Librem 5: First Impressions

  Platform Model                            Life  Size Efficiency
  
  iOS      Apple iPhone 13 Pro Max          21.7  4352       4.98
  iOS      Apple iPhone 13                  16.8  3227       5.21
  iOS      Apple iPhone 13 Pro              16.6  3095       5.37
  Android  ASUS ROG Phone 5                 16.6  6000       2.77
  Android  ASUS ROG Phone III               16.5  6000       2.75
  Android  ASUS ROG Phone II                16.2  6000       2.7 
  Android  Samsung Galaxy S21 Ultra (S888)  15.9  5000       3.18
  iOS      Apple iPhone 11 Pro Max          15.6  3969       3.93

I absolutely love Calyx, and wouldn't move from it. However, the battery life even compared to my old iPhone 6S is pretty bad. I'm not even a heavy phone user, I don't use social media or have many apps sending me notifications (only Signal) and I feel like my phone can't even last two days of light usage. My iPhone I could use similary and I could usually get 3 days of usage.

Specifically: While iPhones are noticeably more power efficient than Android phones the latter have been sufficient for my usecases especially given that there are typically options with larger batteries.

The downsides of larger batteries are increased weight, volume, and cost. Being more efficient gets the same battery life for less of those three, allowing them to either be reduced or the budget to go to other components.

The Purism CSO has been running a Librem 5 as his primary desktop PC for over a year.

When Plasma was first released I was probably running hardware slower than a Librem 5.

One of the things I want to do on my Librem5 is monitor my servers, so that will involve polling things every few minutes. PowerTop says that I can save power by changing the polling for USB, but that changes Wifi ping times from ~1ms to ~350ms. Eventually I'll probably try experimenting with that to get an option with a 10ms ping time that still saves some power.

It is baffling to me that Apple still hasn't implemented this feature even on their top end "Pro" iPhones... but this feature by itself isn't enough for me to switch to Android full time.

Not months worth of battery saving like an eInk screen maybe but still enough to be insignificant compared to general "screen off" battery consumption.

For the moment the Librem 5 seems to be using apps designed to work on PC desktops but at lower resolutions.

> The RYF has a “secondary processor” exclusion that can be granted on a case by case basis. We will leverage this exclusion to load and train the DDR PHY on the i.MX 8. We will use a secondary processor to keep binary blobs out of u-boot and the kernel.

Our perspective is that this has been made immensely harder by the dozens of 'secure' phone companies providing something far worse than iPhones and Pixels. Most of those aren't presented as being 'open hardware' that's not at all open hardware, but they do share a lot in common. A few of them were so bad that they were actually law enforcement stings from the beginning (https://en.wikipedia.org/wiki/ANOM) or turned into them (https://en.wikipedia.org/wiki/EncroChat) which is amusing.

We'd happily target iPhones but Pixels are the only smartphones providing proper alternate OS support. On Pixels we get to have full verified boot covering all the OS images with downgrade protection, very nicely designed hardware attestation, all the hardware keystore capabilities, the disk encryption key derivation throttling feature with insider attack resistance, A/B updates for the firmware/OS, etc. This is the stuff we expect other vendors to provide us with in order to support their phones. Vendors using a flagship Qualcomm SoC, including the alternate OS verified boot / attestation support and doing a good job with security work themselves come close, but we need more than that. There's a fair bit missing without vendors caring enough to go above and beyond with security and alternate OS support.

When we look at the Librem 5, what we see is rolling back privacy/security massively on the hardware, firmware and software levels with no possible way we could support it. We were actually in contact with them at one point but it became clear that they had zero interest in anything more than an empty partnership where out name would be used for marketing without us being able to use the device at all. That was back before lots of the security improvements in the Android smartphone world where Pixels caught up and even surpassed iPhone security in most areas.

Our expectations used to be far lower, and they get higher as devices like iPhones/Pixels get better. For example, Weaver showed up with the Pixel 2, as did the related secure element insider attack resistance where the main owner user has to authenticate for firmware updates to be applied. Before then, the SEP was way ahead in this area, and continued to be ahead in a lot of others until the Titan M. Weaver provides always enabled aggressive throttling for disk encryption key derivation attempts, with full support for the separate per-user encryption keys. It quickly throttles up to 1 day per attempt after 140 failed attempts. We take it for granted that we get all this stuff including the hardware keystore with physical confirmation support, etc. We expect these features from other vendors. It's not a lot to ask that they implement a feature like Weaver which was introduced with the Pixel 2 via open source AOSP code for the OS integration and open source secure element applets for an off-the-shelf Java smartcard secure element. That got obsoleted by the Titan M, which is actually well on the way to becoming open source with OpenTitan and the Pixel 6 shipping a RISC-V implementation of it based on that. It would be great for that to be fully open sourced and usable by other vendors. That's the kind of thing which is exciting in this area. Pixel 6 using Trusty OS for TrustZone isn't that compelling since TrustZone is terrible and largely obsolete beyond being the component responsible for talking to the actual secure element via authenticated encryption. They're also on the path to getting rid of it now that there's proper virtualization support, which is where the nonsense like Widevine is moving to be properly unprivileged/isolated unlike the mess of TrustZone.

It would be great if other vendors actually cared more and tried to actually compete with iPhone and Pixel security. We found a vendor that's willing to put in the effort, and we're optimistic about that. We're more than happy to work with others. The main thing we need to do is point them at what they need to implemented and the existing open source AOSP code for the OS side of it, etc. We have a lot of experience with this including reporting a bunch of firmware vulnerabilities / bugs in Pixels. There are so many things which could be done better but it starts with any other vendor catching up to where iPhones / Pixels where years ago.

You're saying that as if the firewall handled the communication with the modem. There's a Linux driver behind the firewall to do the actual communication, except that's probably not a USB driver.

The attack area is probably comparable, except a misconfigured USB driver doesn't automatically give full memory access, while a misconfigured IOMMU (the firewall) does.

There isn't anything remotely misleading about the correction that kernel drivers are not in the same situation as most firmware in that they're entirely open source. Every ARM SoC is a proprietary SoC with proprietary CPU cores, memory controller, GPU, image processor, radios and all the other massive complexity included in them. There's a ton of firmware involved in that proprietary hardware. It exists whether or not it's updated. Not updating the radio, GPU, etc. firmware doesn't mean it doesn't exist. Whether or not it's open or closed source has little relevance to the need for it to be hardened, properly isolated and updated. Open or closed source does not directly provide any privacy or security properties. Other components outside the SoC are almost always closed source too. We don't think the fact that the Pixel 6 has a TEE OS based on the open source Trusty TEE or a RISC-V secure element based on OpenTitan makes it inherently more secure than Qualcomm's offerings. Pixel 6 having a Samsung cellular radio and Broadcom Wi-Fi/Bluetooth radio as 2 separate chips from the SoC instead of being part of the SoC has not made those more isolated or more open, and if anything we consider moving away from the Qualcomm radios to be a security regression with how well Qualcomm hardens them, although the overall improvements make up for it. If we made a device, it would have a Qualcomm Snapdragon 8 Gen 1, and we'd be quite happy with having the most secure radios, hardware memory tagging support and Qualcomm's great security work elsewhere. Would it be nice if we could have more control and insight? Sure. If a theoretical currently non-existent open source RISC-V smartphone SoC existed and it had comparable privacy/security (which would be very difficult), we'd be very interested. It would take massive work beyond simply having a viable SoC in the performance class with the necessary functionality to make anything remotely competitive on a security level.

GrapheneOS is a non-profit open source project and doesn't produce or sell hardware products. We don't do consulting, exclusive deals with companies or anything like that. There is no intention to ever sell any products, but rather we research hardware, report issues upstream and choose the best available hardware as the officially supported targets. We also intend to work with hardware partners on equal footing with nothing exclusive to help them produce better devices, which will benefit them and will benefit us through having more and better devices to support. We could potentially get more reliable donation revenue through devices with GrapheneOS installed, but that's in no way any kind of requirement for us to work with vendors. We have little to no interest in ever selling devices ourselves. We're fine with the fact that over a dozen companies sell devices with GrapheneOS installed mostly without giving anything back to us. Only a couple give us any form of donations/support. This is not a business.

The claim that there are closed source kernel drivers is untrue for most mainstream Android devices and is a misunderstanding of why devices stick to a specific LTS kernel branch. Those branches receive 6 years of support now because of the model that's used for mainstream embedded / mobile devices. They don't want to port their drivers to newer versions and spend a year getting it working robustly again. It would be entirely possible to do it and it's possible it will happen for Tensor but it can be an overall bad thing for security instead of a good one due to all the added attack surface. A great example is how a bunch of recent vulnerabilities such as 'Dirty Pipe' only impacted the Pixel 6 due to it using the 5.10 LTS branch which was the newest at the time. A Qualcomm device wouldn't have been impacted due to having an older kernel branch.

Not a major handset maker and not Android, but https://www.crowdsupply.com/sutajio-kosagi/precursor. Also, not state-of-the-art, too.

As stated earlier, there are no closed source kernel drivers for AOSP, GrapheneOS or even most mainstream Android devices running the stock OS.

The reason for using an LTS kernel branch with 6 years of support from kernel.org is stability. Porting forward the drivers to each new kernel release is entirely possible and isn't a lot of work when it's done incrementally. Not that many changes are even required. The issue is that there are substantial regressions in each Linux kernel release and it takes at least 4 months or more to get a production quality kernel for a specific hardware target with nothing break, the massive CTS/ITS/VTS passing, etc. Pixel 6 uses the Linux 5.10 kernel branch which was the latest at the time, and those LTS branches have 6 years of support from kernel.org and expanded support with more bug fixes / security enhancements / other improvements via the AOSP common/generic kernel. It's entirely possible to move to a newer LTS branch. There are no closed source kernel drivers. Is it worth the time, when newer LTS branches have substantially more attack surface and tons of regressions that are going to need to be detected/fixed? Bear in mind it would not expand the lifetime of devices at the moment, time several hardware components won't receive more than 6 years of firmware support.

There are already people who have gotten the mainline 5.15 kernel working with the Pixel 6, but from 5.10 to 5.15 there are a lot of regressions, and there's a lot of new attack surface. There's a reason that ONLY the Pixel 6 among the Pixel family has been vulnerable to several serious core Linux kernel vulnerabilities disclosed in the past few months including the branded dirty pipe vulnerability. There are both advantages and disadvantages to using a newer LTS branch. Unfortunately, one of the disadvantages is that there are more bugs overall, including more vulnerabilities overall. Many software projects mature over time and the rate of finding vulnerabilities goes down. That's not the case for the Linux kernel. It's having vulnerabilities introduced at a faster pace than they're fixed. It isn't better from a security perspective to use the 5.15 LTS rather than the 5.10 LTS, especially with the additional changes backported by AOSP including security enhancements like mitigations, not just bug fixes. It may be a good idea to move to the new LTS branch once it has matured for 1-2 years, but definitely not months after release.

Since you're doing that yourself, I don't think engaging with you on the topic is productive. I responded here due to the inaccurate attacks on GrapheneOS from people promoting Purism products. Doubling down on spreading their inaccurate marketing / talking points isn't going to deter us from responding and we're more than happy to post a more detailed response on our site and across platforms. I already gave detailed responses and don't intend to repeat much of what I've already said.

> The problem is that the Android app ecosystem has a very large number of apps which are based on collecting users' personal information and violating people's privacy, and it is hard for a normal user to avoid all the spyware and malware in Android. In my experience using CyanogenMod/LineageOS and the F-Droid repo since 2015, I inevitably fall back to installing some proprietary apps when using AOSP-derivatives, whereas my PinePhone and Librem 5 USA only have FOSS apps and drivers installed on them. If the goal is to use FOSS as much as possible, you are better off buying a Linux phone in my opinion.

There's a far larger and better ecosystem of open source apps for Android than there is for the products that you're marketing, and they can be used on secure devices rather than blatantly insecure ones not even meeting basic standards as I've already detailed in my responses.

> The problem is that Google only sells Pixels in a very limited number of countries. Whereas Purism offers free worldwide shipping for the Librem 5, the Pixel 6 is only being sold in 8 countries (Australia, Canada, France, Germany, Japan, Taiwan, UK, USA), so your security requirements exclude over 90% of the world's population from being able to use GrapheneOS. Plus, many people don't want to financially support a company like Google which is based on Surveillance Capitalism.

Pixels can be purchased internationally. They don't need to be bought from Google. Purism is a company based around spreading misinformation and marketing their products dishonestly which I know people in our community don't want to support. We're not going to support thoroughly insecure devices from a company which is unwilling to even admit to the limitations/weaknesses let alone fixing them and producing something we could ever consider supporting. The experience we had with them is that they only want to use the name of projects like ours to promote themselves as partners without doing anything on their part. They engaged in libel/harassment/bullying targeting our developers in response to us not supporting their phone as a target and explaining why within our community. I see what you're doing here as an extension of their dishonest marketing and inaccurate attacks on other platforms/projects/products. If this is going to be something that's happening regularly, we'll add detailed documentation / articles to our site about the topic to reference so we don't need to keep writing up the same things.

> Not true. Purism has promised to provide updates to the proprietary firmware on the Librem 5, and already provides instructions for how to update the firmware on the WiFi/BT and USB controller.

There aren't full firmware security updates for the Librem 5 and what I said is completely accurate. What's even worse is that they do not ship the incomplete updates that could be available and they did things in a way that makes it impossible to even ship all of those as part of an OS. Please don't claim that my completely accurate description of the situation is not true based on something that's not in any way debunking what I said.

> What are you talking about? Purism purposely designed the Librem 5 to avoid planned obsolescence, so it looked for component suppliers who support their hardware for a long time. For example, NXP guarantees that that it will provide updates for the i.MX 8M Quad for 15 years (Jan. 2018 - Jan. 2033).

They're unable to provide full security updates from day one and the device is already end-of-life in terms of what that means for GrapheneOS. It would have to be marked as end-of-life from day one if we added support for it. We would be unable to declare any Android security patch level for the device due to it not meeting the basic security requirements and not having full firmware security updates available. What I've said is true, and you're just claiming otherwise based on their deliberately very incomplete and misleading marketing.

> In contrast, Google only promises to provide 3 years of OS updates and security updates for the Pixel 3/4/5, and 3 years of OS updates and 5 years of security updates for the Pixel 6. Qualcomm announced in Dec. 2020 that it will support its Snapdragon processors (which are used in Pixel devices) for 3 years of Android updates and 4 years of security updates.

Those are minimum guarantees of full security updates, not end-of-life dates and the number of days you get those for the Librem 5 is ZERO. The only recommended devices for GrapheneOS are the Pixel 6 and Pixel 6 Pro, which means that there is at least 5 years of full security updates for the devices we support. You can see from our site that we continue providing extended support releases which we mark as insecure past the end-of-life of devices. A device is end-of-life as soon as any important component no longer provides the proper monthly security updates. How can we support the Librem 5 even aside from all the missing security features which have already been explained elsewhere, when we would be unable to provide anything close to the March 2022 security update, and would be unable to ship all the updates that are available through the OS?

> Linux phones like the Librem 5 and PinePhone use separate components which are supported for many years by the manufacturers, whereas most Android phones (like the Pixels) use integrated mobile system-on-chips which are only manufactured for 1-2 years and only supported for 3-4 years by the manufacturer. Because Linux phones use components with long-term support by the component suppliers, the Librem 5 is the first phone to be sold with the guarantee of lifetime software updates, and PINE64 promised to manufacture the PinePhone for 5 years, which is longer than any other smartphone ever sold.

This is completely inaccurate. They still use an SoC and the components they've chosen do not provide a longer period of support in the sense that Android expects in order to declare the latest security patch level. Several of their component choices including the radios rule that out, as does the way they are integrated. Your claim of lifetime security updates is completely bogus and demonstrates the extreme lengths Purism goes to in order to mislead people and profit from it. They still need firmware support and all the drivers, etc. still need to be maintained. There's really no point of engaging with people lying through their teeth and pushing all their inaccurate talking points so I'm not going to keep engaging with you much further.

Linux doesn't mean systemd, polkit, glibc, GCC, binutils, GNOME, pulseaudio/pipewire, Wayland/X11, etc. It makes no sense to claim these are Linux phones when the vast majority of smartphones run Linux. It's marketing spin. If you want to call it a GNU/Linux phone, go ahead, but what you're doing is a deliberate attempt at misleading people on their part.

> The Librem 5 doesn't need an IOMMU, because it uses separated components, and it uses serial buses (USB 2.0/3.0, SDIO, I2C and I2S) that don't allow direct memory access, so there is absolute no chance of the WiFi/BT, cellular modem, GNSS and USB controller being able to access the RAM or the SoC's cache. Unlike the Snapdragon processors in Pixels whose hardware is essentially a black box, we can independently verify by looking at the open source schematics that direct memory access is not possible in the Librem 5.

This is not accurate. It still has an SoC with a ton of components aside from the SoC despite your inaccurate claim that it doesn't, and those components still need to be isolated with an IOMMU. The other components which you're talking about using USB are dramatically less isolated than the Qualcomm or Samsung baseband on mainstream devices. You're trying to present something dramatically worse as being better in this regard. Are you trying to claim that the Librem 5 doesn't have components like a GPU and other SoC components? The Librem 5 hardware is also just as much of a black box. It's 100% as proprietary. It does not have firmware or hardware that's any more open and this is a blatant lie. Them marketing the hardware as being more open is thoroughly unethically and dishonest. They've done the same with their laptops and other products, which has done immense harm to projects like Talos actually trying to produce open hardware in any actual sense of the word.

Many other people promoting these platforms often talk about the (limited) support for running Android apps. You can choose to use those apps there, just as you can on GrapheneOS, and it does not make sense to claim that availability of apps is a bad thing. GrapheneOS supports nearly every Android app which would run on the stock OS thanks to the sandboxed Google Play compatibility layer feature. Many of our users including make no use of that feature, and it didn't exist before 2021 so our entire userbase before then was happily using it without Play services. There are more users now, but there were still many before, and many are happy using only the open source Android app ecosystem which goes far beyond F-Droid which doesn't even have apps like Signal, Chromium, Brave, Bromite, etc.

You're trying to make this about AOSP vs. a completely insecure software stack but the post is about a phone which is capable of running AOSP and other phones are more than capable of running the desktop Linux stack. It's a red herring, and you're being thoroughly dishonest and manipulative in how you're presenting the app ecosystem considering that there is a far larger open source app ecosystem for AOSP than there is for desktop Linux on mobile... F-Droid has very incomplete coverage of the overall open source app ecosystem and they don't always do a particularly good job maintaining it. F-Droid itself still targets Android 7.1 (API 25).

I'm talking about the fact that this hardware, firmware and software is a decade behind on security and has almost zero systemic privacy/security work across it. You have the privacy and security situation completely backwards. The fact that Purism blatantly lies about many aspects of their hardware, firmware and software also demonstrates that they're a highly untrustworthy vendor. I would trust a company like HTC far more because at least they aren't blatantly lying about the security patch level, openness of the hardware and they aren't covering up security vulnerabilities, weaknesses and the fact that the firmware/hardware is proprietary.

> Yes, Android/AOSP does have a lot more security built into its design than Debian->PureOS, but it is based on a model of letting all sorts of unverifiable and dangerous code run inside it.

I'm not sure why someone would want to place complete trust in thousands of different fragmented projects which have no real isolation and no systemic privacy/security work on the overall OS or across those projects. Many of those projects are unmaintained, and some of them have even shipping malicious changes either unintentionally or intentionally before. You're also trusting the huge amount of packagers for the OS who set up the builds and patches for these projects. There are still closed source apps availability, but open source is not the magical panacea that you present it as and does not infer any inherent privacy/security properties on the software. It doesn't make the developers inherently more trustworthy or ethical either. It's a development approach, which we can both agree is a great approach and enables people to make changes to the software, fork it or attempt to contribute to it upstream.

> Care to provide any evidence to prove that Purism or its employees are "spreading tons of misinformation and outright lies about the mainstream options"?

You've done a great job of doing that yourself by cycling through a whole bunch of inaccurate talking points promoting their products and attacking other projects like GrapheneOS. Most of it comes directly from Purism, and your comment is an extension of their highly underhanded, dishonest and malicious attacks on projects like GrapheneOS to make sure they keep getting their substantial salaries and profit.

At the moment, I'm not currently interested in investing the necessary time into writing such as an article. If you're going to post another lengthy problematic reply, that's the medium I'm going to use for my response rather than writing another comment on this platform which few people are going to see, which is not a good use of my time.

But basically, it is not secure neither on a hardware front, neither on software. The latter runs everything as the same user so a rouge bash script can encrypt your whole photo gallery, exfiltrate any data, etc.

[0]: https://gitlab.gnome.org/GNOME/libshumate

Being able to receive and place calls is critical for a phone, but I think for many people it's far from the most common activity performed on one. I don't upgrade my phones to make better phone calls, I upgrade them to get better cameras.

But I don't know if a full emulator is a good idea on a smartphone.