Status of Suspend for Librem 5

It has a completely different level of security, compared to Librem 5:

100% inspectable with standard optical microscope

which is why it’s black and white.

Why is it important to inspect an LCD screen with an optical microscope? What could one find?

Don’t forget, it is intended for different use cases, for other things than standard phone, SMS, MMS etc. And on top of that, as it states on their website: " Precursor is an open hardware development platform".


How is any of this relevant to: Status of Suspend for Librem 5 ?

There is an existing topic on the Precursor: and it would be best to continue the digression there.


Yeah no make sense :joy:
Well the Precursor(WIPHONE) as far i know the battery life on demand it is 5 hours.

How far is the development of suspend from becoming a reality (enabled by default) ? I ask because the summer approaches fast here and again L5 will be a burning phone.

I try it after OS updates, though no cigar yet, screen for me doesnt wake on push of power button.

@dos excellent insider information about the 18 sec power button press. where can we find this in the firmware source code or component documentation?

long live open hardware!

Here ( BATFET Full System Reset):


Hello dear L5 users, last time I was monitoring the link kindly sent by @joao.azevedo :

As you can see, all merge requests are now closed (the last one was merged 4 days ago), therefore I would like to know if you could see any new update on PureOS and significant improvement regarding suspend feature ?

Thank you very much !


This is not the whole story. There are several open suspend issues:[]=suspend


Ah ! thank you !

wow, it is amazing that you can read that text and then deduce that holding down the power button for 18 seconds does the right thing. are there any other important techniques revealed in that document?

You might want to check this out, too:


yes, i’ve been through that.

Hello all, just to say that I thank Martin Kepplinger very much for his (very) long and hard work on the suspend issues.
This man works hard, he updates the gitlab issues one by one, contacts linux teams at NXP, kernel teams, etc. and frankly even if he doesn’t see my message here, I thank him a lot for his serious and quality commitment.

I hope that he and the other teams will be able to solve the many concerns about suspending which is an essential feature of any smartphone, especially now when we talk a lot about energy issues and anti-waste of energy measures.


Recently I had to go to sleep without access to a charger for my L5. Luckily, the battery level was at 87% so I figured it would last the night if it would sleep too (systemctl suspend). It did, but not with a wide margin. After 9.5 hours I woke the phone from suspend and the battery level had dropped to 27%. That means it lost ~6.3% per hour during suspend. This is, of course, a lot better than the 10-12% per hour it usually drops when fully awake.

With this background I am curious if there are any hardware limitations e.g. battery capacity, CPU, GPU, storage etc. that determines the minimum battery drain per hour or is it only about software optimizations to reach the level of other modern smartphones?


From what I’ve seen, it seems that the large battery drain is simply due to the fact that Librem 5 currently defaults to s2idle and not deep sleep.
cat /sys/power/mem_sleep returns s2idle [deep].

There’s an old MR saying that suspend to ram is not supported but I don’t know what the current status is.

Am I correct in considering that current power consumption is simply due to s2idle and will get better once deep is supported?
Is there already a way to use deep or is it still unsupported?

1 Like

So what is the difference between s2idle [deep] and deep sleep?

Well, I think that the selected option is the one not within bracket but I could be wrong.

As for the difference between the two options, see the kernel description.

On my L5, pureos, with tlp installed, the current power consumption in suspend is 4% per hour.