Zephyr RTOS 2.x on Fedora: Configure the build environment

It’s been a while since I’ve had time to play with the Zephyr RTOS project and the project has evolved greatly since so I thought I’d document the process I went through while playing with Zephyr 2.1 on Fedora using the Fedora native cross toolchains rather than the various ones suggested by the Zephyr Project docs.

I’m going to do a couple of posts in this series to break it up a little. This first one will be getting a generic build environment setup. I’ll go into more detail on the specific devices I’m playing with but the ones I have handy are ARM Cortex-M based so that’s what I’ll be focusing on even though Zephyr RTOS supports numerous architectures.

As before it’s worth reading the latest Zephyr Getting Started guide. This time around I’m using a AWS aarch64 a1.medium instance running a Fedora 31 cloud instance but I’ve also tested that a DigitalOcean Droplet with 2Gb RAM works with the later ZephyrRTOS releases too.

Once you have a Fedora instance up and running install the required dependencies:

# install core utils, git and cross compilers
$ sudo dnf install git-core cmake ninja-build gperf dfu-util dtc \
xz file python3-pyelftools arm-none-eabi-gcc-cs python3-pip

# requirements for the west (Zephyr meta build tool)
$ sudo dnf install python3-colorama python3-configobj \
python3-configobj python3-docopt python3-pykwalify \
python3-dateutil python3-colorama python3-docopt \
python3-pykwalify python3-packaging

We now install west:

$ pip3 install --user --upgrade west

We now have key build dependencies installed so we can initialise west and clone key repos using west (this will take a little while to do an initial clone run):

$ west init zephyr
$ cd zephyr
$ west update

Setup the Zephyr environment for cross compiling with the distribution tools:

$ export ZEPHYR_TOOLCHAIN_VARIANT=cross-compile
$ export CROSS_COMPILE=/usr/bin/arm-none-eabi-
$ source zephyr-env.sh
$ mkdir builds

By default the above uses the git master branch of the Zephyr git repo. If you wish to use a stable branch you can just check it out. The latest stable release is, currently, 2.1 so to use this you can check out the stable branch:

$ git checkout v2.1-branch

With this we now have a Zephyr RTOS development environment setup for building for Arm Cortex-M based devices on Fedora using the distribution’s cross toolchains.

The state of open source GPU drivers on Arm in 2019

I first blogged about the state of open source drivers for Arm GPUs 7 years ago, in January 2012, and then again in September 2017. I’ve had a few requests since then to provide an update but I’ve not bothered because there’s really been no real change in the last few years, that is until now!

The good news

So the big positive change is that there’s two new open drivers om the scene with the panfrost and lima drivers. Panfrost is a reverse engineered driver for the newer Midguard and Bitfrost series of Mali GPUs designed/licensed by Arm, whereas Lima is aimed at the older Utguard series Mali 4xx series of devices. Panfrost, started by Alyssa Rosenzweig, and now has quite a large contributor base, has over the last few months has been coming along leaps and bounds and by the time Mesa 19.2 is out I suspect it should be able to run gnome-shell on an initial set of devices. I’m less certain the state of Lima. The drivers landed in the kernel in the 5.2 development cycle, which Linus just released. On the userspace side they landed in the mesa 19.1 development cycle, but they’ve greatly improving in mesa 19.2 cycle. Of course they’re all enabled in Fedora rawhide, although I don’t expect them to be really testable until later in the 19.2 cycle, but it makes it easy for early adopters who know they’re doing to be able to start to play.

A decent open source driver for the MALI GPUs from Arm had been the last biggest hold out from the Arm ecosystem we’ve been waiting for and it covers a lot of the cheaper end of the SBC market with a lot of AllWinner and some Rockchip SoCs having the MALI 4xx series of hardware, which will use the Lima driver and other lower to midrange hardware shipping with the newer Mali midguard GPUs like in the Rockchip 3399 SoC.

Other general updates

Since I last wrote the freedreno (QCom Ardreno) and etnaviv (Vivante GCxxx series) have continued to improve and add support for newer hardware. The vc4 open drivers for the Raspberry Pi 0-3 generations have seen gradual improvement over time, and there’s a new open v3d driver for the Raspberry Pi 4 which they use from the outset.

The last driver is one that seems to have transitioned to be in limbo is the driver for the Nvidia Tegra Arm platform. While it has an open driver for the display controller, and the GPU mostly works with the nouveau driver, at least on the 32 bit TegraK1 (the upstream state of the Tegra X-series is definitely for another post) they appear to have yet another driver, not their closer x86 driver, but another one (not the latest rev, which is 4.9 based, but the only linkable version I could find) which is needed to do anything fun from an CUDA/AI/ML point of view, I wonder how it will fit with their commitment to support Arm64 for their HPC stack or will that only be interesting to them for PCIe/fabric attached discrete cards for HPC super computer deals?

That brings me to OpenCL and Vulkan for all the drivers above, for the vast majority of the open drivers support for either is basically non existent or in the very early stages of development so for the time being I’m going to leave that for another follow up in this long winded series, probably when there’s something of note to report. The other thing that is looking quite good, but one for another post, is video acceleration offload, there’s been quite a bit of recent movement there too.

Raspberry Pi improvements in Fedora 29

So Fedora 29 is probably going to account for the largest single improvement to support on the Raspberry Pi support in Fedora since we added initial support in Fedora 25. It certainly wasn’t without issue, but after quite a bit of debug we’ve got the post release issues with the WiFi back to being stable!

WiFi improvements
The support for upstream NVRAM files and the ability to add those files to linux-firmware means we get WiFi support for the Raspberry Pi 3 Series of devices out of the box! No need to grab anything, it just works! Well mostly, we had some issues with WiFi being very intermittent, as well as a missed bug around aarch64 but now with the 4.19.10 kernel everything appears to be working and stable. This makes me very happy and it took longer than I had hoped but we’re there. This device specific NVRAM driver support will also help another bunch of cheap Arm and x86 based devices that ship with Broadcom/Cyprus based WiFi support moving forward.

ZRAM enabled by default
By supporting and enabling ZRAM swap by default we get a more responsive device and less wear on the MicroSD storage. Over all we’ve seen reasonable performance improvements and to no date no major issues.

GNOME performance improvements
In May 2018 the Raspberry Pi Foundation kindly hosted a GNOME Performance Hackfest in the lovely Cambridge. Over a couple of days we managed to fix a number of issues seen, review and document a number of issues and work on a number of ways of reducing the memory usage of GNOME. Of course this improvement is primarily seen constrained devices like the Raspberry Pi but ultimately less memory utilisation by GNOME even helps devices with decent amounts of RAM and CPUs too. The fixes didn’t arrive in time for Fedora 28, but a bunch have landed in Fedora 29 providing noticeable improvements, and the GNOME team is by no means done and there will be more coming in Fedora 30 and beyond! It was an excellent start and I expect there will be ongoing enhancements here into the future especially with devices like the Purism Phone which will have similar constraints.

Initial CPU frequency support
Another of the largest issues around the Raspberry Pi is complaints it was slow, part of the issue here is that there’s no upstream CPU Frequency driver which means all models of the Raspberry Pi run at a glacial, but safe, 600Mhz out of the box compared to the highest speed, which on the 3B+ is 1400Mhz. With Fedora 29 we’ve landed an experimental cpufreq driver which allows us to run the Raspberry Pi 3-Series at much closer to optimal speeds. While this is experimental it might not stay around if we find out it causes issues or ends up being a maintenance burden but to date it hasn’t yet appeared to have caused any issues.

HWmon Voltage Sensor
There’s a new driver that reports when the voltage supplied by the PSU drops below the required voltage. It can be a bit noisy in dmesg but one of the biggest support problems we have with the Raspberry Pis is people using a power supply that’s not powerful enough, this issue is more of a problem with Fedora 29 because with the support for running at faster frequencies due to the cpufreq driver it means we also draw more power and some PSUs that were previously fine now cause issues because they can’t supply enough current.

Enhanced support for config.txt
A lot of the hardware addons are supported in Raspbian are done by enabling things in the config.txt file, this in turn does things like loading DT overlays and merging them with the base DT to enable extra hardware like HAT support. We have enhanced the way Fedora works with this which enables us to be much closer to the way Raspbian handles these things. The advantage this has is that the documentation that’s written for Raspbian is then usable by Fedora in the wider Raspberry Pi ecosystem which in turn makes it easier for end users to get HW up and running due to less differences in process. There’s further enhancements to make here but every step closer is easier for everyone to enable and use their favourite HATs.

Improved bcm283x firmware support
In preparation for grub2 support we enhanced how we deal with the firmware that the Raspberry Pi uses for booting. This deals with the early startup. We never use to upgrade it by default to ensure things didn’t break, but it also meant most users also didn’t by default get the fixes and enhancements. Now we do. The config.txt is also handled directly which means if you never edit the file you now automatically get any changes we make, because rpm handles the file as a config file, if we change it you get a .rpmnew file so you won’t lose your changes.

Camera support
This wasn’t available in the Fedora 29 4.18 kernels, but with the rebase to the 4.19 kernel the support for the camera on the Raspberry Pi CSI Camera interface improved enough we could enable this in Fedora. The early 4.19 kernels don’t automatically detect and load support if the camera module is attached. There’s some patches in 4.20 in rawhide for this, and we’ll bring some of this to 4.19 soon, and we’re working with upstream to further improve the camera support. You’ll also want the latest bcm283x firmware which tweaks some of the config.txt and updates to a firmware with ISP fixes.

Another improvements
There was also a number of general Arm improvements which sped up crypto on the Raspberry Pi, improved the USB, fixed up some issues with the wired ethernet on the 3B+, power and a number of other fixes. As always there’s more coming. The 4.20 kernel rebase should also bring with it analog sound support early in the new year.

Conclusion
Overall I was pleased with the work that landed in Fedora over 2018 for the Raspberry Pi. The WiFi regression was disappointing, but now with that fixed in 4.19.10 we have WiFi support out of the box without users needing to download anything which moving forward will make things a lot more straight forward. The initial support for the camera makes it much more useful in numerous use cases and we’ll really polish up the HAT support in Fedora 30 which for me is the last remaining big ticket item for Raspberry Pi support. There’s still some annoying bits around the EDID detection in the display, but there’s work to improve that upstream, and also there’s work to land the media decode offloading upstream too which will also one of the few remaining bits.

Using ZRAM as swap on Fedora

One of the changes I did for Fedora 29 adding using ZRAM as swap on ARM. The use of compressed RAM for swap on constrained single board computer devices has performance advantages because the RAM is an order of faster than most of the attached storage and in the case of SD/emmc and related flash storage it also saves on the wear and tear of the flash there extending the life of the storage device.

The use of ZRAM as swap isn’t limited to constrained SBCs though, I also use it on my x86 laptop to great effect. It’s also very simple to setup.

# dnf install zram
# systemctl enable zram-swap.service
# reboot

And that’s it! Simple right? To see how it’s being used there are three commands that are useful:

# systemctl status zram-swap.service
● zram-swap.service - Enable compressed swap in memory using zram
   Loaded: loaded (/usr/lib/systemd/system/zram-swap.service; enabled; vendor preset: disabled)
   Active: active (exited) since Tue 2018-10-09 22:13:24 BST; 3 days ago
 Main PID: 1177 (code=exited, status=0/SUCCESS)
    Tasks: 0 (limit: 4915)
   Memory: 0B
   CGroup: /system.slice/zram-swap.service

Oct 09 22:13:24 localhost zramstart[1177]: Setting up swapspace version 1, size = 7.4 GiB (7960997888 bytes)
Oct 09 22:13:24 localhost zramstart[1177]: no label, UUID=d79b7cf6-41e7-4065-90a9-000811c654b4
Oct 09 22:13:24 localhost zramstart[1177]: Activated ZRAM swap device of 7961 MB
Oct 09 22:13:24 localhost systemd[1]: Started Enable compressed swap in memory using zram.
# swapon
NAME       TYPE      SIZE   USED PRIO
/dev/zram0 partition 7.4G 851.8M   -2
# zramctl
NAME       ALGORITHM DISKSIZE   DATA  COMPR  TOTAL STREAMS MOUNTPOINT
/dev/zram0 lz4           7.4G 848.3M 378.4M 389.9M       8 [SWAP]
#

When I was researching the use of ZRAM there was a lot of information online. A lot of implementations sliced up the zram into multiple slices to enable the balancing of the slices across CPUs, but this is outdated information as the zram support in recent kernels is now multi threaded so there’s no performance advantage to having multiple smaller swap devices any longer, and having a single larger swap space allows the kernel to be more effective in using it.

In Fedora all the pieces of the Fedora implementation are stored in the package source repo. So those that are interested in using zram for other use cases are free to test it. Bugs and RFEs can be reported as issues in pagure or in RHBZ like any other package.

Fedora on the UDOO Neo

Some time ago I backed the UDOO Neo Kickstarter as it looked like a nifty, well featured, IoT device. I got the full option which came with 1Gb RAM and both wired and wireless Ethernet and some add-on sensors. It was a well run kickstarter campaign and the device was well packaged with a fab box. It has both a Cortex-A9 processor to run Fedora and a Cortex-M4 embedded processor to enable you to do Arduino style functionality which should be interesting to experiment with.

For various reasons it has sat around gathering dust, it’s been a bit of a long drawn out process with me randomly poking it as time allowed.. Primarily this was because there was no decent upstream U-Boot and kernel support, and I’d not had the time to hack that up myself from various downstream git repositories, but even without Fedora support their forked Ubuntu distro in the form of UDOObuntu has an experience that is truly terrible!

Late 2016 the problem of a lack of upstream support for U-Boot and kernel changed with initial basic support landing upstream for all three (Basic, Extended and Full) models so with a few cycles over a weekend it was time to dust it off to see if I could get Fedora 26 (did I mention this has been long running?) running on it and to see what worked.

The first thing for me to do was to setup a serial console for easy debugging. The UDOO Neo documentation is generally outstanding and the pins for the UART1 TTL are documented. Two things to note here is that the headers are female rather than the usual SBC male pins so I had to bodge my usual usb to serial TTL with some male-male jumper wires and you’ll need a ground for the TTL which is undocumented on their page, I used one of the GNDs as documented on connector J7 and all was good.

So after an initial set of fixes to the U-Boot support it saw my Fedora install and started to boot! Success! Well sort of, as mentioned above the initial support is rudimentary, it started to boot the kernel and very quickly managed to corrupt and destroy the filesystem not making it much beyond switch root. That wasn’t good. In the last week or two I’ve had a little time to look again, similar issues, it was better than it was a year or so ago but it still ended up with corruption. I reached out to one of the maintainers from NXP that deals with a bunch of the i.MX platforms and I got directed to a handful of patches, a test kernel and image later and a test boot… all the way to initial-setup! SUCCESS!

The core support for the i.MX6SX SoC and the UDOO Neo is pretty reasonable, with the MMC fixes it’s been very stable, all the core bits are working as expected, included wired and wireless network, thermal, cpufreq, crypto and it looks like the display should work fine. There’s a few quirks that I need to investigate further which should provide for a fun evening or weekend hacking. There has also been recently merged support for the i.MX6SX Cortex-M4 land upstream in Zephyr upstream for the 1.13 release, so getting that running and communication using Open-AMP between Fedora and Zephyr should also be an interesting addition. I think this will be a welcome addition to Fedora 29, and not a moment too soon!!

The Raspberry Pi 3 B+ in Fedora

So I’m sure none of you are surprised to hear that I’ve been asked a bunch about support for the Raspberry Pi 3 B+ in Fedora. Well the good news is that it’ll be supported in Fedora 28. Most of the bits were there for the official Fedora 28 beta, it just needed a minor work around, but nightly images since Beta have had all the bits integrated so the upcoming Fedora 28 GA release will support the Raspberry Pi 3 B+ to the same levels as the original 3 B on both ARMv7 and aarch64. The Fedora Raspberry Pi FAQ has now been updated with all the details of both the RPi3+ and Fedora 28.

WiFi

As with the original 3 there’s files with the firmware we can’t redistribute. The details are documented in the Fedora Raspberry Pi FAQ.

You can grab the files like for the 3 although there’s now an extra one, which you don’t really need, but it gives you all the 802.11a frequencies:

$ sudo curl https://fedora.roving-it.com/brcmfmac43455-sdio.txt -o /lib/firmware/brcm/brcmfmac43455-sdio.txt
$ sudo curl https://fedora.roving-it.com/brcmfmac43455-sdio.clm_blob -o /lib/firmware/brcm/brcmfmac43455-sdio.clm_blob

I’ve also done a rpm of the files for both editions of the RPi3 (plus a few other brcm adapters used on ARM boards). You can either just grab the rpm or setup the repo if you want to get the latest if I update it:

$ sudo curl https://fedora.roving-it.com/wireless.repo -o /etc/yum.repos.d//wireless.repo
$ sudo dnf install brcm-firmware

Issues

Like all new things there’s often a few teething problems. Over all the support for the new RPi 3 B+ is relatively solid but like all new devices there’s some bits still bedding down, and the combination of a new device and new OS there’s been a few minor issues that have been seen in some circumstances. We pulled in the latest firmware just before GA to fix some issues but no doubt as it gets wider testing both in Fedora and in the wider Raspberry Pi community more issues may well come up.

The ones we’re aware of are:

  • The new USB hub and gigbit ethernet interface have seen a few issues in some cases. We’ve pulled in quite a few patches to help stabilise the NIC in Linux and it now mostly works in the vast majority of use cases.
  • The USB hub in u-boot, uEFI and grub on aarch64 can cause some issues. If there’s too many USB devices connected it sometimes won’t boot or will do so really slowly. The work around for the moment is to disconnect all the USB devices until Fedora has started to boot and then plug them in.

In Fedora we’ll deliver updates and fixes by the usual updates, in particular as fixes land upstream we’ll review and land them where useful into Fedora, more than likely via a kernel update. If you see a uboot-tools or a bcm283x-firmware update you’ll want to run the rpi-firmware-update command to update the firmware and then reboot for it to take effect.

Older releases

We won’t be supported the new device in the older releases. Why I hear you say? Well it’s possible but it needs update to the firmware, U-Boot and kernel to work. The Raspberry Pi foundation respun Raspbian to support it and basically it’s not straight forward. Much better to have a new shiny Fedora for a new shiny device!

A small 2017 retrospective

I don’t have a huge tendency to do new year resolutions, I’m more the continuous integration type of person where I make a resolution at any time of the year when it makes sense. One thing that did want to achieve as 2017 was starting out was to blog more, with an aim of at least one post a month, and preferably a post every two weeks. I didn’t quite make it with a total of ten posts for 2017, a total of two more than in 2016, so it was a slight improvement! If you count the fifteen draft posts that I wrote in the year, which in most cases just needs some tech details, or a couple of bug fixes to polish the details to actually post, I actually wasn’t that far from the a post every two weeks goal. Let’s see how I get on with this in the new year!

I also started full time into my IoT platform role, and to say it hasn’t been a completely expected roller coaster would be an understatement…. I love a good roller coaster and I think this is the biggest one I’ve ever climbed upon. I got involved in areas of IoT and components of the entire stack that I never thought I would be involved in. I seem to wear about 8 different hats, at last count, and it’s certainly been fun and interesting but busier than I expected, getting pulled into different things that I and others hadn’t planned or anticipated. It’s been a lot of fun, in the Fedora IoT space I didn’t achieve nearly as much as I had hoped but I had also not expected a few of the big blockers and other issues that slowed that down, thankfully it looks like a lot of that is pretty much resolved so I can start driving that forward early in the new year. I have lots of ideas here and this year we’ll start to build the IoT community in Fedora and by the end of the year I believe it’ll be fun and useful!

In the ARM space there was quite a lot of achievements. The big one being the initial support of aarch64 SBCs (finally!), I was very proud of the work we achieved here, it’s a single install path with uEFI/grub2 and a single install path. More work in the short term, by a team of cross team distro people, which took us a lot longer than I’d hoped, but the outcome is a lot better experience for end users and a much more supportable platform for those that need to support it moving forward! It was no means our only achievement with a lot of other ARM improvements including on the Raspberry Pi, accelerated GPUs, initial support for the 96boards platforms. Three is of coarse already LOTS of work in motion for the ARM architectures in 2018 and I’m sure it’ll be as fun and insanely busy as always but I feel we’re now going into it with a good base for the aarch64 SBCs which will rapidly expand in the devices we support moving forward!

Other than that I had a lot of travel, meetings, talks and other things. AFAICT I took around 35 flights, attended around a dozen conferences, numerous meetups and gave around 20 talks! A long with other Fedora and work commitments it was an overall insanely busy year! I somehow, with some of the bangs that 2018 has already shown us (and TBH I blame 2017 for meltdown/spectre) I doubt the coming year will be any quieter than the last… lets see if in among all of that I can meet the ~26 blog posts goal this time around?

Getting started with Fedora on the 96boards Dragonboard

Support for this board has been a long time coming, it was originally announced in March 2015 and shipped later that summer. Two years on we can finally add support for it to Fedora. The enablement here will also assist us with supporting the newly announced 600c and 820c boards more quickly. We’re not all the way there yet, there’s still some firmwares that needs to go upstream into linux-firmware, but the improvement is fantastic and it’s been a pleasure working with the 96boards and Qualcomm teams getting to where we are today.

At the moment we support running Fedora off either an micro SD card or a USB stick. We don’t currently support running off the eMMC and currently basically treat that as the location of the firmware. Anyway lets get started!

Updating the firmware

You’ll want to update to the latest firmwares, my board originally had an old firmware without support for PSCI and so it didn’t bring up all four cores or support reboot. OOPS! You’ll need the latest linux rescue images from the 96boards download site. As I write this the latest is the 17.09 release (version 88). Create a directory for this file before you unzip it because it’ll expand all into the current directory. While there we also need a u-boot build that’s prepared for flashing, the upstream support isn’t quite complete, we add a few patches to the Fedora build to get everything working nicely. You can grab a pre-built version here and also get LK firmware build which enables display output.

You’ll need a host with the fastboot utility, in Fedora this is found in the android-tools package, and a micro USB cable. This process is very similar to flashing a phone with a new image, not surprising given the chipset really. If you have a serial console on the board you can follow along on the console but it’s not required for this board.

To put the board into fastboot mode we hold down the volume down button, labeled as ‘(-)’ near the middle USB port and then power it on. Wait around 30 seconds to ensure it’s booted to fastboot. You can test this with the fastboot devices command. You’ll likely want to run the next commands as root, or use sudo, and be in the directory you created with the extracted firmware and u-boot build:

sudo ./flashall
sudo fastboot flash aboot emmc_appsboot.mbn
sudo fastboot flash boot u-boot.img
sudo fastboot oem select-display-panel adv7533_1080p

The flashall command runs a series of fastboot command to write out various early boot firmware to the eMMC, then we write u-boot out to the boot partition, and finally ensure that output is configured to appear on the HDMI port. Assuming you don’t get any errors from fastboot that should be all the firmware done and in place.

Fedora image and further setup

Next up is the Fedora image. I chose the Workstation image, but we also have a Minimal Image and a traditional Server image. GNOME not the fastest in the world as 1Gb of RAM isn’t really enough for GNOME-3 anymore, but it works well enough. On a USB stick or Micro-SD card (I’ve tried both). We need to write out the image, then expand the rootfs (Note: update XXX for the device you’re writing to):

xzcat Fedora-Workstation-27-1.6.aarch64.raw.xz | sudo dd status=progress bs=4M of=/dev/XXX
sudo gparted /dev/XXX (expand the last partition)
partprobe

Next up we need to adjust the kernel command line slightly, mount up the first partition and edit /EFI/fedora/grub.cfg and search for the string cma=256MB and delete it, then add in it’s place the following console=tty1 console=ttyMSM0,115200n8. Next mount the boot partition (partition 2) and create a sym link

ln -s dtb-4.13.9-300.fc27.aarch64 dtb

. Unmount the partitions and we should be good to go on the Dragonboard.

Plug in a keyboard, mouse (and/or a usb cable for the serial console if you’re going that route) and a HDMI cable, plug in the USB stick or SD card and power it up. If you’re following along on the serial console you should see output straight away, screen might take a little longer.

Once you’ve booted you should be able to complete initial-setup (text or the one from Workstation) and login. To get the WiFI and Bluetooth working you need to install a Radio (WiFi and friends) firmware package which I’ve made into a rpm you can grab from here until it lands into linux-firmware.

What next?

The DragonBoard 410c is pretty functional. I’ve not widely tested sound, the Venus media offload components (we have all the firmware and kernel bits for this), the GPS or some of the other more advanced components but I’ll have more details about those soon. I’ll be documenting the above plus other bits on the Fedora ARM wiki so keep an eye on that or get involved and help out 😛

Overview of aarch64 SBC support in Fedora 27

Support for ARM 64 bit (aarch64) Single Board Computers (SBCs) has been one of the most highly requested features along side the Raspberry Pi. It’s something I’ve been working towards almost as long too. Finally with Fedora 27 I felt we would have enough of the bits in place for a Minimum Viable Product (MVP).

You’ll note from the Fedora change linked above I was very cautious with what we planned to achieve. The change has a very focused list of images: Server, Workstation and Minimal and a limited list of devices: basically the Raspberry Pi 3, the 96boards Dragonboard 410c and HiKey, and a handful of AllWinner devices with a focus on the Pine64 series of boards. The reason for this was I knew there was going to be a lot of low level boot and kernel bits that needed focus and polish and the Fedora 27 cycle was severely limited time and resource wise so the plan was to focus on getting all the core bits into place for Fedora 27 and have a couple of well polished devices and then expand that rapidly for Fedora 28.

The key functionality we were aiming for was a well polished uEFI implementation in u-boot to enable a single install/boot path in Fedora on aarch64 using uEFI/shim/grub2 to boot Fedora on both SBCs and SBSA compliant aarch64 platforms. We now have that platform in place, primarily due to Herculean efforts of Rob Clark and Peter Jones, as well as many others who have provided insight into the deep dark details of the uEFI specification. Fedora 27 will ship with a quite heavily patched, well by Fedora’s standards anyway, u-boot 2017.09 which provides us the core of this functionality enabling us to use a vanilla upstream shim and grub2 to boot a standard Fedora. All this work is already upstream, or making it’s way there in 2017.11. In Fedora 28 there will be even more improvements that will enable us to do a bunch of other cool stuff (that’ll be a post for later!) and also enable much quicker upstream board enablement now all the core bits are in place.

So what do we actually support? Well all the usual bits that you would expect on a standard Fedora install, whether it be x86_64, ARMv7 or aarch64, like SELinux, containers, desktops and all the other bits. There’s a few bits and pieces that are a little rough around the edges but overall the feature is pretty robust. On a board by board feature set lets break the this down across the boards:

Raspberry Pi 3

The support for the Raspberry Pi3 is the equivalent to the ARMv7 support but with boot via uEFI/grub2. The memory isn’t quite as good as on 32-bit but that’s to be expected, overall it’s pretty reasonable for a device of the specs and cost. Like on 32 bit support we’re seeing regular improvements each release and throughout the releases. The aarch64 support for the RPi3 is just an evolution to this.

DragonBoard 410c

The support for the DragonBoard 410c is looking pretty decent. Qualcomm has been doing a pretty decent effort to get stuff upstream, we have firmwares for the GPU and for video decode/encode upstream as well, along with kernel drivers and the open freedreno 3D drivers, HDMI audio should work as well. The WiFi firmware isn’t yet upstream but I’ll document how/where to get that and hopefully that should be in linux-firmware soon as well. Overall I’m quite happy with the status of this device, although like all devices with 1Gb RAM it’s a little constrained, but that should make the newly announced 820c with 3Gb of RAM a decent device ;-). All the details for getting it running will soon be in the Fedora 96boards wiki page.

HiKey

Most features and functionality of the HiKey are supported, note this isn’t the HiKey960 (look to F-28 for support for that), except accelerated graphics due to the use of a MALI GPU. Other than that the functionality is pretty decent. You’ll likely want the latest tianocore firmware and the details for that can be found on the Fedora 96boards wiki page.

Pine64 (AllWinner A64 SoC)

We actually should have a number of devices based on the AllWinner A64 SoC working here but we’ve only tested the 3, 2Gb/1Gb/512Mb, Pine64 device sizes. The support for these devices is headless and you will need a serial console else you’re on your own as none of the display bits in the kernel have made it upstream, and of course the GPU is a MALI 400 series so when it does it won’t be fast. The support for the rest of the device is basic, it’s usable for a headless server style device, we support network, USB, KVM, RTC and a few other bits. Other than display we don’t yet support the SDIO attached wireless, sound, crypto offload or any of the other media interfaces. A lot of this is under review upstream so I think Fedora 28 should look much better for this series of devices and 4.15 might even bring very basic console output. Speaking of series of devices which ones should actually work other than the three Pine64 devices? Well the following A64 SoC devices have a Fedora built u-boot and kernel DT support so should work as well as the Pine64: BananaPi-m64, OrangePi Win, SoPine baseboard (PineBook boots if you’re happy with serial console), NanoPi-A64 and the A64-OLinuXino. We had some troubles with the AllWinner H5 SoC devices earlier in the cycle but I’ve had a couple of reports that it seems to be resolved so they should work too and that adds the Orange Pi PC2, Prime and Zero+ 2 as well as the NanoPi NEO2. So that’s around a dozen or so devices! 🙂

Other ARM64 SBCs

I’ve had reports that other aarch64 SBCs boot on Fedora just fine. I’ve not listed those where I can’t verify whether they boot with our uEFI enabled u-boot. Looking around on my desk I do have a number of devices that I expect us to be supporting in Fedora 28, or maybe even just enabling u-boot bits in a F-27 update.

Overall I’m pretty happy with the state of Aarch64 SBCs for Fedora 27 and what we’ve managed to achieve is such a short cycle!

Why I’m not backing the Purism Librem 5 phone

NOTE: This is a post about my opinion on the device, hence the title of “Why I’m not backing….”, people have been explicitly asking me why I’ve not backed it, this documents it so I don’t need to keep repeating myself!

Numerous people have come up to me and asked “So will you get Fedora to run on your Librem phone?” and when my response is “No, I’ve not backed it” I get weird looks with a question of “Why?” I had thought it was time to do document my concerns with this laudable venture. This was certainly further qualified when I had to inform someone from the EFF that they’re delirious about the “it doesn’t need closed source firmware” on the i.MX chips that are being proposed. While I applaud the general principals and ideas of a fully open rights protecting phone I can’t help but feel that the group doing it either are being false at worst or naive at best with some of their statements.

Firmware

Their site claims “The i.MX 6/8 CPU will be completely free software without any binaries whatsoever!” while this could be sort of true if you want a hobbled device it’s not really the case at all. There’s a number of firmwares that are needed to make a number of pieces of functionality of a mobile device useful. Firstly the SDMA driver needs a firmware to run at any level of reasonable speed. Secondly the accelerated media decode, which will be required if you actually want to consume meda on your phone and have more than moments of battery life, also needs firmware. You’ll note on the media decode I don’t reference the actual firmware! Why? Because it’s not actually distributed in the linux-firmware repository so you have to request it from NXP or somewhere with appropriate signups (while I’m not 100% sure I believe having a driver in the linux kernel without required firmware in linux-firmware is a breach of the requirements of said driver in the upstream kernel).

i.MX6 or i.MX8 SoC

The i.MX 6 or i.MX 8 option concerns me. Basically make a choice! The core issue I have is the i.MX6 SoC is ancient, being announced back in Jan 2011, and based on a Cortex-A9 SoC. It doesn’t support USB-C so to ship as promised they’ll need to add a PCI-e attached USB-3 controller, charging circuitry, and maybe even a LVDS to Display Port option plus a chip to MUX the three through the actual USB-C port if they want to be able to ‘dock’ which will make power consumption on the phone even worse! If I’d spent $600 and received a phone in January 2019 based on an eight year old 32-bit chip design I’d be seriously pissed off. To be quite honest the i.MX8 is no spring chicken either, being announced in September 2013, it’s based on the already quite long in the teeth Cortex-A53 SoC design which was the original aarch64 “little” low end design but it does at least support USB-C on the SoC. The i.MX SoCs are generally quite nice, and the i.MX6 is quite well supported upstream, but the whole line of SoCs are more targeted towards embedded applications like cars, so they do have a long support cycle. They’re not a mobile phone focused SoC though, they also tend to be quite slow to get moving in the market/availability and the i.MX8 has little upstream support in the kernel as yet. Sure the etnaviv driver enables an open 3D accelerated driver, one that isn’t supported by the vendor so doesn’t enable all the features upstream, but you can run this today on Fedora 26/27 on the i.MX6 SoCs now but a free GPU driver is not the only reason to choose a platform.

General concerns

I can’t help but feel that there’s going to be a lot of disappointed people that will end up receiving an expensive sub standard device, probably late, that ends up not taking us much further along the road, to a fully open rights protecting phone, than the days of the Nokia n9xx series phones running Maemo. The HW, if based on an i.MX6, will certainly not be much further along that route and still stuck in the past in terms of HW. I personally believe the project would be better off engaging with the Qualcomm community team to use the Qualcomm 820c/600c SoCs because there’s an open driver that Qualcomm are working to improve, and while there’s a need for firmware for GPU and media offload, in reality it’s no better/worse than the i.MX devices with the bonus that their devices that are more current and aimed at mobile workloads with the vendor actively working to upsteam the enhance the support support of their SoCs.

Ultimately I think a device that closely resembles the specs of a phone of recent history, than that of ancient smart phone history, is likely to get a better following and hence a better software ecosystem than one that’s the same era as the Motorola DROID 4.

Update: So just to clarify a few things that were bought up in a couple of threads:

First this reddit thread:

  • I don’t question the value of hardware isolation of the various wireless interfaces that they claim, or the ability of them being able to deliver that bit on what ever SoC they choose, that is why it wasn’t addressed above.
  • I bought up USB-C because it is explicitly claimed as a feature.
  • I bought up display port because given USB-C above and their “It can be a desktop computer and phone all-in-one” claim, including a nice picture, then DP over USB-C is the only real option for that functionality.
  • Qualcomm chips were an example of other SoCs I believe might be a better fit, yes they do provide options in their APQ line without built in radios, they were just an example of another option, there are other possibilities, it wasn’t meant to provide a guarantee, it was an alternative example.
  • Yes, all ARM processors have onboard boot firmware, it’s generically referred to as “PBL” (Primary Boot Loader), so do most other processors.

Second this purism thread:

  • Sure you can disable the media engine with an e-FUSE but media is kind of useful for a lot of use cases like video config, audio calls and music, not just watching videos. If they choose this route, I would hope they leave the fuses unblown and document how a end user can do it else the $600 is a whole lot less useful for a lot of people

Ultimately the main points is I have to make about all of the above is two fold, firstly it’s my opinions and why I didn’t back the device, and secondly there appears to be some large discrepancies in the statements they make about the device and SoCs which I would not have expected and that is the key problem for me because it causes concerns over the ability to deliver a working and useful device to me. $600 is not a small amount of money, for me at least, to hedge on what *might* be a i.MX8 or might be an old 32 bit i.MX6.