What about if OSI focus on Modem then FSF to WLAN/WPAN for Libre Firmwares?
from ryf-x200
2 Likes
That could work too.
1 Like
Any specific reason people are averse to blobs so much? Given that the rest of the system is properly sandboxed, who cares what the modem is trying to do? It’s an uphill battle too, things get released and obsoleted these days way faster than they can be repurposed, so maybe all the effort should go into sandboxing. In an ideal world, sure, all components should be open source. But realistically this is not where we’re moving to at the moment.
2 Likes
Because you can stop producing e-waste and modifiy things to your behalf as reducing energy usage or finding/fixing zero-day-exploits. There are many things sandboxing alone is not enough. Let’s say we will get a 5G modem in 4 years that is free of blobs, it can be used a few years for Librem 5, but even on a Librem 5v2 or 5v3, because 5G will exist longer than 4G will. It is not required to get hands on the newest modem, only because it is free. So you have another 10-15 years to remove blobs of one 7G or 8G modem etc.
3 Likes
It is doubtful that you can fully sandbox anything that has access to a radio and which unavoidably has access to that radio.
3 Likes
Yeah, naturally since the modem is literally the outermost layer. But, I really wonder from a technical perspective what even a rogue modem can do, hypothetically.
1 Like
Marketing glitter aside, you know, there’s a certain threshold after which all the tech becomes slightly irrelevant. Because we as a species don’t evolve. Our psyches don’t evolve. So when a tool is good enough it’s good enough. 4G is more than good enough. The only thing that is hurting us all is that people (and companies in general) keep chasing novelty instead of productivity.
2 Likes
While I am totally fine with 4G, there are good reasons for 5G (and those are not the marketing things I do not care at all). For example you need fewer towers outside the cities to connect everything to the network.
However, that is not the reason why I wish a freed 5G modem. The reason is more practical nature: Whoever wants to reverse engineer it and develop a free software solution needs a lot of time to realize it. In additional at some point providers will just disable 4G, no matter what you want. Freeing a 5G modem allows it to run as long as possible without freeing another new modem. 4G would not be usable for a long time and makes Purism and other companies stuck with 4G modems, if they care about freedom.
So it’s just pragmatic to free 5G, nothing more or less.
3 Likes
Providing that the 5G modem also supports 4G (which I expect to be the case) and providing that freeing the modem includes all functionality of the modem.
As a small example … the law in my country requires a tower to record and retain metadata associated with communication from a mobile device but it does not require a mobile device to make such communication continuously so as to provide a complete record of the movements of the operator of the device.
A rogue modem could obviously subvert even this very limited human rights protection by intentionally making unnecessary communication so as to make the location tracking more complete. (I understand that anyone with a strong desire not to be tracked would use the off switch for the modem, in the case of the Librem 5.)
Of course, if you have a spyphone, you have about zero chance of being able to control when communication does and does not occur - because blackbox operating system software and applications just do things unpredictably and uncontrollably in the background.
2 Likes
You can use a Faraday bag or sleeve.
1 Like
Well I took some time to educate myself on the subject and those modems aren’t what I thought they were - they are instead complete solutions along with their own RTOS and a network stack, all that working independently from the main device/OS. The consequences are obvious.
5 Likes
Interesting post, @what !
So I went deeper to your post and asked to AI about it!
Here’s what it said:
Why modems are not completely open source
What a modem does (in simple terms)
Router = manages your internal network (WiFi, assigns IP addresses to devices, etc.)
Modem = translates the ISP’s cable/fiber signal into data that the router understands
Technical issues
- ISP certifications
ISPs (TIM, Vodafone, etc.) must approve every modem on their network:They certify that it does not damage the infrastructure
They verify that it complies with standards
They test compatibility and security
This is a costly and lengthy process (thousands of euros, months of testing)
→ No hobbyist can afford this for an open source project
- Proprietary chipsets
Modem chips (Broadcom, Intel, Qualcomm) have:Closed binary firmware (blobs)
Documentation under NDA (Non-Disclosure Agreement)
Proprietary algorithms for signal modulation
3. Complex standards
DOCSIS (cable), VDSL2, GPON (fiber) are extremely complex
It takes years of engineering to implement them
They cost millions in R&D
Is it possible to make one that is 100% open?Technically: YES, but…
Practical obstacles:✗ Prohibitive cost - requires a team of engineers for years
✗ Available chips - almost all require proprietary firmware
✗ Certifications - the ISP may refuse to activate it
✗ Complexity - standards are technical mazes
Existing attempts:
Partial projects:Some attempts to make open ADSL/VDSL modems → very limited
OpenBTS (for mobile networks) - similar but simpler example
LiteX - FPGA framework that could theoretically do it
Reality:No mature, working projects
Too expensive for open source communities
ISPs are not cooperating
Practical exampleImagine you want to build a completely open source car:
Engine = modem (ultra-complex, strict regulations, proprietary chips)
Interior = router (easier, in fact OpenWrt exists)
You can make the interior open, but the engine requires factories, emissions certifications, crash tests → impractical for a community project.Conclusion
100% open source modem: technically possible, practically impossible due to costs and legal barriers.
Realistic alternatives:
Use commercial modem in “bridge mode” (only passes the signal)
Completely open router (OpenWrt) behind it
This way you control everything except the “translation” of the ISP signal
Does this make sense? Would you like to explore any specific points in more detail?
What do you think?
1 Like
That information is very generic. We are talking here specifically about cellular modems and, in the case of the Librem 5, cellular modems that are available as discrete compatible modules.
2 Likes
I think is primarily important privacy and security about modem, mainly on Librem 5 that is created on those principles! No? 
1 Like
I think whatever LLM you were using got trained on a lot of corporate marketing, and delivered little of technical relevance.
Here’s a better (academically and environmentally) source Modem - Wikipedia
5 Likes
The Librem 5 is created on the principle that the cellular modem is blackbox and hence by definition can’t be verified and shouldn’t be trusted. Therefore a range of design concepts in the Librem 5 are used to mitigate that. The approach (theoretically) available through the project of this topic would be to make the cellular modem open source and hence verifiable, which would then suggest different design concepts in the Librem 5.
3 Likes
I don’t see how it matters whether it is a blackbox or not: Security through compartmentalization also works with free software. So is hardware upgradability. There is no need to design a different concept if the modem is free.
3 Likes
That’s right. There is no need.
But a libre modem would open the door to alternative designs. The current design has trade-offs. Those trade-offs could be revisited. Out of that “revisit” you can certainly make the case for no design change anyway but you can also make the case for design change.
However this is largely a theoretical discussion, based on two independent hypothetical events: 1. This project actually delivers a libre modem 2. Purism has the willingness and means to change the design.
2 Likes
It would super-incredible if @dos make a Libre Firmware for Librem 5 Monitor/displayUnit.
s.krzyszkowiak a full reverse-engineering capabilities. 
2 Likes
It looks like the project website (https://librephone.fsf.org/) has been updated with a bit more content and there’s even a link to the Codeberg project: https://codeberg.org/rsavoye/librephone.git.
The licensing is unclear to me whether it is AGPL (https://codeberg.org/rsavoye/librephone/src/branch/main/docs/LICENSE.md) or GPL (https://codeberg.org/rsavoye/librephone/src/branch/main/LICENSE). The Python source files show it’s AGPL-3.0-or-later.
2 Likes