 linux-backports-modules-5.15.0-18-generic N/A
 linux-firmware 20220217.git6342082c-0ubuntu1
SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 12/11/2019
dmi.bios.release: 5.14
dmi.bios.vendor: American Megatrends Inc.
dmi.bios.version: F5
dmi.board.asset.tag: Default string
dmi.board.name: X399 AORUS XTREME-CF
dmi.board.vendor: Gigabyte Technology Co., Ltd.
dmi.board.version: x.x
dmi.chassis.asset.tag: Default string
dmi.chassis.type: 3
dmi.chassis.vendor: Default string
dmi.chassis.version: Default string
dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvrF5:bd12/11/2019:br5.14:svnGigabyteTechnologyCo.,Ltd.:pnX399AORUSXTREME:pvrDefaultstring:rvnGigabyteTechnologyCo.,Ltd.:rnX399AORUSXTREME-CF:rvrx.x:cvnDefaultstring:ct3:cvrDefaultstring:skuDefaultstring:
dmi.product.family: Default string
dmi.product.name: X399 AORUS XTREME
dmi.product.sku: Default string
dmi.product.version: Default string
dmi.sys.vendor: Gigabyte Technology Co., Ltd.

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This change was made by a bot.

[Kai-Heng Feng (kaihengfeng)]

Is the swap space big enough? Furthermore, the hibernation failure isn't logged in the dmesg.

[bsdz (blairuk)]

> Is the swap space big enough?

A good point. This worked on previous versions on Ubuntu with swap (18GB) only 50% of RAM (32GB). I recreated my /swapfile with 38GB and the same issue persists. ie.

$ free -m
               total ...
Mem: 31964
Swap: 38911

> Furthermore, the hibernation failure isn't logged in the dmesg.

That's correct. The machine just stalls with a blank screen. It's still awake as is apparent by motherboard lights. It cannot be awoken in any way and needs to cold restarted using power button.

Curiously if I run smem before I attempt to hibernate, it does not show any processes using swap. ie

$ smem --columns="swap" --no-header | tr -d " " | awk '$1 != 0' | wc -l
0

[James (morris-570)]

Same here. I have 48GB Ram and 76GB swap partition. Where did you see the error code -16 is EBUSY?

[bsdz (blairuk)]

For me, when I ran journalctl (see above) the following message was present: "kernel: PM: Image not found (code -16)". The minus sign is prefixed to the error code and 16 is a standard linux kernel error code, i.e.

$ grep 16 /usr/include/asm-generic/errno-base.h
#define EBUSY 16 /* Device or resource busy */

[zyf0330 (zyf0330)]

I am lucky that find the same problem here.
I use Ubuntu 21.10, and the error info is same `kernel: PM: Image not found (code -16)`.
This is my swap info

NAME TYPE SIZE USED PRIO
/dev/nvme1n1p5 partition 32G 0B -2

[zyf0330 (zyf0330)]

Sorry, I am wrong.
Previously, I set "resume=LABEL=Swap" at grub `GRUB_CMDLINE_LINUX` param, and got error msg `kernel: PM: Image not found (code -16)` and resume failed.
After change to set `resume=/dev/xxx-to-swap`, resume works.

But although resume succeeds, error msg `kernel: PM: Image not found (code -16)` still appears in the syslog sometimes.

[Fabio (fabiofs-br)]

I'm getting this same error after reinstall my system (Linux Mint 20.3 Una base: Ubuntu 20.04 focal).
No workaround which I tried to fix it worked until now, and I'm unable to hibernate my system.

~$ journalctl --no-hostname -b -1 | tail -6

jun 25 16:32:04 kernel: PM: Image not found (code -16)
jun 25 16:32:04 systemd[1]: Starting GRUB failed boot detection...
jun 25 16:32:04 systemd-sleep[6466]: Suspending system...
jun 25 16:32:04 kernel: PM: hibernation: hibernation entry
jun 25 16:32:04 systemd[1]: grub-initrd-fallback.service: Succeeded.
jun 25 16:32:04 systemd[1]: Finished GRUB failed boot detection.

Hope anybody can come up with a solution.

[bsdz (blairuk)]

I had another look at my hibernation settings and realised my resume offset was wrong in my /etc/default/grub. After updating this with the value from command (found on https://wiki.archlinux.org/title/Power_management/Suspend_and_hibernate):

~$ sudo filefrag -v /swapfile | awk '$1=="0:" {print substr($4, 1, length($4)-2)}'

Then hibernate appears to work again. Perhaps this offset changed during an upgrade of my base system to 22.04 and/or changed again when I increased my swap size during this bug.

All this got a little forced on me when my suspend capability stopped working several days ago (after an apt upgrade!). I'll take another look at that again now :-)

[Fabio (fabiofs-br)]

Well... I may have found a workaround to this problem. It worked in my case at least. The new NVIDIA driver was the culprit.
I had my share of problems with it in hibernation configurations before, so I try to install a old version. Based in this thread found in NVIDIA forums.
https://forums.developer.nvidia.com/t/resume-from-hibernation-fails-with-510-60-02/211025

As soon as I downgraded to NVIDIA Driver 510.54 my hibernation start to work. NVIDIA Driver 470 works too, but it has an issue with my 3 screen setup.
Basically, looks like NVIDIA Driver 510.60.02 introduced this regression.