WingNews logo WingNews GitHub [2]
top | item 26739220

Ask HN: Why is the Linux community struggling to implement hibernation?

322 points| cyka178 | 5 years ago | reply

Today I have found out that hibernation is by default disabled in my Ubuntu 18.04 distribution. After this, I found this 11 month old post https://discourse.ubuntu.com/t/re-visiting-hibernate-on-ubun... where I realized that the Linux community does not seem to be able to implement a working implementation of hibernate. Is there any reason why this is a difficult problem? I would like to have an option in my OS like VM's have where everything that is currently running is saved on disk and can be resumed later without issues.

332 comments

order
[+] gvb|5 years ago|reply
Hardware (peripherals) have state. Reading the hardware state is difficult and sometimes impossible so recreating that state correctly after powering down is very difficult. Then multiply that difficulty across all the different hardware that can be in a PC and it becomes extremely difficult to do on an arbitrary laptop/PC.

Add to that difficulty the fact that many hardware (peripheral) vendors provide incomplete documentation making it difficult or impossible to implement "off-nominal" situations like hardware state reading/writing-restoring.

VMs have an advantage that the peripherals are limited in types and numbers and the state is "virtual" so the state is directly accessible (r/w) in the virtualizing driver software, not buried in physical hardware. When you "hibernate" a VM, no physical device actually powers down and no (buried in hardware) state is lost.

[+] addicted|5 years ago|reply
I think the non server Linux community needs a reference device.

Basically, you develop against that device, and it's upto hardware manufacturers to try and match the reference device to make their laptops/desktops compatible.

If you could get a couple of distro owners such as Red Hat, Ubuntu, Arch, etc. to agree upon such a device it would go a long way to resolving a lot of the issues that the desktop community faces.

Disclaimer: I have no expertise in this on either the hardware or software end so this idea might be completely nonsensical for all I know.

[+] rcxdude|5 years ago|reply
Also the amount of stuff the firmware does nowadays, especially on a laptop, is quite high. The OS-UEFI interface is quite complex and involves a fair amount of firmware code which keeps running after boot. This code is heavily involved in operations like suspend and hibernate, and is usually developed by the laptop OEM themselves, so quality is not very high and they have little incentive to test with anything other that windows. This results in a tendency towards fragility (even with windows), as well as worse battery life under linux.
[+] cesarb|5 years ago|reply
And besides all that, hibernate is fundamentally more complex than sleep. With sleep, when waking up the computer resumes directly to the already loaded operating system in RAM, and the operating system "only" has to restore all the hardware state. With hibernate, the computer does a clean boot of a fresh copy of the operating system, which has to initialize all hardware, read the hibernation image, de-initialize (quiesce and/or power off) all hardware, and finally jump to the copy of the operating system found in the hibernation image, which only then will restore all hardware state.
[+] liquidify|5 years ago|reply
how does winndows resolve these issues?
[+] ironmagma|5 years ago|reply
When does restoring state become a necessity as opposed to just disconnecting the devices upon hibernate and reconnecting them afresh when the device wakes up?
[+] test6554|5 years ago|reply
Why not create a low-level hardware state interface layer that linux uses to save/restore hardware state. The attitude could be "it's not linux's job to fix your hardware" At first nothing would use it and it would be basically useless, but eventually manufacturers would add support. In the meantime, Another layer on top of this low layer could be created to make older hardware compatible. For hardware that is worth the effort, people could write drivers or other things that could bridge the gap between hardware and the low-level interface.
[+] m463|5 years ago|reply
I remember things like getting initialized by at boot cleanly vs coming back from sleep and having to remember the no-boot-state setup.

It's like savegames - I think it's hard and bug-ridden so some folks just add checkpoints and don't bother with complex restores in the middle of the bossfight with rockets in the air.

[+] whatever1|5 years ago|reply
Have no clue from hardware, but why we cannot take a full memory snapshot to the hard disk and then just load it back bit to bit ?
[+] saltedonion|5 years ago|reply
Can’t this also be said about windows. How do they solve this problem ?
[+] Kliment|5 years ago|reply
After setting up a swap partition of sufficient size, hibernate worked with no issues whatsoever on my very recent hardware and kubuntu. The main reason it's disabled by default on ubuntu is that the default partition setup isn't compatible with hibernation for most memory sizes. You need a swap configuration that's large enough to fit your entire RAM and also your current swap. I find it works at least as well as on any other OS. I currently have it configured to auto-hibernate on low battery or after some number of hours of sleep. From my point of view the Linux community is doing absolutely great on that front, and I've never had a machine where hibernate didn't work correctly. The main issue here is the defaults on one distribution (ubuntu) make it a massive pain to enable.
[+] bbulkow|5 years ago|reply
And you need to find some buried comment on hacker news to figure out that partitioning is the problem! Thanks!
[+] bouncycastle|5 years ago|reply
Same here - the swap partition to match the memory size did the trick. I'm on a Lenovo 480. To hibernate, I just type in "sudo systemctl hibernate" and it works, I'm happy.

Incidentally, I've just upgraded my RAM to 64GB and it's in the process of resizing the swap partition right now!

[+] herodoturtle|5 years ago|reply
This was a fascinating comment to read - thanks for the insight.

> You need a swap configuration that's large enough to fit your entire RAM and also your current swap.

I'm struggling to wrap my head around this one - please help me understand?

How can a swap be made large enough to fit your entire RAM as well as your current swap?

In other words, what if your current swap is already being used near capacity?

Or is that not really likely with a really large swap?

[+] deckard1|5 years ago|reply
It's going to depend on your hardware and probably your distro.

I have a ThinkPad running Archlinux. I have suspend-then-hibernate mode enabled, which means when I close my laptop lid it suspends for 2 hours (configurable) and then puts itself into hibernate. If I decide to open the laptop within 2 hours, it's available instantly. Otherwise, it takes maybe 20 seconds to reanimate from hibernation. People often say, why not just shutdown and reboot? The obvious answer is that I don't want to close out a dozen Chrome tabs, close all my terminals, all my files, and open everything up every time I need to walk away from the computer.

I have secure boot turned on. Hibernation is done to a swap file (I can simply delete the file if I need the space). I have LUKS encryption enabled. I'm dual booting with Windows. And yes, it all does work perfectly fine.

You can have your cake and eat it too. It just takes a bit of research and a bit of time to setup.

[+] k__|5 years ago|reply
Not even MacOS gets it right if you confront it with the same problem that Linux is facing.

I switched the SSD of my MBP 2 years ago and since then everything is out of control. Sometimes the thing keeps running when I close the lit and when I open it again the battery is empty. Sometimes it simply shuts down entirely and has to boot from scratch when I open it again.

I tried reinstalling the OS and whatnot, nothing worked.

[+] hospadar|5 years ago|reply
Personal hot take: It's super hard and no one cares very much.

With SSDs, booting is fast. Most apps that I care about which have complex state (browser, IDE) do a much better job of saving their state than the OS can do (browsers & IDEs can and do easily restore sessions, etc).

[+] jules|5 years ago|reply
For a laptop it is important to be able to close the lid, go to zero power state after X minutes, then resume where you left off.
[+] pkulak|5 years ago|reply
Plus, with full-disk encryption, I don't want my key getting written to disk so that anyone can get it until the end of time. I'm okay with my laptop powering down fully after sleeping for some amount of time.
[+] pixelmonkey|5 years ago|reply
One reason is that a big community effort to rewrite the suspend-to-disk layer, which made it more reliable, called TuxOnIce, was kept out of the mainline kernel for a number of reasons, and thus a lot of "community energy" for solving the problem has been sapped out. Also, this is a uniquely "desktop Linux" problem, so exactly the kind of problem that rarely gets commercially sponsored developer hours thrown at it.

https://en.wikipedia.org/wiki/TuxOnIce

[+] dorfsmay|5 years ago|reply
As a side note, if you have a recent laptop, there is probably a setting in the BIOS for sleep mode, Windows 10 vs Linux.

This makes a huge difference on Linux, and I hear the windows mode is somewhat experimental and not working so well yet even with Windows.

https://mobile.twitter.com/dorfsmay/status/13629639921094778...

[+] techrat|5 years ago|reply
This tends to also apply to whitebox models. I have a no name ODM laptop which has options in the Bios to indicate the system is booting as a Windows, Linux or Android device.

Oddly, enabling Android got the touchpad working in Ubuntu.

[+] terinjokes|5 years ago|reply
This switch is gone on late 2020 Dells and forward. It's the new experimental Windows way only.
[+] rickspencer3|5 years ago|reply
I was involved in these discussion for Ubuntu ... back in 2011.

You can see some discussion here: https://bugs.launchpad.net/ubuntu/+source/policykit-desktop-...

and even more here: https://bugs.launchpad.net/ubuntu/+source/policykit-desktop-...

net/net even back then, no matter how much effort we put into it, we could never make it work on even most machines because there was just too much variability. The OEM team could make it work for specific machines that Ubuntu was pre-installed on, but otherwise, hibernate was just a bug farm.

[+] _joel|5 years ago|reply
Hibernate with LUKS (albeit after a lot of hacking) works pretty well on Debian Buster. I've resumed and hibernated for many cycles with no issues. Wifi comes back etc. It never used to be the case, in my experience at least
[+] dathinab|5 years ago|reply
I just wanted to write it works without any hacking on Arch Linux, then I realized that 100% of the Arch Linux installation process would normally count as hacking. So I guess hacking it is. ;-)
[+] yaantc|5 years ago|reply
Same here, I'm using it all the time with Debian on two laptops (Latitude and Thinkpad) and it's stable for me.

Still, with the current implementation one must disable secure boot to enable hibernation (which I do). Having hibernation and secure boot is the (complex) missing part, see [1] for why.

[1] https://mjg59.dreamwidth.org/55845.html

[+] wing-_-nuts|5 years ago|reply
I have to be honest, I struggle to think of a scenario where simply putting the computer to sleep isn't the better option. If power draw is a concern, just power the computer off? It's certainly not worth having a 16 or 32 GB swap partition on my machine just to enable hibernate
[+] jcheng|5 years ago|reply
I was hoping to replace my laptop with a NUC running Linux, that I would transport between home and office. Since a NUC is a (small) desktop machine it doesn’t have a battery. Hibernate would have been great.
[+] alerighi|5 years ago|reply
For desktop computers it makes sense. I find that hybrid hibernation gets the best of the two world. I can have the PC in sleep and have it turn on faster, or if the PC looses power you don't loose anything.

A thing that I do often is that I forget my desktop PC in sleep mode but still powered, and since I know that hibernation works I can simply cut power to the whole power strip to save energy without having to wake up the PC just to shut it down correctly and then remove power safely.

By the way, you can also have a swap file, but it's not as simple to configure (and still the file size is kind of fixed and that removes the point).

[+] throwaway823882|5 years ago|reply
> If power draw is a concern, just power the computer off?

I have about a thousand browser tabs and 20 terminals open at any given time. Re-setting all that up would be a pain. I haven't shut down my laptop in 6 months, but I hibernate and resume it regularly.

[+] detaro|5 years ago|reply
a bit of storage is easily worth the benefit for me. Now if I had some superthin laptop with soldered-in SSD at exorbitant pricing that might be different, but other than that disk space is not that much at a premium.
[+] ashneo76|5 years ago|reply
I don't think this is true.

Regularly my MBP dies with the lid closed with no option to hibernate. After which, it refuses to start for 15 minutes after being plugged in, before it agrees to turn on. To me, that is user hostile.

Windows hides the hibernate option until you do some magic incantation. Often upon restart, wifi driver is disabled or cannot be enabled and I have to restart.

On Linux, At least with a decent swap size, all I have to do is `systemctl hibernate` and it works like magic. LUKS and sd-boot are what I use as well.

[+] ineedasername|5 years ago|reply
My very naïve understanding of hibernation was that basically RAM was written to disk as some sort of image, and coming back from hibernation means loading that back into RAM... I'm getting the impression that this is very much not true.

I can't, however, easily find an explanation of how actual hibernation is different from this simplistic understanding.... would someone care to explain?

[+] Syonyk|5 years ago|reply
At a very, very high level, you're exactly correct. The state of RAM is written out to disk, and on restore, it's loaded back into memory, and stuff resumes.

But the devil is very much in the details, and you may not have access to any of the details because the datasheets are NDA'd or just flat out not available to anyone outside the OEMs.

The problem is mostly around hardware state. I've written some sleep patches for Linux on a PineBook Pro (I should finish those and try to upstream them...), and the problem is that on restore, you have to put the hardware exactly as it was when you put it to sleep, or things just don't work. The details of both querying the hardware state, and putting it back, range from "fairly straightforward if inefficient" (copying out device register state on sleep, restoring state on resume - but which of them actually matter?) to "really, really nasty, involving going through the init sequences again, then updating state in the proper order."

Resuming from hibernate is almost the same as a cold boot, except that it's a different shutdown state, so the boot firmware may or may not do things the same as on a cold boot. If you've moved PCI devices around to where you want them instead of where the boot firmware put them, you have to shuffle them back to where you want them - and do this in the correct order, which may or may not be the same as the ordering you did it when the kernel first came up.

If any of this isn't done properly, the system will be in a weird state and things might not work on resume. The patches I wrote for my PBP work fine, unless there's audio playing when the system goes to sleep. Then it still doesn't restore audio state properly and I've not taken the time to dive into the Linux sound system enough to understand why... and this is sleep (suspend to RAM), not hibernate (suspend to disk). Coming up from hibernate, the device state might be entirely different.

Like just about anything with modern computers, a simple idea turns into an absolute minefield of implementation details, and the Linux side of things is hindered because you often simply don't have the datasheets. So now you have to reverse engineer Windows drivers if you want support, or guess that, well, this device is in the same family as that similar device we have a leaked datasheet for from some OEM's FTP site, so maaaaybe it uses the same weird register layout, and... etc.

[+] pedro2|5 years ago|reply
Thanks to Secure Boot (AFAIK), you can't implement hibernation until you implement hibernation file signing.

Sleep works fine for ages. Works better on Linux on my 2015 laptop than on Windows (hint: if you have a KIRA laptop, suspend stops borking after you update the drivers for the touchpad to a ~2019 version).

[+] dathinab|5 years ago|reply
You can, by having LVM on top of a LUKS or a swap file on a LUKS partition.

E.g. for a relative secure setup you can:

- use a custom platform key

- use EFIStub boot and bundle you kernel initramfs, flash screen image, kernel options, etc. into a single efi bootable blob

- sign that blob with your custom platform key (or enroll it's hash if you don't use a custom platform key).

- copy your secure boot signed early boot environment into the efi partition.

- have everything else in encrypted partitions including /boot (which your are not using)

- easiest way to have this is make a EFI partition and a single large encrypted partition and then use LVM (logic volume manager 2) on it to split it into swap, root, /home etc. Added benefit is you can easily resize them.

- now you just need to setup a way to decrypt on boot, e.g. by having a boot password, secure key or similar. You will need to provide it even if you boot from hibernation. (or you use the platforms TPM!)

Naturally you can also just ignore how hibernation breaks secure boot and set it up anyway. Your computer doesn't know you are braking the secure boot spec. But a company (like Ubuntu) might have committed to not brake the spec.

[+] dfox|5 years ago|reply
uswsusp can both sign and/or RSA encrypt the image essentially from the beginning of its existence. On the other hand you in fact do not need that and can simply store the hibernation image on otherwise encrypted device (either as LV on encrypted LVM or as a file somewhere).
[+] hobonate|5 years ago|reply
Frustrated with my Fedora Linux configuration, I recently got hibernate to work with my Gentoo Linux system.

I have a LUKS (but not LVM) encrypted root and swap partition on /dev/sda3 and /dev/sda4 respectively.

I use the command "su && mount /boot && genkernel --luks initramfs" to generate the initramfs.

I give the initramfs the following parameters: "crypt_root=/dev/sda3 root=/dev/mapper/root crypt_swap=/dev/sda4 resume=/dev/mapper/swap". This makes my initramfs ask for my password for the partitions, then decrypt the partition, then mount them and resume from hibernation if possible.

I add this parameter by setting GRUB_CMDLINE_LINUX in /etc/default/grub, and then running "grub-mkconfig -o /boot/grub".

When I want to hibernate I use the following command: "exec loginctl hibernate". My window managers let me bind this to a keyboard combination.

I am using openrc as my init system (with elogind), and sway/GNOME as my window manager. The computer uses legacy BIOS (SeaBIOS) and the storage uses GUID partition table with the grub bootloader. Hope this helps.

[+] ka0lin|5 years ago|reply
You should add this to the Gentoo Wiki.
[+] rurban|5 years ago|reply
You are misleaded. Only the Debian and Ubuntu communities are struggling with technical issues, such as hibernation. What do you expect from lawyers running a distro?

The technical linux communities, like Redhat, Suse or Arch figured it out. See eg https://www.ctrl.blog/entry/fedora-hibernate.html

[+] fsflover|5 years ago|reply
Hibernation and suspend work fine in Linux on hardware designed for Linux.
[+] smhenderson|5 years ago|reply
I think this sums it up pretty well. As other posters have mentioned, Windows works great until it doesn't, Mac does it well unless you replace some internal hardware with something other than their branded gear. O

OpenBSD hibernate, and power management in general, works great on Thinkpads but the developers specifically like and use them regulary, at least so I've heard.

So I don't think it's any one OS that works well or badly or better than another; it comes down to the combination of the OS and the hardware and the compatibility between the two...

[+] rurban|5 years ago|reply
Partially. You can also assume that something that does not work on Ubuntu will work fine in a RedHat distro. Esp. something kernel, firmware, systemd, compiler or qemu related.
[+] phendrenad2|5 years ago|reply
Maybe some important context that's missing for a lot of people is - on Windows hibernation is disabled by default too. So perhaps the question is, is hibernation still relevant in 2021?
[+] twobitshifter|5 years ago|reply
Reading replies, it seems hibernation is not something that users ask for. What they ask for is to restore the program(s) to the state that they left it in. Files open, state loaded in memory etc. This should be an easier problem than a special boot sequence that is restoring hardware state. There’s probably one or two programs open at a time, that I don’t want to lose my place in, maybe some that I’ve written myself. If we wanted to add a “save state” option in the title bar would that be a tractable problem?

Of course app developers can make a restart less painful too. Web browsers have already implemented restoring tabs from the last use, and Outlook will automatically save any of your drafts. For any work in the cloud, you really just need a pick up where you left off cue.

[+] chme|5 years ago|reply
Not even Microsoft get hibernation right all the time. So this isn't a issue that only Linux has.

Every time my desktop PC running Windows 10 tries to wake up from a suspend-to-disk or even suspend-to-RAM state, it resets its BIOS and RTC.