Hacker News new | past | comments | ask | show | jobs | submit login
Librem5 Hardware Update (puri.sm)
292 points by spotman on Nov 10, 2018 | hide | past | favorite | 105 comments

It’s hard to overstate how important Librem phone is. Currently, it’s pretty much impossible to buy a high end phone that you can actually own. All Android based phones have their firmware locked and root access disabled. This effectively means that you do not own the device, and you don’t get the final say as to what will be running on it. I’m aware that it’s possible to crack the firmware, but I think that’s completely beside the point.

We desperately need an open alternative that puts control back in the hands of the user. I think open computing is of fundamental importance. If we lose the ability to decide what code runs on our devices, we’ll be moving a step closer towards a totalitarian dystopia where the governments and corporations get to decide what’s good for us.

> crack the firmware

In most cases, you don't have to crack anything, you can officially unlock the bootloader and flash your own OS build.

Yes, you can't touch the early boot initialization, you're still required to use vendor blobs for using wireless/modem/camera/etc things, but you can control most of the software that matters.

Yea using adb/oemunlock. Sometimes you have to get a code (like with Sony, you input your IMEI number into their website and get one). I've had one Sony that was unlockable; a condition of the original cell provider.

Still, it's bizarre considering you can install Linux on any PC/laptop you buy.

The bigger issue of course, is that you can't just install Linux on a phone. They're all random pins soldered to random chips and all use patched to hell Kernels with binary blobs. Linux distros can release images that are designed on boot on x86/x86_64/PPC/Sparc and they'll booth up and install on most of those machines. That's impossible with nearly all Android/ARM devices.

PostmarketOS is trying to change that from the other direction, but Librem is a huge step in giving developers embedded hardware that doesn't require carefully modified and patched kernels/bootloaders.

Yeah, I think the fact that you have to request your IMEI number and then use additional tools to unlock the device you bought is absurd. Nowadays, a phone is no different from any other computer, and you should be able to buy one that lets you run whatever you want on it without jumping through any hoops.

Some of them make you wait weeks/months before you get their approval to unlock.

> Still, it's bizarre considering you can install Linux on any PC/laptop you buy.

Apple is trying to change that with their new T2-line of machines where you are effectively locked out of using the device’s storage if you try to boot a OS not blessed by Apple (like Linux).

You don’t own your new MacBook or Mac Mini.


You can disable all of that and boot any code you like. For the 99% that don’t want bootkits to be possible on their hardware, you can require a signed bootloader. Best of both worlds.

Inaccurate. Boot yes, but you can’t access your SSD from Linux, so you can’t actually install anything on your (very expensive) internal storage.

For a Mac Mini with external storage that may be passable, but for a laptop like the MacBook that’s pretty much a no go.

This move by Apple should not be confused with generic UEFI secure boot with user-loadable keys, which yes, IMO is a good thing for most people.

Apple however has locked down the hardware in a way which only benefits them.

This move is actually what everyone feared Microsoft would do with UEFI (which they didn’t) and literally caused endless FUD and uproar on the Internet (and it’s still going on).

Now that Apple is actually doing it nobody bats an eye, because Hey shiny!

TLDR: I don’t get people.

Linux distro support for Apple hardware has always been pretty weak; I doubt many people run linux on apples anyway. They ship with a viable unix and better UI, and getting Ubuntu or similar to work well on them is an uphill battle. I imagine most “linux people” are buying thinkpads.

I run Linux on an old Macbook Air, and it's a great machine. But it's getting a bit long in the tooth (2GB RAM doesn't cut it anymore) and I've been looking around for my next laptop. That locking move by Apple ensures I won't be buying a Macbook this time, it will likely be a Dell XPS or Thinkpad.

>Still, it's bizarre considering you can install Linux on any PC/laptop you buy.

It is kinda not true. There are windows 2 in 1 and recent macbooks on which it is almost impossible to install linux.

> recent macbooks on which it is almost impossible to install linux

That's FUD, and you know it. You can disable secure boot on Mac and boot whatever you want.

You can boot, but not install, as Linux won't see the internal storage that's controlled by the T2 chip.

So Linux need some drivers or it's a fundamental restriction from Apple?

Drivers won't work. According to my understanding, any unsupported operating system (those which are not signed by Apple -- even Boot Camp doesn't use the Windows UEFI certificate so Linux's shim won't work) cannot access stuff protected by the T2 chip. So it's definitely not a driver problem.

And apparently disabling secure boot still doesn't fix the issue -- see the Phronix updates[1]. I only found one article claiming you can get past over all the T2 "security" features[2], but they explicitly say (at the end of the article) that they haven't actually tested booting or installing Linux.

[1]: https://www.phoronix.com/scan.php?page=news_item&px=Apple-T2... [2]: https://www.imore.com/no-apples-not-locking-you-out-linux-ma...

Could the "AppleSSD" driver for Windows [1] be reverse engineered?

[1] https://twitter.com/tperfitt/status/1060995265694449664

It's probably not impossible to develop a driver?? But it'd involve a lot of reverse engineering.

> Still, it's bizarre considering you can install Linux on any PC/laptop you buy.*

* Unless it's a 2018 Mac ( https://www.theregister.co.uk/2018/11/06/apple_mac_linux_woe... )....

The device tree is the magic feature for describing the hardware layout of a device, thereby moving "wiring"-related modifications out of the kernel: https://en.wikipedia.org/wiki/Device_tree

If you can't use mainline Linux and/or have no GPU acceleration that's no use at all.

It's sort of theoretically useful whilst being practically pointless unless you want to produce your own OS.

How do I get into a Linux userland with X11 and a touch driver on my Galaxy Note 8? The answer is pretty much "write it all yourself, quickly before the hardware becomes obsolete".

I hope that the librem5 gains traction for that reason alone - producing a critical mass of developers.

You can chroot into a normal ARM distro and run the X11 server directly on Android. Works pretty well unless you need non-ES OpenGL.

This still requires running Android, most likely with an old kernel and nonfree blobs.

Without some pretty strong UX like iOS this effort is not worth it. There is great UX in the FOSS world unfortunately.

Priorities sometimes differ :).

There is place in the market for reggae and death-metal.

You couldn't PAY me to take an iOS device as a main personal phone.

(although as a caveat: I might consider owning one to test cross-platform mobile applications)

> All Android based phones have their firmware locked and root access disabled

Xaiomi A1 and A2 come with stock android and official bootloader unlock on the website, takes 10 mins to unlock and install LineageOS. LineageOS root access is an option in the settings.

I don't really have anything of value to add to the conversation, but in case the Librem5 team is here I just wanted to thank you for your work and applaud your openness to the current status of everything.

Another project which is taking on a similar challenge and is similarly open is the Dragonbox Pyra.

> and is similarly open is the Dragonbox Pyra

The Pyra is taking forever and it's pretty much a one-two men's team, which is why it takes so long I guess.

It is taking forever, yes. Unfortunate as the hardware will be out-of-date by the time it is actually fully released. Still, it'll be fast enough for its purposes and will really be a triumph for Michael Mrozek. It's amazing that he's been able to do all this.

The problem is that there are now alternatives out there like the GPD-Win 2 and GPD Pocket, which are way more "open" (and actually using the mainline kernel now) than the Pyra can ever hope to be (The Pyra will use to use a number of binary blobs to function. Such as the GPU support...). Of course, the GPD may be inferior in some aspects, but they release hardware very quickly and the market is pretty much captured by their efforts.

And one more project is https://neo900.org.

It's dead, and I'm pretty sure all of the cash has been spent.

i wonder if smaller screen variants will ever be on the radar. i'm on a bootloader-unlocked, rooted, lineageos sony z5 compact (4.6" 720p display) that you'd have to pry from my cold dead hands. but i would love to buy a small hackable, privacy-first phone with open hardware instead.

i imagine pcb layout becomes a much harder and more expensive problem at small sizes, not to mention most people wouldn't touch "only" a 720p display in 2019, even at a 4.6" display size (unfort), even if you told them it adds 30% battery life.

ill be sticking with the xz1 compact 4life! (or at least until someone makes something the same size, waterproof, with a headphone jack, camera button, notch-less and a fingerprint scanner on the side. so not very likely the way things are going :)

I just sent mine in for (hopefully) warranty replacement, I've never had a phone I loved more.

Lots of power, easy to handle, and I don't think about battery life ever. If they don't replace it, I'm just going to buy another. I was very sad to see they increased the screen size and resolution on the new models.

I've been completely impressed with the Sony compact phones - I just got my second one a few months ago and I have no plans to move back to the larger form factor of current flagships

agreed. i just replaced the battery on my z1 compact, it's still fine.

I'm quite impressed by their company manifesto:


More companies should follow this example.

Indeed. This is an encouraging idea. But in a broader sense -- applied to a whole sector of corporations, for example -- I'm curious how legally enforceable this kind of charter is, under the cited statute or under other laws governing incorporation.

Under what conditions can articles of incorporation be changed? Can't their investors force such a change? If so, it seems a large burden must be placed on choosing investors who are somehow legally bound under similar rules.

And what mechanism of enforcement exists? Presumably investor lawsuits are the main mechanism, which again puts a large burden on finding investors whose goals are not returns maximization. If this state statute specifies a distinct tax classification, perhaps there would then be state-level prosecutorial or administrative enforcement.

While the Librem5 is step into right direction, it is far away from open hardware and will likely need even software proprietary bits - so I would recommend to stay in realistic zone and help the project in any way you can.

From their manifesto:

> The Corporation will release all software written by The Corporation under a free software license.

> The Corporation will release all hardware schematics authored by The Corporation under a free hardware license.

If you are talking about the IP blocks that are bought/licensed from third parties, then you are right, but I suppose that these are replaceable by equivalent blocks in future versions, without the user noticing. In other words, users don't depend on these blocks, so for all the user knows, these blocks are commodities.

I know the manifesto, I am among original ones proposing it.

And for things that might happen in future, I would leave them for future. Building hw takes time and proper dedication - all what I am saying is to be realistic towards current situation. The blocks are no commodities but some essential parts in today's use of smartphones. And no, things don't move that fast that you will just replace it.

If done right, things could finally go into right direction or at least have some new light on it.

> software proprietary bits

Only the DDR4 memory training blob (which someone could eventually rewrite).

Where are they getting the LTE baseband software from?

They're not of course. They just treat the modem as walled off peripheral black box like all the other rest of the hardware. They may load firmware on boot for LTE/Wifi/Bluetooth like most Linux laptops do. Depending on your opinion on the topic that may be free enough or not but it should be safe as none of those can access the CPU/RAM directly.

I see. The question here is not if there will be proprietary software: there will be. The question is which of that proprietary software runs on the CPU.

Is the idea that all of the software in the kernel to interface with these devices is open source?

I also don’t get the hangup over PCIe (vs USB). DMA with an IOMMU can be made fairly secure (and has obvious perf benefits).

> I see. The question here is not if there will be proprietary software: there will be. The question is which of that proprietary software runs on the CPU.

I'm pretty sure no closed source software runs on the CPU. As for the hardware that depends on your definition. Hardware often has firmware in ROM. That you now feed that firmware on bootup instead of it being in ROM just allows you to update the firmware from the manufacturer. If that's running proprietary software depends on your definition.

I think the main reason is that powering off a device (for their hardware kill switches) connected via USB is more reliable than powering off a device connected via PCIe.

There's also just more layers of security when using USB as opposed to a single layer with PCIe (IOMMU -- which is secure as far as I know but I'd prefer to be safe rather than sorry in this case).

Besides IOMMU bugs, which silicon vendors often subtly add to their errata too much for my comfort; an IOMMU adds security against DMA attacks in theory. I specifically say in theory because oftentimes vendors either don't configure it correctly or leave it totally unconfigured. Additionally, things like multiplexing on the same bus further complicates things

It depends on the board design. It costs more, but you can have controlled power line for a single device, controllable with a gpio.

Edit: it has a "radio hardware killswitch" ; so it's well designed.

That makes sense.

I thought of the kill switch thing after commenting… eGPUs (necessarily) have the ability to be unplugged while running, but I think it was a lot of work on the part of the OSes to make that work well.

What would be a benefit to putting the modems on PCIe bus? I don’t see any positives, only drawbacks.

Reminds me of a joke I heard:

What’s the difference between a person and a cellphone?

A person only has two arms.

They don't have the baseband on the same chip, the modem is connected over something like USB (and there's probably gonna be a physical kill switch).

IIRC this is fine even under FSF "Respects your Freedom" rules

Kudos. The most interesting thing in smartphones for my needs since the Nokia N900, not counting the SailfishOS efforts. I'll buy one if it works "a la Maemo" (Debian derivative).

Their OS, PureOS, is a Debian derivative.

> Almost right after that was a Chinese holiday, the Golden Week, which is in practice a two week holiday.

This doesn't match with my experience. Golden Week is a one-week holiday that gets you less than 5 days of vacation, because -- although companies are legally required to give Monday through Friday off -- the norm is to work on weekends around the holiday to make up for the lost time.

Does anyone have an idea what sort of Golden Week implementation Purism ran into? Why is it so different?

My experience with ODMs at Shenzhen was that it's usually ten days for engineers. Could be down to one week when production was halted. But yeah that means workers only had 7 days. I didn't witness five days though. And they shut down power for the whole building, so people couldn't really get there early

I'd be interested in hearing anyone's experience with holidays and China B2B.

Is it just kind of known that things stop there for certain weeks? Or are these formally set out in schedules?

My only experience has been B2C, where the experience is generally {no response} -> {a week later: "Oh, everyone was on holiday"}

Well everyone there has the same holidays, and yes they mostly assume everyone they work with are well aware of it. So just ask them nicely every three months what are their next holidays

China is a very big country.

True enough. What Golden Week practices have you encountered?

For the last few months I have been super excited about the Librem 5. I check way too often for their updates.

Maybe I’m alone but I wish I had more visibility into their progress. Might not be helpful for them though.

I also wish they would embrace Vala as the language for the apps. ElementaryOS has had pretty decent success so far building an app ecosystem and the choice of Vala has contributed to that

I really want a Linux phone that has the same kind of specs and build quality as Samsung flagship phones.

I’m not so worried about that any more. But if the premium hardware is what you want I’d support postmarketOS.

Hope you're willing to pay $x0,000 to amortize the design / code / build cost over a smaller-than-Samsung userbase then.

Does it have a hardware switch to turn off the mic and camera?

Yes. You can see the switch on the side render here: https://puri.sm/posts/librem5-roadmap-to-imx8/

Those are just mockups and not the end product which might heavily change due to various things.

Of course. But you can see three switches on the side of dev boards which would seem to indicate that this part of the plan hasn't changed. (yet)

Ah that's great!

When you think about it, it amazes me. So many people, including me, would love to have these switches. Like physical switch for the mic, camera and radio. And yet, the companies that are supposed to listen to the customer seem to do exactly the opposite to what the customer wants, like installing all these unremovable crappy apps nobody uses or removeing the headphone jack. Are these people so out of touch with reality?

Hold up.

I don't have stats but I'm incredibly confident that almost no users want hardware switches.

Does your mum want a hardware switch? Does the guy working at Starbucks?

I think it might be the techheads who think that installing Linux and having hardware switches are normal - they're the ones who are out of touch with reality.

Companies are basically dictating what users "want": from social media to personalized ads.

The problem with regular users is: they don't care until it's too late.

Can we blame them? Not really, because life is too diverse to worry about every consumer product.

It's good, however, if there is a bunch of good guys who do the thinking for these users. Still, most users may not care, but eventually consumer advocacy groups will pick this up, and they will educate users.

> The problem with regular users is: they don't care until it's too late.

But it shouldn't be that way, should it? At least it's not what I learned during my marketing lessons. You do market research, listen to your potential customers, discuss the results and proceed based on that.

In the mobile world you have an absurd situation that you remove an important component of the device, you release the device, everybody tells you it was a very stupid decision, you ignore everybody and still rake a ton of cash. Amazing.

I mostly see what you're saying here

The thing is, if these things resulted in increased profit over the medium term, they'd probably have happened already.

So yes there's user-hostile decisions made that put profit over users. Like ad tracking.

But there's other things that I think convince a user to purchase a product, and purchase the next product from the same company (loyalty maintained) that may not be in the user's best interests.

An example is a larger battery. My current theory is a manufacturer could easily have a 2 day battery and it wouldn't make for a terribly unwieldy phone.

But they don't, because customers are going to buy the phone next to it that is sleeker and weighs less.

I think customers are more often than not short sighted when making decisions like this. When they're in the store they think "but this one's so much nicer, and the battery probably lasts enough"

I think purchases are often made emotionally, not rationally, and a phone with 7 extra switches and more weight or thickness just seems clunky and less "high tech"

"I don't have stats but I'm incredibly confident that almost no users want hardware switches."

How "incredibly confident" can you be about this claim without published stats?

I would like hardware switches ala Purism's laptops but I see that Hacker News prioritizes steering discussions away from these switches (such as down-moderating the grandparent post for making a similar claim with no pointer to evidence but leaving your evidenceless post relatively highly moderated). I'm not sure if other people would have thought about having such switches therefore I'm not sure if they would value them or even consider asking vendors to include such switches.

As you say we both provide no evidence, but I suppose my claim aligns better with HN user's anecdotal evidence

It's true I could be wrong, it might be something really unintuitive and for some reason people want hardware switches and I've just literally never heard it until now.

But in my experience 99% don't even know about, let alone care about advanced features. Even simple features you and I take for granted. They want to launch Instagram. That's about it.

Anecdotally, I do see a large number of lay people with tape over their laptop cameras. Enough stories have filtered into the press about RAT malware and whatnot that there is likely to be some endemic insecurity about a camera pointed in your face across a fairly broad section of the populace.

Yet those tinfoil folks who cover the camera do nothing about the microphone. Since ~2009, it’s been impossible to turn on the camera on an Apple laptop/monitor without the light turning on, yet there’s no indication of something using the mic.

> Yet those tinfoil folks who cover the camera do nothing about the microphone.

You say this like it's a "gotcha," but what exactly could somebody with no technical knowledge do about the microphone? If you're trying to say that if there were an equivalent to electrical tape but for the microphone, people wouldn't use it, then I don't even think you honestly believe that.

Also, it's bizarre that in a world that ratting has existed, you'd compare covering the camera on your laptop with a piece of electrical tape to paranoid schizophrenia.

> but what exactly could somebody with no technical knowledge do about the microphone

Take an old pair of wired headphones with mic, cut the cord off, plug in connector. There’s better solutions, but that’s the easiest if you are nervous.

> paranoid schizophrenia

I made no such comparison

To clarify for the down voters, I didn’t mean to insult, I meant to imply it was an unfounded fear (as most tinfoil conspiracies are). Most folks I’ve talked to who do it, didn’t know about the impossibility of turning on the camera without the light on current gen hardware. When I told them about the very real risk of listening to audio without any similar visual indication they were rightly alarmed. I’m actually calling for the ability to either have a hardware mic cutoff or at minimum a visual indication similar to how the camera does.

Regardless, that average people are cautious about their electronic privacy is something to applaud, not discourage. If you are trying to train people out of fearing their cameras, then you're doing a bit of a disservice.

In particular, your claim is false for phone and tablet cameras. These devices do not have visual indicators of camera operation. Furthermore, the only reason for an industry-wide adoption of laptop camera LEDs was consumer fear of them. If that fear erodes, there is no incentive to maintain the hardware feature and no governmental regulations governing such hardware.

> your claim is false for phone and tablet cameras

Yet nobody seems to care about them (unless forced to care by their employer). The exact same folks I see in work environments with the tape on the Apple laptops are the same folks with nothing on the phones/tablets (and previously mentioned mics, where as at least on iOS you always have a visual indication of that). If you really care about privacy, you learn the different ways to protect yourself. If instead you have a different motive, you instead just grab a piece of tape and call it done.

> I'm incredibly confident that almost no users want hardware switches.

I'm incredibly confident that almost everyone does. They'd explain it as the ability to turn off the microphone or camera without someone else having the ability to turn it back on without asking you, but it amounts to the same thing.

One thing I know nobody is asking for is microphones and cameras that can be turned on without their consent after they've explicitly turned them off.

> Does your mum want a hardware switch?

Actually, she does. She read some news and put some tape over the camera of her notebook. Just as Zuck does - and I imagine there is an enormous gap in technical knowledge between the CEO of FB and my mum. So yes, maybe the demand is higher than the marketing departments are willing to admit. Apple's years-long reluctance to cater for people looking for a phone with two SIM card slots also comes to mind.

I mean, maybe. I freely admit I have no evidence other than the anecdotal.

It's just generally been my experience that the vast majority of users don't care about privacy issues or "freedom" the way techies do.

im Pretty sure it does

As an expat living in HK, I did not quite understand the explanations about delays at the beginning of the article.

The golden week is Japanese, not Chinese. Nothing closes in China at that time. And it’s in may/april.

Mangkhut did not really hit China mainland, it was already mainly just a storm when it reached mainland, and it died in 1/2 days once leaving the sea. I guess most of the factories are near Shenzhen, they should not have really suffered from Manghkut.

Just wanted to rectify that...


>The "National Day Golden Week" begins around 1 October.

>Three days of paid holiday are given, and the surrounding weekends are re-arranged so that workers in Chinese companies always have seven continuous days of holiday.

Any experiences with their laptops?

Typing to you from a Librem 13 that I've owned for ~3 weeks. So far, the hardware is stellar. I can't say the same for PureOS, I replaced that after ~36h (never made it to desktop). Battery life is consistently ~8-9h. The machine is more powerful than the desktop I replaced it with -- that box is basically NAS now.

Would I buy again? Yes, for the ethos. Would I recommend it to a casual user (the intended audience)? No. It's still very much a techie device. If I push the power button and don't see desktop within a minute, that's broken. You need to be comfy at the command line to maneouver -- which I notice has been almost entirely convenienced away in Fedora 29. If Librems shipped F29, I could at least recommend it for my partner who is the definition of a casual computer user.

I like mine (a 13). Hardware seems very solid, battery life is good. Screen, keyboard, touchpad are fine for me (some people are very picky about these things). Having the kilswitches for mike/camera and wireless/bluetooth are nice. I was on a 2013 13" MacBook Pro previously (an amazing piece of hardware as well) and can't say I miss it.

The only thing I miss actually is the 16x10 aspect ratio of the MBP, where the Librem like all other laptops I know of is 16x9.

Software-wise I've stuck to PureOS (which is Debian-based) despite temptations to install Arch or something. It's worked fine for me.

>18650 battery

I am tired of subpar battery life being a de facto industry standard

I am almost certain that battery only applies to the dev board for testing and development. There is no way a 18650 is going to fit in the final Librem 5 product.

Sadly, I think you're right. Hopefully they manage to go with some sort of standardized (cheap and easy to replace) battery for the final product.

Sadly? I wouldn't really want a phone that had a battery physically larger than a AA battery.

Why is there a substantial heat sink on what's supposed to be packaged down to the size of a smartphone?

Because it's a platform under active development, and they're probably more concerned with functionality than some super sexy tiny cooling solution? At some point they'll most likely test the production cooling system, but for now it probably doesn't matter. You'd be surprised how many things (cooling, PCBs, etc) are much larger early on than their production counterparts...

Software perspective: Fatal error, I'll just make a change and recompile.

Hardware perspective: CPU cooked itself because frequency regulation was accidentally turned off, time to wait for new parts from China + assembly.

It always amuses me how laisez faire folks are about things like ESD. "It's never caused a problem before" isn't a reason to ignore best practices and not work defensively.

> "It's never caused a problem before" isn't a reason to ignore best practices and not work defensively.

It is if working defensively is costlier than replacing the odd broken part.

So the i.MX 8M SoM package will be smaller and dissipate less heat in the final handset?

I wouldn't expect the same chip going into a smartphone to require a heatsink or any thermal consideration at all on a sparsely populated, open-air circuit board.

It could be that power management is not enabled yet, or perhaps other thermal protections are not fully enabled yet. I doubt the SoM package itself will get much smaller, but I'm just basically guessing since I'm not an employee of Purism.

Hate to be the party pooper here but does anyone believe this thing will ship in January?

Applications are open for YC Winter 2022

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact