Thanks to volunteer effort and NLnet funding we have had developers working hard the past couple weeks on making graphics work on the Samsung Galaxy S3 for the upcoming Replicant 9 release. Three different software renderers have been successfully used with the current Replicant 9 development version: softpipe (Mesa), llvmpipe (Mesa), and SwiftShader.

The Linux kernel and Mesa required some small tweaks to allow using the Exynos DRM driver for software rendering. For the Linux kernel we had to allow in the DRM subsystem the creation of dumb buffers for render nodes as discussed previously on LKML. The devfreq module was also disabled as a workaround for graphics corruption issue that happened due to too low clock speeds. Mesa on the other hand required us to whitelist Exynos DRM driver to be used with a software rendering driver called kms_swrast. After these tiny changes we were already able to use the Mesa’s software rendering backend called softpipe! Unfortunately, softpipe turned out to be unusably slow even after using HW overlay planes to offload some of the buffer compositions to the Exynos 4412 display controller found on the Samsung Galaxy S3.

Next the community members Putti, dllud and GrimKriegor worked on a new revision for a patch originally used in the Android-x86 project for enabling llvmpipe support in Mesa. When the compilation fixes and slight adjustments for the rebased original patch were done we had yet another software renderer to use. The new revision of the patch enabling llvmpipe is now submitted again to Mesa and hopefully this time around it will go through.

The rendering speed with llvmpipe turned out to be only slightly faster than softpipe and still unusable for everyday usage. There are plans optimize llvmpipe for the ARM Cortex-A9 processor in Exynos 4412 so hopefully soon it will be usable for everyday usage.

Before diving into the world of NEON optimization for llvmpipe and so forth, we decided to give SwiftShader software renderer a go together with hwcomposer.ranchu and gralloc.default. Other than having to add support for UDIV and SDIV instruction emulation in the Linux kernel, we got it running by simply setting some system properties and adding the SwiftShader and hwcomposer.ranchu modules to the system image. The speed of SwiftShader felt like it was at least twice as fast compared to llvmpipe and softpipe but still just a tiny bit too slow to be enjoyable for everyday usage. We are now hoping to find people that could work with us to replace hwcomposer.ranchu with drm_hwcomposer. Using drm_hwcomposer with SwiftShader would allow us to take advantage of HW overlay planes to speed up the compositions.

If you are interested in helping or want to learn more you can get in touch with us on IRC on the #replicant channel on Freenode, or via our mailing list.

A Replicant contributors meeting will take place in or near Paris in France the 27 and 28 July 2019.

As we are still looking for a place to host the meeting, we don’t have a definitive address yet.

It will be open to anyone who contributes or wants to contribute to the Replicant project.

The meeting will most likely be in English as not all the Replicant contributors who plan to attend speak French.

More details will be posted on the wiki page dedicated to this event over time.

Last minute information, if any, will also be posted on that wiki page.

The Replicant project has been looking forward to support devices with free software bootloaders. While Replicant is a fully free software Android distribution, many freedom, privacy and security issues are orthogonal to the operating system. The hardware design of each computer (smartphone, tablet, laptop, etc.) people use, and the architecture of the cellular network also have their set of issues. For more information on these issues, the Replicant project has some documentation on the topic.

So far all the devices that are (or have been) supported by Replicant use a nonfree boot software (the bootloader). These devices also use hardware restrictions to deny users the freedom to replace them completely with free software, effectively forcing them to run nonfree software. This is a very serious freedom issue that prevents users from being in control of their devices.

There were several attempts to add support for devices with free software bootloaders in Replicant:

• The LG Optimus black: this smartphone doesn’t prevent users from replacing the bootloader. Paul Kocialkowsky did a lot of work to add support for this device in upstream u-boot (a free software bootloader) and added minimal support for it in the upstream Linux kernel. However support for some of its most important hardware components like the display are still missing in the Linux kernel. This device can probably still be found second hand
• The GTA04 smartphone from Golden Delicious: this smartphone has a free software bootloader which is based on u-boot. The smartphone was designed to run GNU/Linux and has almost complete support in upstream Linux. There were attempts to add support for it in Replicant 6.0, however a lot of time was spent to try to make suspend to RAM work with Android. However older Replicant 4.2 images are available. Several hardware revisions of the GTA04 have been made and shipped to customers and developers over the years. However this has stopped due to manufacturing issues. Another issue is that the revisions before A5 only have 512M of RAM and a high DPI display: This combination makes running Android 9 potentially challenging. Fortunately the A5 revision has 1G of RAM, but not a lot of working units were produced.

There is also some ongoing work to specifically add support for smartphones that are currently supported by Replicant like the Galaxy SIII (i9300), the Galaxy Note 2 (n7100) and their 4G versions (i9305 and n7105). The 4G versions could also be supported by Replicant if the work to support their modem (through QMI-RIL) is resumed.

The Replicant project will receive a mobile device, the NC_1 (formerly called Necuno Mobile) from its manufacturer (Necuno Solutions), which will have a free software bootloader

This device has the size of a smartphone, but doesn’t have a broadband modem: while users will not be able to use a built-in modem for phone calls, SMS or to access the Internet, it is still the best way to be completely sure of avoiding any freedom privacy and security issues related to broadband modems and the cellular network. It will also require less work to add support for this device in Replicant.

Even if it’s possible to disable the modem on some of the mobile devices currently supported by Replicant by not loading the modem’s code, some nonfree software still run on these mobile devices. This includes the bootloader and potentially any other nonfree software that it may load. Because of that we cannot be 100% sure that the modem is completely disabled.

The Necuno Mobile will use an I.MX6 Quad system on a chip (which is a chip that contains the main CPU, the microSD card controller, the GPU, etc.). Its free software support is better than for many other system on a chip: the only functionality of the I.MX6 Quad that requires nonfree software is the video decoding acceleration. The article on single board computers has more details on freedom issues affecting various system on a chip and by extension the single board computers that use such components.

A Replicant developer (Joonas Kylmälä) will receive a Necuno Mobile to work on it.

The Necuno Mobile should have a Linux kernel that is very close to upstream: this is a good opportunity for a new attempt to enable Replicant to use upstream kernels. This has many advantages. One of them is that in the long run, it should decrease the amount of work required to maintain the devices and potentially increase their lifetime.

This should also enable the Replicant project to more easily add support for other devices that can use an upstream kernel, like the GTA04, or devices like the Galaxy SIII (i9300) and the Galaxy Note 2 (n7100) that are starting to have good support in upstream Linux.

It is also very interesting in the long run as we could share some of the work with other smartphones projects like postmarketOS who are also trying to support mobile devices with upstream kernels. It could also enable the Replicant project to more easily support future mobile devices that will have free software bootloaders, as some of them will also use kernels that are meant to run GNU/Linux.

A crowdfunding campaign was launched over a month ago by Bootlin in order to fund the development of an upstream Linux kernel driver for the Allwinner CedarX VPU. The VPU (Video Processing Unit) is in charge of offloading video decoding and encoding to a dedicated hardware block, relieving the main CPU. While Replicant does not support Allwinner devices at this point, the project has acquired a number of Allwinner tablets a few years ago, that helped with the advancements of Allwinner platforms support in upstream projects such as the U-Boot bootloader and the Linux kernel.

Recently, Replicant was a candidate for the Google Summer of Code program and we came up with a list of tasks for the occasion. Although our application was not accepted, we are still interested in completing the tasks that we picked up. We put a deliberate focus on supporting mobile devices in mainline U-Boot and Linux, with a particular emphasis on Allwinner devices, the Optimus Black and the Kindle Fire (first generation). We believe that supporting mobile devices and using standard driver interfaces in the upstream Linux kernel is the only sustainable way for freedom on mobile devices. Instead of writing device-specific code specifically for Android for each of the supported devices, this would allow using generic Hardware Abstraction layers (HALs), reducing the amount of work for hardware support on the Replicant side in the long run. This also allows running other operating systems that integrate the upstream Linux kernel interfaces, such as standard GNU/Linux distributions.

In spite of this, I have been dedicating more and more time to contributing to upstream projects such as coreboot, U-Boot and Linux for supporting devices of various form factors, including mobile devices, laptops and single-board computers. Thus, I became less and less active on the technical side for Replicant, where Wolfgang and others have picked-up the work. There is still a lot of room for contributions and everyone is warmly encouraged to join-in and help with the upstreaming effort for devices, especially regarding the Optimus Black, Kindle Fire (first generation) and Allwinner devices.

As a student approaching graduation, I have joined Bootlin (formerly Free Electrons) in Toulouse, France for an internship focused on supporting the Allwinner VPU in upstream Linux and userspace. It definitely fits perfectly with the logic behind focusing Replicant towards upstream Linux support. In order to accelerate the development of the driver, Bootlin has decided to start a crowdfunding campaign in order to fund Maxime Ripard, who has been working for the company and maintaining Allwinner platforms in the Linux kernel for a while.

As the main goal of the campaign was reached within its first week, Maxime will be able to work with me on the VPU. His in-depth understanding of the sun4i DRM video driver’s innards will also reveal very useful for accelerating the processing of the frames coming from the VPU (without unneeded copies of buffers) and implementing scaling in hardware. In order to support the VPU hardware efficiently, a number of changes have to be introduced to the Linux kernel. It currently lacks an interface to provide coherency between setting specific controls for the media stream and the input/output buffers that these controls are related to and should apply to. This API has been implemented by Alexandre Courbot (who’s working at Google on the Chromium OS project) as the V4L2 request API, that fits the requirements for the Allwinner VPU driver. Other VPU drivers, such as the tegra-vde driver that supports the Tegra 20 video decoder engine, also require this API in order to implement a proper V4L2 mem2mem driver.

The crowdfunding campaign still has 10 days to go and two stretch goals to meet (while the first stretch goal, about supporting newer Allwinner SoCs was already met):

• H265 video decoding support
• H264 encoding support

As I am not directly impacted by the funding received through the crowdfunding campaign, we believe that there is no direct conflict of interest writing this blog post on the Replicant blog.