As the mainline porting workflow becomes more streamlined, we found ourselves able to automate and simplify a lot of things that previously required a lot of manual work. A great example of this is the mdss-panel-driver-generator which is able to convert vendor devicetree panel control sequences into a driver which meets the requirements for upstreaming into mainline.

More recently, tools like the new msm-firmware-loader for postmarketOS, and the potential for scripts which can extract and package firmware simply given a link to an OTA update, we might now have a generation of porters with little/no knowledge on how firmware works in the context of downstream / mainline. Qualcomm platforms differentiate between mdt firmware files where the firmware is split into many files, and mbn files where it is squashed into one larger one.

The OnePlus 5 is a popular high-end phone from 2017, featuring the Qualcomm Snapdragon 835 SoC and Adreno 540. It has a 1080p display and up to 8GB of RAM.

Mainline support for the device has been around for a while, since early 2020 in fact. Unfortunately, the Snapdragon 835 SoC it is based on lacks the same interest upstream that has benefited SDM845 devices so much, requiring a lot of work to reach a usable level of functionality.

Despite that, after a lot of time and effort from Jami Kettunen and several other Snapdragon 835 developers, we can finally welcome a postmarketOS port based on a close-to-mainline kernel! A surprising amount of the hardware already works, although there are a few known stability issues, such as the requirement to run diag-router, a tool meant for debugging the modem to prevent WiFi from crashing.

The RISC-V ISA is commonly divided into two architectures, namely RV32 and RV64, covering 32-bit and 64-bit register sizes respectively. And although the RV64 standard is not yet fully RV32-compatible, the two are, except from some minor divergences, closely associated products of the same family.

In fact, it took some years to see the Linux kernel running on RV64 (and partially RV32). The first obstacle was predictably the lack of suitable and powerful enough silicon for the new software, and the next - which was thankfully solved quite fast - was porting the entire kernel to a new architecture.

But now, with Linux support approaching full stability on a variety of RISC-V chipsets, distributions like Alpine starting to add support for RV64 packages, and an increasing quantity of chips powering Linux single board computers which retail for prices as low as $12, another challenge for most consumer producers occurs: how to turn this into a competitive ecosystem for Android smartphones?

As most readers will know, the PinePhone is essentially considered to be the reference mobile Linux hacker's playground. And while this achievement may be for the most part thanks to its flexible software, its hardware structure, and the Quectel EG25-G baseband in particular, are no exception. The modem firmware can be flashed through fastboot over USB like a regular Android phone, and in recent months, even an open-source firmware for it was developed, for what could be the first time ever on this kind of device, including audio, GPS and proper power management. In fact, what stopped this impressive operating system from spreading further, or being shipped by default on new Pines, is not really its lack of maturity, but rather the set of legal problems that are associated with re-flashing of mobile broadband modems. (There are still instructions on how to flash it, of course)

It was not three years ago that the PinePhone was announced. With entry-level, but relatively modern specifications, video output through Type-C, and full openness to custom software and mainline Linux, the device grew to be possibly even more popular than expected. Name it, and most Linux enthusiast will either know it, or own one themselves.

While its humble specifications did not put it directly on the same league as Purism's Librem 5, the community grew large enough to test most pieces of modern open-source software to its form factor.

As many will know, the main difference between commercial phones and Linux-native alternatives is in how the latter is more of a "pocket computer" that can be used without firmware limitations. Coupled with often very open, if often modular, hardware, this category of devices kind of takes on the PDAs which we missed since the smartphone era. Having proper desktop-ready apps and advanced hardware in your pocket brings not only somewhat better freedom but also improved possibilities. This, naturally, if you can accept the still limited usability of most mobile Linux software, which is far from the ease of use and stability of commercial models.

If choosing a graphical environment for traditional Linux computers is already confusing for some users, deciding one for Linux phones is twice as hard. Most mobile interfaces still lack the maturity and usability of their traditional counterparts, also because integrating all kinds of desktop applications without keyboard can be tricky, given the graphical toolkit fragmentation and the challenges of functionally scaling most windows to very small sizes.

Here we will briefly showcase the main mobile Linux projects, most of which we mentioned in the past, to give an overview of the reasonable choices for most users.

UNISOC (and formerly Spreadtrum) chips have always been a very popular choice for Android device manufacturers. However, they were not commonly known for having wide mainline compatibility. Similarly to Rockchip and AllWinner, these chips became popular for their fast performance and relatively low prices, to the point of being used on devices ranges from the (just released) JingPad A1 to mainstream Android Samsung and Teclast tablets.

In the last years, the number of Linux distributions aimed at smartphones and tablets has grown considerably. What was initially a land of projects like UBPorts, Armbian and postmarketOS has grown to interest projects like Manjaro, and several Debian and Arch spins.

However, the newborn Kupfer does not look like yet another fork for Linux phones. While at an early stage, the ambitious project's infrastructure mimics that of postmarketOS, and even shares some early testers and users. The aim is not only to build a ready-to-flash Arch with some added mobile packages, but rather to have a complete pmbootstrap-like suite of tools (the Docker-based kupferbootstrap) to easily port new devices, and maintain existing ones. The first devices to be supported are Snapdragon 845 phones (OnePlus 6T, Poco F1 among others) and the BQ Aquaris X5 – and, interestingly, no PinePhone yet.

You have most likely heard of Fairphone, who iFixit claim to be "the world's most repairable phone", and certainly lives up to it. But now, there's a new kid on the block, and they have a different approach.

SHIFT are a very small phone manufacturer based in Germany, they produce quite a few models and even have a fully convergent device planned, sort of like the Motorola Lapdock of yonder years. In this article, I will discuss my experience with the SHIFT6mq, and why I think you will hear a lot more about SHIFT in the future.

The Pocket P.C., short for Pocket Popcorn Computer, is a handheld created by the independent hardware manufacturer Popcorn Computer, and originally announced in the fall of 2019. This PDA comes with an integrated keyboard, and runs a pure Linux OS based on the Allwinner A64 platform.

By now, most readers will be aware of the quick progress of Qualcomm phones in the mobile Linux ecosystem. This became especially true with the SDM845 platform, powering the near-stable Linux ports of devices like the Oneplus 6/6T and Poco F1 which we often mentioned in the past.

However, after some work, a new family of popular Android devices is entering the kernel: the Xiaomi Redmi Note 7 from 2019, was finally shown to boot the 5.15 kernel, and should be soon entering the mainline Linux kernel tree in a set of new patches. The Note 7 (not that one) is based on the Snapdragon 660 chipset, a 14nm SoC which is focused on high performance and power efficiency. Powered by eight Kryo cores (custom ARM64), it can be clocked as high as 2.2GHz, supports LPDDR4 memories, top-speed LTE modems and high-speed I/Os while keeping a close eye to the power consumption.

There have been some recent waves in the Pine64 community, as user Federico Izzo (@_symmetrist_) found a PinePhone mod which can roughly double the eMMC speed simply by moving a single resistor. But, anyone can write a flashy headline, how the heck does this actually work?

Federico did a write-up on their blog which explains where they got the idea from and how to perform the mod for yourself, as well as how distros can support the mod. Thanks to this it is now already supported in postmarketOS!

To summarise, the PinePhone eMMC ships in the "DDR52" timing mode by default, which modes are allowed depends on the eMMC controller voltage (this voltage is supplied by the VCC-PC power rail and provides power to the eMMC controller). DDR52 is the highest timing mode supported at 3.3v, which is what the PinePhone uses. To run this any higher, the supply must be at 1.8v or 1.2v.

The cold months are here, people are looking forward to the holidays, and of course there is a new FSF Ethical Tech Giving Guide. This is the yearly Free Software Foundation article giving advice what hardware to avoid for reasons of freedom, and which ones are instead recommended.

This year, the guide starts at the top with advice for smartphones, which is great since these account for a large chunk of computing use these days. Let's take a closer look at the advice in this screenshot below.