This seems like a basic feature, especially for a laptop, but the MNT stock image defaults to disabling both suspend and hibernate, and I’m not anxious to re-enable them until I’m sure why. Additionally, if I run the reform-suspend script, it complains that that only works on the MNT Reform 2 with i.MX8MP module.
Surely there is some way to put it into some sort of power save mode besides completely shutting it down? There is a “Wake” option in the keyboard menu, which must do something.
The reason for that is that the only platform where suspend+resume is known to work is i.mx8mq classic reform.
resume from suspend will not work
The reform-suspend script used to be a script installed into /lib/systemd/system-sleep but was removed from there in early 2022. Maybe you mean reform-standby? And that script should complain that it only works on the reform 2 with i.mx8mq not the i.mx8mp.
This used to work also on the pocket reform with i.mx8m+ but that was with a very old kernel (i think it was 5.12?) and it was before the first units started to get sold. These days, suspend/resume only works on some i.mx8mq boards.
Yes, that is for i.mx8mq classic reform. I cannot tell you how it found its way into the pocket reform keyboard. Maybe from back in the day when i.mx8m+ suspend/resume worked?
@minute will be able to shed more light on the historic details of this than I.
So… is this all to say that there is no sleep mode for the RK3588 version of the pocket reform? I was trying to figure out how to make this machine sleep myself, since the portability of it is greatly hindered without the ability to do so.
That is correct. If this is important to you, please consider helping to make at least hibernation work. Only few blockers seem to be left for hibernate to work and on other rk3588 platforms there are reports that it already does work. There are some PCI reset patches which are getting tested.
Adding the inexpensive larger battery mod and using a powersave mode works well for extending the battery life until hibernation can be implemented. If you power off the screen it works well as a pseudo-sleep, though I wouldn’t leave it running in a hot car like this.
Boy, if I knew how to even begin helping with something like that, I would love to contribute. Is there an ongoing thread about the issue I can look into?
My RK3588 Pocket Reform also has a “Wake” option in the oled keyboard menu, at it appeared after upgrading to RK3588, I am quite sure it didn’t have it when I was using the i.mx8m
shigeru@mntpr-0:~$ sudo reform-check
I: Contents of /proc/device-tree/model: MNT Pocket Reform with i.MX8MP Module
I: `uname -a` output: Linux mntpr-0 6.16.3-mnt-reform-arm64 #1 SMP PREEMPT Debian 6.16.3-1+reform20250830T222127Z (2025-08-3 aarch64 GNU/Linux
I: Version of linux-image-mnt-reform-arm64: 6.16.3-1+reform20250830T222127Z
I: Version of reform-tools: 1.78-1+reform20250901T183925Z+1
I: Version of system image: System Image v4: 2024-08-20
I: Version of system controller firmware: 20250623
I: Version of keyboard firmware: 20250623
I: probably booting via /boot/boot.scr (/boot/extlinux/extlinux.conf does not exist)
I: Mount source of /: /dev/reformvg/root (LVM vg 'reformvg' on LUKS device 'reform_crypt' on SSD)
I: Mount source of /boot: /dev/mmcblk2p1 (eMMC)
I: the following files differ from how they are shipped by reform-tools (ignore /var/lib/alsa/asound.state):
??5?????? /var/lib/alsa/asound.state
I’m am trying to give you options. With a closed hard+software projects, when something does not work your options are: deal with it and do nothing. With this hardware, at least we can fix things. And even if you are not into kernel-level development (I’m not either) there are a lot of things that you can do if you find a part of the system lacking. Reporting these things as problems is a valuable contribution as well!
To be fair: there are a lot of computers you can buy these days where you cannot even install a FOSS operating system on to begin with. And on those where you can, not all parts will work even though you might’ve paid much more than 1500 EUR for a proprietary system.
And to be fair: you paid for the hardware. The software was available to you free of charge to begin with. MNT didn’t have to give you the software that runs well with it but they did and continuously invest vast amount of time to make it work better. And in contrast to many other Linux compupters out there, most parts work. Unfortunately suspend/hibernate does not (yet).
I’m sorry my message offended (?) you. I was trying to give you options instead of saying: well, it doesn’t work, you cannot do anything. MNT is very good at collaborating with volunter contributions (ask me how I know) and I’d recommend it to others who are looking for fun projects to spend their free time in, yes.
Could you elaborate on how you are selecting a powersave mode and what method you are using to turn off the display, such that it is doesn’t glitch out and not come back on when you want to resume use?
I think a powersave and turning the display off are great ways to reduce the powerdrain such that it extends the battery life some, which is good.
Josch, you’re a super star around here so I am not disagreeing with you. The point that really matters here is that MNT products are truly yours to do what you want with, but you need the know-how to do that.
However, I just wanted to point out that ALL of the Linux hardware companies that sell laptops have hardware that can handle standby, and to my knowledge all of that laptop hardware does that.
So, if the hardware that is being used on MNT products does not support suspend and this is just the reality of a completely open product, that is a different story.
However, I feel that MNT is on the hook to clearly define this. Finding out AFTER you have purchased a product is not that great. And if this hasn’t been defined yet, because there is still hope for hibernation to work on the RK then that is also awesome, and I REALLY hope there is success there.
For those wondering suspend is a very low-level function that must be supported by the hardware in question. Its implementation varies between hardware, and it is not something that is super easy to correctly accommodate for. I’m mentioning this because the task of creating a truly open piece of hardware is truly enormous! MNT has made giant strides in this effort as well.
That said, the Pocket Reform is awesome, especially with the RK upgrade!
Honestly I haven’t touched this since I upgraded to RK3588 and replaced the batteries with larger ones, as that resulted in adequate battery life. I used to use the default GNOME powersave controls and screen blanking with the iMX processor, but these don’t seem to work at present. Does anyone know how to get the GNOME powersave settings working with RK3588?
I installed cpupower-gui and it allows tweaking the power settings, although the default is already set to powersave. I was able to save a small bit of power here (but not much) by disabling extra cores.
I haven’t tried, I was only testing for the sake of determining possible ways to reduce power use. The power savings from disabling the cores wasn’t all that significant, so I’d only really recommend it if you absolutely must squeeze out every minute of power from the system, especially since it comes with some serious disadvantages (fewer cores available for quickly handling high demand). I guess the RK3588 must already be fairly efficient at minimizing power draw from unused cores. It would be better to start with low hanging fruit like lowering screen brightness and disabling wifi/Bluetooth, unused services, and keyboard LEDs.
I agree with what you said, but the MNT store pages and promo videos don’t give any indication of these issues and give the impression that this is an easy system for anyone to use and enjoy — it’s not until you start looking at the forums and so on that you realise what you’re buying into. And I love my PR — hibernate (and maybe a couple of sound quirks) is the only thing I feel like is missing now.
I hear you and I also really want this feature to work. When we shipped imx8mplus Pocket very initially, sleep/resume did work, but something got broken in kernel updates and I didn’t have time to go back to fixing it. I’m working on debugging hibernate on RK3588. Sebastian Reichel at Collabora has recently noticed that suspend/resume isn’t working and I expect some patches to land from that direction, too.
Also for the Quasar project, sleep/resume/power mgmt is a main priority from the beginning. I really want to get this right.