Progress update from the Librem 5 hardware department:
[...] Making a non-Android mobile phone that will run an FSF-approved OS that supports all the features that we've all come to rely on (cellular, WiFi, Bluetooth, touchscreen input) has and continues to involve a lot of pathfinding, given that a RYF phone has never been attempted before and discovery involves solving issues as they come up.
The industry offers all the hardware to create a smartphone on a fast path, as the SoC vendors typically provide the modem (cellular and wifi) integrated directly on the SoC. Like a recipe in a cookbook—take an SoC, place it on a PCB, add RAM and flash chip on top of it (called a package on package—PoP), add antennas and finally power. The difficulty comes down to the firmware and the software that run these devices. The necessary firmware to operate the cellular modem, WiFi, BT etc. is provided by the chip maker, including the drivers for the GPU and more. The firmware and software included is proprietary with no source code with little to no alternatives.
Related Stories
Purism has finally released the specs of the up and coming Librem 5 smartphone!
Librem 5 Specs
What are your initial reactions?
Though I'm not a hardware expert by a long shot, I'm not incredibly impressed with the specs. I do feel that smartphone hardware has been "good enough" for most uses for a while now and I know they have to start somewhere. What has not been good enough is freedom, flexibility, and, you know, actual ownership of the device. Sure you could get some level of freedom by jumping through a bunch of hoops, but who has time for that? Also, in case you haven't been paying attention, most of the work-arounds are becoming more and more difficult, if not impossible to implement.
Like it or not, smartphones are the way most people interact with computers, and beyond the basics for survival, are probably among the most important of our possessions. I want devices that I control and I want my kids and grand-kids to live in a world where they don't have to be the "product". In the wake of so many failed open-smartphones, is there any way Purism has a shot?
Previously:
Lockdown Mode on the Librem 5: Beyond Hardware Kill Switches
Librem 5 Dev Kits Are Shipping
Progress Update From the Librem 5 Hardware Department
(Score: 1, Insightful) by Anonymous Coward on Tuesday September 11 2018, @11:13AM (3 children)
This cant come out soon enough imo. Im so tired of every phone being otherwise cool thing controlled by an adversary.
(Score: 0) by Anonymous Coward on Tuesday September 11 2018, @11:43AM (2 children)
There will always be plenty of fresh side channel attacks.
(Score: 2, Informative) by Anonymous Coward on Tuesday September 11 2018, @12:32PM (1 child)
At least it wont come preprogrammed with this stuff. Did you read that wikileaks where google was (not even being paid to) run a project to show the optimal search results, hashtags, etc to people to brainwash them into voting for hilary clinton? All based on their phone data... Thats what they are doing for free, think about whats going on for business reasons.
Its the same reason multiple people I know started saying "bitcoin is dead" recently even though its selling for insane prices. They got brainwashed by some fake news facebook/news/whatever feed that someone paid for to take their money.
(Score: 0) by Anonymous Coward on Wednesday September 12 2018, @03:24AM
Where's the black helicopters mod when you need it?
(Score: 2, Informative) by pTamok on Tuesday September 11 2018, @12:04PM (7 children)
Reading the article shows that Librem have been unable to avoid using non-open software in the cellular modem.
Librem are doing their best to isolate the modem, so that it does not have direct access to RAM, but is instead accessed over an internal USB bus, but there is nothing to stop undocumented over-the-air interface access to the modem accessing undocumented capabilities.
(Score: 5, Interesting) by mth on Tuesday September 11 2018, @12:43PM (1 child)
It is possible to use a whitelist for USB devices. So if a modem suddenly decides to become a keyboard, the OS can ignore it. That limits the possible attacks considerably.
It would still be possible to eavesdrop on communication, but that is possible anyway at the telco. You would have to make sure communication is encrypted at the OS level, before it even reaches the modem, but that's a pretty basic requirement.
Denial of service by turning off the modem remotely is also theoretically possible, but again this type of attack could also be done at the telco.
Having a USB port exposed to an untrusted device is an additional attack surface. I don't know how secure the average USB stack is. I remember there being a PlayStation 4 exploit some years ago that used a weakness in the USB stack.
With the likely alternative being no libre phone at all, I think a black box on a USB bus is an acceptable compromise.
(Score: 2) by mth on Tuesday September 11 2018, @02:58PM
I misremembered: it was actually the PS3 that was exploited via USB.
(Score: 0) by Anonymous Coward on Tuesday September 11 2018, @05:55PM
In other words, patents and proprietary software thugs are screwing us out of our freedoms again. Yet, when Free Software has an issue or doesn't have some feature, many instantly blame the developers. Sometimes that is warranted, but in cases such as this, our corrupt political system and the proprietary software companies are directly responsible for the issues facing Free Software developers. I wish people would remember that next time they decide to sneer at Free Software.
(Score: 2) by jmorris on Tuesday September 11 2018, @08:50PM (3 children)
Putting the modem on USB is a mistake. My LG talks to the modem over a direct bus with i2s on the side to let the modem do audio encode/decode. The modem can't attempt very many shenanigans over such a connection so no real worries. We have to accept that for regulatory compliance issues the modem is going to run a blob. So wall it off and get on with opening up the rest. We desperately need an ARM SoC with open graphics that isn't an obsolete gimped Pi.
(Score: 2) by mth on Tuesday September 11 2018, @10:02PM (2 children)
According to TFA, they picked the i.MX series SoC specifically because there are open GPU drivers for it, from the Etnaviv project.
(Score: 2) by jmorris on Tuesday September 11 2018, @10:57PM (1 child)
Last update was eleven months ago. Yeah I'd plan a product launch around them coming through and delivering the goods.
(Score: 3, Informative) by mth on Wednesday September 12 2018, @01:29AM
Last update of Etnaviv, you mean? As far as I know several people are still working on that, but since the parts have been upstreamed, the updates are now spread over different repos. The kernel driver is in the mainline kernel, for example, other parts are in Mesa, libdrm etc.
(Score: 2) by RamiK on Tuesday September 11 2018, @12:20PM (6 children)
The bottom shouldn't have ports and you should leave one side clear from buttons, ports and switches. The reason is that people like propping their phones on their chests when laying back or at their laps while sitting in vertical and horizontal positions depending on content (browsing / gaming / watching films / reading) WHILE cables are connected (power/headphones). So, if the speaker and ports are at the bottom, they'd surf with the phone upside down and then need to flip when a call comes through. And if there buttons and switches at both sides, they'll accidentally trigger stuff incorrectly when propping the device horizontally.
There's also issues with docking stations if the cables are coming in from the bottom...
If short on room you can put the card slots on the bottom bellow the mic and led.
compiling...
(Score: 0) by Anonymous Coward on Tuesday September 11 2018, @12:24PM (3 children)
Is that really your priority for this phone?
(Score: 2) by RamiK on Tuesday September 11 2018, @12:45PM (2 children)
Priority suggests needing to compromise something in order to get better agronomic. That's not the case. You can retain the same features and have the same ports, switches and slots without sacrificing usability by arranging the PCB differently.
compiling...
(Score: 0) by Anonymous Coward on Tuesday September 11 2018, @01:06PM (1 child)
I'm sure there was some strategy applied to the arrangement. Eg in this case they need to put the hardware switches in a convenient place.
(Score: 2) by RamiK on Tuesday September 11 2018, @02:13PM
What's convenient about them being buried in your palm? At best, they're deliberately made inaccessible to prevent accidental switch-offs: https://puri.sm/wp-content/uploads/2018/06/2018-07-26-dev-kit-blueprint.jpg [puri.sm]
But if that's the case, why not put the switches on the top and everything else on the left? This way you'd still get the bottom and right clear and the switches separated from the buttons. And you could still put the SIM and microsd tight on the right/bottom if you can't fit it all on the left since those aren't extended with cables...
Really, lots of way to go about this better.
compiling...
(Score: 1, Funny) by Anonymous Coward on Tuesday September 11 2018, @03:33PM (1 child)
As long as the crop is plenty and healthy, who cares how questionable is the agronomy?
(Score: 3, Touché) by RamiK on Tuesday September 11 2018, @04:43PM
Ergonomics... The sad part is that I learned Latin and when the spell check corrected argonomics to agronomic, I was wondering where did the etymology go from farming to comfortable design.
compiling...
(Score: 3, Informative) by Unixnut on Tuesday September 11 2018, @12:28PM (3 children)
Seriously, they are not the first to try this. There is a whole slew of projects attempting an open-hardware/software phone. From the Openmoko years back, to the current Nokia N900 type phone (https://neo900.org/), the GTA04 open phone hardware (http://gta04.org/), to the "freephone", and others I have probably since forgotten.
What makes these guys special? And more to the point, why are there so many projects that don't seem to work together at all. If all the "open phone hardware" projects actually got together and worked on a single motherboard spec that they could all build upon/use, they might reach critical mass to actually be able to bring something to us to buy.
What happens instead is they all go and do their own thing, and then struggle to find the funding and enough people willing to buy the phone to be able to pay the upfront costs to build it, resulting in either the project stalling and ending up as vapourware, or abandoned, or the per unit cost being so high (up to €1000 for the neo900 for example), that it becomes unaffordable to all except the richest geeks out there.
Constantly forking and doing your own thing works with open source software, because the "cost" of replicating bits/bytes is so small to be negligible, so everyone can do their own thing, fork/merge all they like and let the ecosystem develop and see which forks end up popular.
However this does not map to the real world, with physical hardware. The costs of circuit design, testing, signal analysis, regulatory approval (for things that transmit/receive like phones, it is mandatory, and expensive), PCB design (and manufacture), and having all this soldered together is a huge upfront cost.
I think everyone going off and doing their own thing just isn't going to work out, and so far it seems I am right. I've been waiting since my old n900 died round 2012 for some kind of replacement "pure Linux" open phone I can hack on and not feel like I am being constantly spied on and "monetised" by some corporation. 2019 is around the corner, and I am still waiting, with nothing but promises of "real soon now" by all the separate projects.
(Score: 2) by urza9814 on Tuesday September 11 2018, @02:45PM
The Neo900 (from 2014) is an update of the GTA04 (from 2012), and the GTA04 is an updated version of the OpenMoko (from 2007 -- OpenMoko was code named GTA01 and GTA02). And a quick search did not find the Freephone. So of the ones you named which I can identify, they're all essentially different generations of the same project. And they're not fully open source, which is what the Librem is trying to accomplish. And the Librem can't use the same hardware because they need to do things like isolating the modem chip, and they're likely picking different hardware anyway since they've got pretty significant differences in their objectives.
Also, it seems that Purism will have the first open source phone release in *three years*. You are remembering the names but forgetting how freakin' old some of these are. A lot of modern smartphones don't even last three years:
https://en.wikipedia.org/wiki/List_of_open-source_mobile_phones [wikipedia.org]
Nobody bought them, the projects all but died, and now we've got a new contender that's trying something a little different and hoping it'll stick this time. Sounds like a reasonable plan to me. Yes, there's value in building on the work that's already done...but at the same time, I don't see much point in taking the same old board and trying for at least the *fifth time* to get that sucker to sell.
The other projects were trying to imitate the big guys. Take the same chips, the same components, drop 'em on a board with a few customizations and focus on the software. But if I want open software, it's probably cheaper and easier to just flash LineageOS onto a Galaxy. Librem is FINALLY building hardware that we just can't get from anyone else. It's the first "open smartphone" that I know of where "supporting open source development" isn't the sole reason to buy. That's a pretty big deal IMO.
(Score: 2) by c0lo on Tuesday September 11 2018, @03:40PM
What are you waiting for?
Come on, quick, start a project to design one!
(grin)
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 2) by PartTimeZombie on Tuesday September 11 2018, @09:08PM
I had forgotten about that Openmoko device. I so wanted one, it looked so futuristic and sleek, but I couldn't really afford one at the time.
It looks like it has died, which is a shame.
I am of the view that the more people who try this, the better chance someone will succeed. It doesn't need to be a mass-market hit like the iPhone, but just a supported, well made phone I can use.
I am wondering if they are using the Raspberry Pi way, making compromises to get usable hardware into customer's hands?
I would buy one, if they can get it to market for a reasonable price.
(Score: 0) by Anonymous Coward on Wednesday September 12 2018, @12:33AM
So I'm not holding my breath. I want open hardware so I can run a trusted operating system and I don't want a baseband with memory interface/Hypervisor access to the SoC. While the Librem promises to deliver, what I don't want or need are sweet promises. Show me the hardware when it's ready and we'll see if this is a US IT product deserving of my trust - or just another shiny gadget that's just a tiny little bit less shit than the competition.