on endeavouros when I attempt to hibernate, I am greeted with Hibernate inconsistent memory map detected! The hibernation then fails, and the system shuts down. I’ve looked around and none of the solutions seem to be applicable to my situation. I’m using a swap drive, I am also using systemd boot and not grub, and remounting the swap did not do anything in particular. Kind of at a loss here as to what could resolve this. The end goal is for a sleep then hibernate setup, but that obviously doesn’t work currently as when it hibernates it simply shuts down the computer due to aforementioned reasons.
Does anybody have any information on this? This is quite literally the only issue I have with this laptop. The standby time with sleep is awful but the boot time is extremely fast. Being able to hibernate solves all my issues, but I cannot get it to work consistently whatsoever.
There currently is no hibernation under Linux if you are using the brtfs file system.
Interesting. I’m using brtfs, but have a swap drive. If this is the issue I guess I’ll reinstall Arch.
I do believe that you can change the file system during install, but I haven’t tried that myself as I wasn’t aware of this until after the fact.
Might be worth a try if you are reinstalling another one anyway.
Are you sure about the no hibernation on btrfs?
1 Like
I am hibernating to btrfs constantly, this has been fixed ages ago.
1 Like
Well, what is the setup? Kernel? Anything? Because I currently cannot hibernate constantly, or at all.
Plain arch, luks encrypted main btrfs partition with a swapfile, various kernels up to at least 7.0.10 (this really has been working for a while).
There is a special command to tell btrfs to create a swapfile and to get its offset and then you are basically done (also have to add it to fstab and boot parameters of course).
Not even sure you still need to do the offset thing, I’ll have to try without it at some point.
Are you using the AI 350 with Nvidia GPU variant? I’m going to try to set it up again, this time on ext4, but as I said last time with a swap partition and on brtfs, no encryption I was not able hibernate. Maybe once out of 20 times it would successfully do it, but generally I would get inconsistent memory map error and the system would just boot without hibernating properly
No but I don’t see how that would impact anything about btrfs ability to hibernate.
Have you tried if it works better without the nvidia gpu?
Just did, still nothing.
Jun 08 09:01:01 kunta-mkFrame kernel: platform MSFT0201:00: AMD-Vi: [Firmware Bug]: No ACPI device matched UID, but 1 device matched HID.
Jun 08 09:01:01 kunta-mkFrame kernel: rust_binder: Loaded Rust Binder.
Jun 08 09:01:01 kunta-mkFrame kernel: hid-multitouch 0018:093A:0274.0003: Returned feature report did not match the request
Jun 08 09:01:02 kunta-mkFrame kernel: ucsi_acpi USBC000:00: unknown error 256
Jun 08 09:01:02 kunta-mkFrame kernel: Hibernate inconsistent memory map detected!
Jun 08 09:01:02 kunta-mkFrame kernel: PM: hibernation: Image mismatch: architecture specific data
Jun 08 09:01:02 kunta-mkFrame kernel: PM: hibernation: Failed to load image, recovering.
Jun 08 09:01:02 kunta-mkFrame systemd-hibernate-resume[402]: Unable to resume from device ‘/dev/disk/by-uuid/566a2558-788b-4f16-891d-36279e0bb69e’ (259:2) offset 217257984, continuing boot process.
@Matt_Hartley Sorry to ping you, but I see in another thread as the only mention of a solution related to hibernate at all. Does anything I’m doing look weird to you, and what script are you using to hibernate? If you have confirmed the hibernation working on the ryzen AI platform with nvidia graphics in fedora, if I couldn’t adapt your script (which I’m not sure where it is) I would even be willing to switch to fedora. Any help on this is greatly appreciated.
My only hesitation to installing fedora is if the bug I’m experiencing is from the bios specific to this model, I would much rather not. On this page linux-docs/hibernation/hibernate-fedora-automatic.md at main · FrameworkComputer/linux-docs · GitHub my laptop is not listed under the “tested successfully,” which doesn’t inspire confidence.
Have you worked through the steps here to try to narrow down why it is failing? Debugging hibernation and suspend — The Linux Kernel documentation
I have not touched on hibernate for probably since like Fedora 42. It’s just not something I’ve had time to focus on.
You could open a ticket if you want and ask to have it escalated to me and I can take a look at what’s happening.
In addition to asking to have it escalated to me. Also, please make sure to include this thread for context.
Our officially supported primary focused area is of course suspend, which is different than hibernate.
1 Like
Thank you for the reply Matt. The only reason I do not want to use suspend is that when I had the laptop in suspend, it used what felt like to me an unusual amount of standby power. It would use over 10% overnight while sleep. I don’t see any indication this isn’t expected behavior, which is fine, but then hibernate becomes a necessity to preserve battery life when the lid is closed for extended periods of time.
Have you checked if the uuid and offset are correct?
This is a good place to start for sure.
The offsets are correct. The first thing I did before using offets was use a swap partition that was 42 gigabytes in size, versus my 32 gigabytes of ram in the system. Essentially no matter what I did there was no consistency to if it worked. The only thing I can gather is that the offset error is real, there is a small offset from the mapped memory when rebooting that causes hibernation to fail. Actually writing the hibernation image is consistently successful, but when attempting to load it, there is the offset present which causes the kernel to panic, give up, and just do a normal cold boot.
I’m not really sure if it’s some offset allocations the acpi subsystem is doing that aren’t consistent or what, but it only is successful when there is no offset, which is expected. I just have no clue how or why the offset is happening ultimately.
Have you tried shutdown mode for hibernation vs platform mode?