In the hunt for a development machine, I got to the next phase. I did some shopping, and there it is: my own Ampere Altra-based system.

Why?

As you may have read in my previous post, I used several AArch64 systems for local development. And the latest one, an Apple MacBook Pro, is nice and fast but has some limits — does not support 64k page size. Which I need for my work.

Let’s go cheap

So I have decided to buy myself an Ampere Altra system. As cheap as possible.

AArch64 server parts

The only part I needed to buy brand new was a motherboard. And the only “affordable” one was AsrockRack ALTRA8BUD-1L2T, which was a product for data centres (so I was told).

Next, a used processor. At first, the idea was to buy a Q64-22 (64 cores, 2.2 GHz clock), but when the seller on eBay was not responding, one of my friends decided to upgrade his Altra systems and offered me a Q80-30 (80 cores, 3.0 GHz clock).

The CPU requires cooling. There are not many options for the LGA 4926 socket. I found an Arctic Freezer 4U-M in one of the online stores here in Poland and bought the only one they had.

Each computer needs memory, right? So I went with used ones again, as there is a huge market for used server parts. I bought eight sticks of 16GB each; the model is SK-Hynix HMA82GR7CJR8N-XN. It is not present on the official Ampere Altra memory Approved Vendor List.

Working fine (after re-seating three sticks):

DRAM populated DIMMs:
  SK0 MC0 S0: RDIMM[ad:80] 16GB 3200 ECC 2R x8 RCD[32:86] HMA82GR7CJR8N-XN
  SK0 MC1 S0: RDIMM[ad:80] 16GB 3200 ECC 2R x8 RCD[32:86] HMA82GR7CJR8N-XN
  SK0 MC2 S0: RDIMM[ad:80] 16GB 3200 ECC 2R x8 RCD[32:86] HMA82GR7CJR8N-XN
  SK0 MC3 S0: RDIMM[ad:80] 16GB 3200 ECC 2R x8 RCD[32:86] HMA82GR7CJR8N-XN
  SK0 MC4 S0: RDIMM[ad:80] 16GB 3200 ECC 2R x8 RCD[32:86] HMA82GR7CJR8N-XN
  SK0 MC5 S0: RDIMM[ad:80] 16GB 3200 ECC 2R x8 RCD[32:86] HMA82GR7CJR8N-XN
  SK0 MC6 S0: RDIMM[ad:80] 16GB 3200 ECC 2R x8 RCD[32:86] HMA82GR7CJR8N-XN
  SK0 MC7 S0: RDIMM[ad:80] 16GB 3200 ECC 2R x8 RCD[32:86] HMA82GR7CJR8N-XN

Ordinary parts

The rest of the parts are ordinary ones available in any random store.

Computer case

The case was a challenge. The ALTRAD8UD-1L2T is a “deep MicroATX” case which means I needed a case that can take an EATX motherboard (they are a bit deeper than ATX).

I looked through pictures in online stores and selected about 10 candidates. Then I started watching reviews and crossed several of them out. Usually because the holes for handling cables were too close to the edge of a board.

Finally, bought an Endorfy 700 Air case. It came with five 120mm fans (3 at the front, one at the rear, and one on the top). There was a lot of space behind the motherboard’s plate for cabling, and a fan splitter so I could connect all five case fans as one to the motherboard.

And the PSU compartment has an extra hole for PCIe power cables!

Power supply

The ALTRAD8UD-1L2T motherboard is expected to be powered by a 12V PSU only. There is no connection port for the ATX 24-pin plug. Instead, there is an adapter that takes only power-on and power-good signals from it and connects to the small 4-pin port on the motherboard.

There are three EPS12V connectors available. I used two of them. The MSI MPG A850G power supply I bought comes with two such cables and has an option for connecting the third one.

The rest

What else is needed? Some NVME for storage (Lexar LM790 2TB) and a random graphics card (Radeon Pro WX2100, remembering the old times).

And I was ready to build the system.

How does it look?

The motherboard feels small once the CPU cooler is mounted. And a low-profile graphics card is almost invisible:

The back side of the case shows where I hid the ATX 24-pin adapter and most of the cables.

Price

Of course, if I went cheap, then how cheap did it end up being?

Component	Price in PLN
motherboard	4 068
CPU	1 341
CPU cooler	191
memory	696
case	395
power supply	529
NVME	522

In total: 7 732 PLN (around 1 800 EUR). About 500 EUR more than I anticipated at first.

The CPU upgrade to 80 cores was extra 100 EUR. Memory came from another seller, as the first one ended their sale before I was ready to buy, resulting in an extra 40 EUR to the cost.

At first, I wanted to have 2GB per core, but with the CPU upgrade, I would have needed to go to 256 GB, and that would have been another 250-300 EUR extra.

The case had a 30% discount due to a promotion, and the NVME came with 7% cashback.

Plans

My current plans for this system are:

• install Fedora 42
• create some VM instances:
  • RHEL 9/10
  • CentOS Stream 9/10
• put a Radeon RX 6700XT inside and check whether it would work
• try to use it as a desktop
• do some crazy experiments

All Linux systems will have both 4k and 64k page size kernels.

Since I started working on AArch64 in 2012, I had been searching for a hardware system capable of meeting my development needs at home.

Pinkie Pie

In 2014, I got Applied Micro Mustang for home use. I named it “pinkiepie” because I use cartoon-character names for all systems I manage. As both mustangs and ponies are types of horse, I decided that Pinkie Pie from My Little Pony was a good name.

This system served me for several years. With 8 cores, 32 GB RAM, and a SATA SSD for storage, it was enough to build RHEL, Fedora, Debian packages — both for AArch64 and 32-bit armhf.

It failed in 2020, was briefly operational again, but has remained non-operational since then.

Jagular

In 2021, I acquired Solidrun HoneyComb. With sixteen Cortex-A72 cores, 32 GB RAM, and SATA SSD from Pinkiepie, it served as my development machine.

Jagular’s name came from Winnie the Pooh, as this system was my own (unlike Mustang, which was Red Hat property).

Applejack

In 2023, I took a chance and replaced my work laptop with an Apple MacBook Pro. This gave me more modern Arm system with 10 fast CPU cores.

The system’s name came from My Little Pony again.

After the installation of Fedora Asahi Remix on the MacBook, I moved Jagular to storage. And sold it in 2025.

Radxa Orion C6

At the end of 2024, Radxa announced “World’s First Open Source Arm V9 Motherboard”. Several people sent me a link to it. I checked the details and decided to wait for any delivery of source code for this platform.

It took them months to release anything. An old (6.1.44) version of the Linux kernel, a copy of EDK2 sources with 23 megabytes of binary blobs.

The good part was that this SBC was capable of running mainline Linux in ACPI mode. Still, the situation regarding it was far from what their slogan promised…

There were “Arm V9” systems before it, so it was not “World’s First”. There were several blobs needed to run it, so it was not “Open Source”.

I was following discussions on their forum and had a feeling that the more I read, the more I postpone my order.

RHEL and 64k page size

After leaving Linaro, I started working on Red Hat Enterprise Linux (RHEL) again. And while most things can be tested in virtual machines on Applejack, some tasks require booting 64k page size kernels — which the Apple M1 Pro is not capable of doing.

As a temporary solution, I used remote boxes from the Red Hat lab, but it was not a good solution in the long term.

So I started searching again.

Ampere Altra?

In May 2024, I was one step from spending 3000 EUR on Ampere Altra-based workstation — the typical 64-core system with 128 GB of RAM.

I did not buy it because at that time I was looking at it as an Arm desktop. Which, at that time, it was far from — issues with AMD cards under Linux required patching the kernel with out-of-tree patches.

Time passed, and I looked at it from different perspective — as “a tool”.

You see — I split computer systems into two categories: tools and toys.

Tools

Computer systems that have to do their job are tools. My x86-64 desktop is a tool. The same with my x86-64 NAS, or my work laptop.

Those systems have their purposes and have to work. No tricks with firmware or operating systems, no self-built kernels.

Toys

Some of my computer systems are toys — I do not expect them to be stable, I do not keep any important data on them. This is where most of my Arm SBCs (Seriously Bad Computers) are.

FriendlyElec NanoPC-T6 LTS on my desk is a perfect example. At this moment, it is offline and waits for firmware reflashing because I flashed the wrong file two weeks ago. And many times I have run my own kernels on it to check some out-of-tree patches adding support for hardware features.

Ampere Altra then

In March, I decided to go and buy an Ampere Altra system. But this time, on a budget: a brand new motherboard, a used CPU, used RAM.

I got a few people interested, tried to find a way for some kind of discount and started ordering parts. During the next few weeks I will build myself yet another development machine. For my own use, at home.

But this will be another blog post.

The June syslog-ng newsletter is now on-line: Installing nightly syslog-ng arm64 packages on a Raspberry Pi Working with One Identity Cloud PAM Linux agent logs in syslog-ng Testing the new syslog-ng wildcard-file() source options on Linux It is available at https://www.syslog-ng.com/community/b/blog/posts/the-syslog-ng-insider-2025-06-arm64-pam-testing syslog-ng logo

A few days ago, I was asked “Do you use AI?”. I also noticed some recent blog posts and emails about it.

So let me explain how I use AI on this blog.

Grammar stuff

If you look through this blog’s archives you may see how terrible my English grammar used to be (not that it is particularly good nowadays). Back in 2010 I started attending English lessons to improve both my grammar and vocabulary. And it paid off.

Still, my writing style can be described as ‘Hrw English’. I try to keep it British English rather than American (I learnt the British version throughout my formal education). My vocabulary includes words from American, Australian and other variants.

Let me ask ChatGPT

My previous post about my arm problem went through ChatGPT:

Take text below, fix grammar to British English and show changes side-by-side

So let take some examples from the result:

Original	ChatGPT
Some time ago my right arm started to hurt. In the beginning it was a bit annoying but then got serious. So I went to orthopaedist.	Some time ago, my right arm started to hurt. At first, it was a bit annoying, but then it became serious. So, I went to an orthopaedist.
MRI was the other thing. My daughter told me already that I need to be prepared for loud annoying noise. What I was not prepared was some shitty pop played over some shitty headphones. zOMG…	The MRI was a different matter. My daughter had already told me to expect loud, annoying noise. What I wasn’t prepared for was some dreadful pop music played over equally dreadful headphones. zOMG…
I know that those headphones have to be as plastic as possible to survive inside of MRI machine but next time I will ask for not playing anything.	I know the headphones have to be made entirely of plastic to survive inside the MRI machine, but next time I’ll ask them not to play anything.
BTW: Can anyone recommend some good Linux application for watching DICOM images (other than AlizaMS)?	By the way, can anyone recommend a good Linux application for viewing DICOM images (other than AlizaMS)?
Most of cases are related to professional sport people. A group where I do not belong.	Most cases are related to professional athletes—a group to which I definitely do not belong.
I promised my daughter some trips during summer vacations. One set of attractions will slip for 3rd time.	I promised my daughter some trips during the summer holidays. One set of attractions will be postponed for the third time.

Review of suggestions

So what do we have here? Some vocabulary changes like “professional sport people” -> “professional athletes”, the usual missing “a/an/the”, more commas, work order changes etc.

I review the suggestions and choose some of them: vocabulary tweaks, missing articles and some word order changes. Most of the time I ignore punctuation related ones as I always had problems with commas — never mind which language I was writing in.

I am leaving several errors as they were — I wrote that text so let it show my mistakes instead of hiding them. And I prefer to use ‘some’ rather than ‘few/a few’ as I always forget when to use which ;D

Style variations

After some updates I got question:

Let me know if you’d like it shortened, rephrased in a more humorous or technical tone, or formatted for a specific blog engine like Hugo, Ghost, or WordPress.

Fun starts with mixing all of them:

updated text below. give me British English version, differences side-by-side and then more technical tone version and more causal one plus more humorous one. render markdown

Look how one sentence can change:

• So, let me explain how I use AI on this blog.
• In this post, I’ll outline how I integrate AI into the workflow of this blog.
• So, here’s a quick explanation of how I use AI on this blog.
• Yes, I use AI. But not for evil. Mostly.

Those are original, technical, causal, humorous versions.

I have to admit that the ‘more technical tone’ version was a bit too much for me. It looked completely different from my writing style. Anyway I read it twice and took some wording from it. The ‘more causal’ version looks what I could write but still not quite my style. And the humorous version is pure fun to read but nothing to copy from it.

Conclusion

Will I continue to use of AI on this blog? Probably yes. It gives me good suggestions and fixes some of my language/grammar mistakes. And I do not have to apply all the changes so text keeps my style.

Is it a way for other people? No idea but I recommend giving it a try and seeing it for yourself. The results can be interesting.

Some of my readers may say that I’ve always had a problem with Arm. The company, the architecture, the ecosystem etc.

But today I have some other problem. With arm.

Beginning

Some time ago my right arm started to hurt. At first, it was a bit annoying but then got serious. So I went to an orthopaedist.

Orthopaedist

“Please stand up, please stand up” and then started moving my arms in various ways. We wrestled a bit, I did that “fasten your bra” move and the diagnosis was clear: I had a problem with my arm.

I received some sick leave and a list of imaging tests to do.

Imaging

I went through the standard imaging flow:

• RTG (X-ray)
• USG (Ultrasound)
• MRI (Magnetic Resonance Imaging)

I visited a few places and received some optical media with images (and some in physical form).

The RTG and USG were simple. I have done them many times before.

The MRI was another matter. My daughter had already told me to be prepared for loud, annoying noise. What I was not prepared for was some awful pop music played through dreadful headphones. zOMG… I know that those headphones have to be as plastic as possible to survive inside the MRI machine but next time I will ask them to not play anything.

BTW: Can anyone recommend a good Linux application for watching DICOM images (other than AlizaMS)?

Second visit to orthopaedist

“Looks like you have SLAP (Superior Labrum Anterior Posterior).” Yay…

Most cases are related to professional athletes — a group to which I definitely do not belong.

I think I need to rethink my life desk setup.

But first have to sort out stuff and how my summer time will look like. Rehabilitation, exercises, handling things with only one arm (not my dominant one).

So if you would like me to do something for you before September then consider postponing as I may not be ready for it.

Summer holidays

I promised my daughter some trips during summer break. One set of attractions will be postponed for 3rd time. We have some conventions to attend and they should be safe.

The trip to visit my mother will be more difficult. No way to drive a car means that I will have to rely on public transport. In area where it sucks. Should be fun (for some sick definition of “fun”).

While most Java-based drivers have been deprecated in syslog-ng years ago, we have recently removed all of them in preparation to syslog-ng 4.9.0. Right now, the only Java-based driver remaining is HDFS, so we want to ask the syslog-ng community if the HDFS destination is still needed for them. Read more at https://www.syslog-ng.com/community/b/blog/posts/deprecating-java-based-drivers-from-syslog-ng-is-hdfs-next syslog-ng logo

Last year, syslog-ng 4.8.0 improved the wildcard-file() source on FreeBSD and MacOS. Version 4.9.0 will do the same for Linux by using inotify for file and directory monitoring, resulting in faster performance while using significantly less resources. This blog is a call for testing the new wildcard-file() source options before release. Read more at https://www.syslog-ng.com/community/b/blog/posts/testing-the-new-syslog-ng-wildcard-file-source-options-on-linux syslog-ng logo

One Identity Cloud PAM is one of the latest security products by One Identity. It provides asset management as well as secure and monitored remote access for One Identity Cloud users to hosts on their local network. Last year, I showed you how collect One Identity Cloud PAM Network Agent log messages on Windows and create alerts when somebody connects to a host on your local network using PAM Essentials. This time, I will show you how to work with the Linux version of the Network Agent.

Finally, a new syslog-ng release! As you can see from its version number, this is a bug fix release. It took a bit longer than expected, as we wanted to release it in sync with syslog-ng PE, the commercial variant of syslog-ng. 4.8.2 serves not just as the foundation of the new syslog-ng PE release, but also provides fixes to 4.8.1, which is included in major Linux distributions. This update ensures that all our recent bug fixes reach the majority of our users.

While no dates are set to stone yet, we expect a couple of syslog-ng releases in the near future. As version 4.8.1 is used in major Linux distributions and has a couple of known bugs, we will release 4.8.2 to address those. However, we are also working on 4.9.0, which will bring many changes. Read more at https://www.syslog-ng.com/community/b/blog/posts/a-call-for-testing-the-upcoming-syslog-ng-releases syslog-ng logo

From my previous Active Roles blogs, you could learn how to forward regular Active Roles logs from Windows Event Log to a central syslog-ng server, where it parses, filters, stores and forwards the logs. In this blog, I show you how to work with Active Roles debug logs, that is reading them using syslog-ng Agent for Windows and forwarding them to a central syslog-ng server for long(er) term storage. Debug logs are typically huge and the Active Roles debug logs are no exceptions, so you must make sure that you collect them only when really necessary.

The April syslog-ng newsletter is now on-line: Testing Elasticsearch 9.0.0 beta1 with syslog-ng Working with parsed Active Roles logs in syslog-ng Running syslog-ng PE in RHEL UBI It is available at https://www.syslog-ng.com/community/b/blog/posts/the-syslog-ng-insider-2025-04-elasticsearch-beta-active-roles-rhel-ubi syslog-ng logo

I regularly visit Seoul, and for the last couple of years I've been doing segments from the Seoul Trail, a series of walks that add up to a 150km circuit around the outskirts of Seoul. If you like hiking I recommend it, it's mostly through the hills and wooded areas surrounding the city or parks within the city and the bits I've done thus far have mostly been very enjoyable. Everything is generally well signposted and easy to follow, with varying degrees of difficulty from completely flat paved roads to very hilly trails.

The trail had been divided into eight segments but just after I last visited the trail was reorganised into 21 smaller ones. This was very sensible, the original segments mostly being about 10-20km and taking 3-6 hours (with the notable exception of section 8, which was 36km) which can be a bit much (especially that section 8, or section 1 which had about 1km of ascent in it overall). It does complicate matters if you're trying to keep track of what you've done already though so I've put together a quick table:

Original	Revised
1	1-3
2	4-5
3	6-8
4	9-10
5	11-12
6	13-14
7	15-16
8	17-21

This is all straightforward, the original segments had all been arranged to start and stop at metro stations (which I think explains the length of 8, the metro network is thin around Bukhansan what with it being an actual mountain) and the new segments are all straight subdivisions, but it's handy to have it written down and I figured other people might find it useful.

About a year ago I wrote a post about a need for a good keyboard. And that I am thinking of making mechanical one.

During last week I built one.

Repetitive strain injury

Twenty years ago I had a problem with repetitive strain injury (RSI in short). Typing on normal PC keyboard was quite impossible for me. Someone recommended buying one of those “fancy” ergonomic ones. And that’s how I ended with Microsoft Natural Keyboard Elite.

It was a great help. Got used to layout and my RSI problems were going away. Then bought second one to have one at home and one at work.

Time passed, keycaps were worn up so I got another keyboard — you can read the story in last year post about keyboards.

Let’s get mechanical

During last year I looked at several keyboards. Checked KMK, QMK, ZMK and other solutions of keyboard firmware. Watched countless videos on how to make keyboard and read many articles about it.

One layout caught my eye: TGR Alice as it was quite simple ergonomic one. Then found Arisu which added cursor keys and did some other changes. And finally Adelheid which added function keys.

My Adelheid

As those layouts are available on MIT license I took Adelheid and altered its layout a bit:

• added Meta key
• moved Backspace to were it is on PC105
• moved Backslash to above Enter key

Parts and costs

I had some parts already (ordered then to my MS4KMech project which I abandoned in meantime) and decided that my first keyboard will be as cheap as possible:

• Redragon switches (11€ on Aliexpress)
• cheap ABS keycaps (5€ oin Aliexpress)
• Ultimate Pico RP2040 (2.7€ on Aliexpress)
• hundred 1N4148 diodes (2€)

So cost of parts was about 20 EUR.

As a way to keep costs down I printed plate in two pieces (Ender 3 table size was a limit) and used some screws and plates to keep them together.

Soldering rows and columns

Mounted switches and started soldering. Simple diode to diode to create rows. And it is visible that I watched some video after first 3 rows of one half ;D

Instead of forming diodes I soldered them and then use small pliers to form rows. Later started cutting legs before soldering.

Columns were easier. Spool of kynar and stripping isolation using nails did the job.

Added a bunch of colour wires to be able to test did my soldering resulted in something working. Grabbed Raspberry Pico from a drawer, flashed it with QMK firmware and testing bench was ready:

And it almost worked! One key was not working. Turned out that diode was soldered in wrong direction. Happens.

For second half I took a bit different strategy. Decided to solder columns first as they took me more time than rows. Whole half went much faster then first.

Then all that left was finding a way to connect both halves (I did not wanted to make a split keyboard) and connect rows. I used pieces of old universal PCB and screws:

Connecting controller

I bought Ultimate Pico RP2040 especially for keyboard project months ago. Compared to Raspberry Pico this board had USB-C port and one line more (GP29). Also less GND pins so 17 GPIO lines are on one side. My keyboard has 16 columns so this allowed me to have all columns connected on one side of board and use second side connections for rows:

I may do not look nice but it works :D

Powering up

I flashed QMK and started testing. Turned out that one column is not working. Soldered wire and it was fine.

3 other keys were not working. One needed soldering but other ones looked fine.

QMK has a solution for it: “Console debugging”. I added a bit of code to keymap.c file:

bool process_record_user(uint16_t keycode, keyrecord_t *record) {
  // If console is enabled, it will print the matrix position and status
  // of each key pressed
#ifdef CONSOLE_ENABLE
    uprintf("KL: kc: 0x%04X, row: %2u, col: %2u, pressed: %u, "
            "time: %5u, int: %u, count: %u\n",
            keycode, record->event.key.col, record->event.key.row,
            record->event.pressed, record->event.time, record->tap.interrupted,
            record->tap.count);
#endif
  return true;
}

This allowed me to notice that both Enter and End keys are seen. Turned out that I wrote wrong row/column values for both of them.

Mistakes

I made some mistakes with this keyboard. They were expected as usual with first attempts.

First mistake: 2mm plate. Should had do 1.5mm to make switches keep better. Total thickness can be even 3-5 mm but with holes done in a way that switches hold properly. Also it may make keyboard less bendy.

Second ones - holes for stabilisers. I have a few 2u - 2.75u keys and they should get stabilisers. But I did holes in wrong way so the ones I have do not fit. Will see how it affects keyboard use.

Third one: soldering. I do not remember when last time I soldered so many points. So some of other are shitty. Just had to fix one more…

Tools

Community behind mechanical keyboards created many tools. Let me try to list ones I used.

First comes QMK firmware which allows to forget about programming microcontrollers as keyboard is done as simple JSON file and one C file with keymap definition.

Then Keyboard Layout Editor which allows to define layout of keys in any crazy way you want. Raw data from it is a base for several other tools.

Keyboard Firmware Builder is a simplest way to manage wiring of keyboard. You paste raw data from previous tool and start simplifying wiring. It can also generate ready to use QMK firmware images for several popular micro controllers (but no RP2040 I used).

Plate & Case Builder also takes layout data. And allows to choose switch/stabilizer types plus a bunch of other options. Then generates plate for your layout with option to download it in SVG, DXF, EPS file formats. I took SVG, extruded to 2mm and had a base plate for 3d printing.

Was it worth the time?

The question for the end: was it worth the time? I think that it was. Learnt new things, got something working. Something I did on my own. Sure, it is far from being perfect but gives me area for improvements.

I have two space keys and already see that for most of time I use right one. There are no layers defined yet and I have a key for it. Some keycaps may get replaced with other ones (like Calculator one next to left Space begs for lower one).

I have links to many interesting layouts so who knows, maybe I will make a new keyboard soon. Split, orthogonal one. With volume knob. Will see.

About seven years ago I visited friends. They were playing some game and told me that I may like it. I asked some questions, looked at their play etc.

Few days later I spent 70 PLN and bought Factorio.

3500+ hours later

Beginning

The idea of Factorio is simple: you crash land on some planet and do everything needed to send rocket to the space. Mine resources, turn ore into plates into electronic components into assembly machines. And use those machines to build more of everything.

And walk around to look for more resources.

And fight local fauna as they do not like you or your effect on their planet. Pollution made by your factory speeds up their evolution so you need to develop stronger weapons etc.

Addiction

It is easy to get addicted to the game. You start game ‘for an hour’ in the afternoon and ‘one hour later’ you see sunrise through the window. And your legs remind you that you were sitting in front of computer for 12+ hours straight.

Been there, done that. Then learnt how to make pauses and how to put some limits.

Achievements

Like many games today Factorio has set of achievements. Most of them you get during normal play. Some require special preparations.

I remember doing “Getting on track like a pro” one. It requires building a locomotive within the first 90 minutes of the game. Done on second attempt cause I did not noticed that locomotive needs to be put on the rail and I went out of time on first attempt.

The one I liked the most was “Lazy bastard” where you can manually craft no more than 111 items. This requires counting each move at the beginning of the game as there are only 10 items reserve. Great way to learn how to make factory produce things.

Took me over six years to get all achievements.

Rocket ones

To win the game you need to send rocket to space. And there are three achievements for it:

• “No time for chitchat” if you finish within 15 hours
• “There is no spoon” if finished within 8 hours
• “So long and thanks for all the fish” if you send fish into space

And there is funny one — you can send yourself into space and game handles it properly. The way is to put vehicle (car, tank, locomotive) into a rocket and then ‘take a seat’ in vehicle. You get the usual animation of starting rocket but from different perspective ;D

Space Age DLC

In October 2024 Factorio team released 2.0 version of the game. And “Space Age” DLC.

I decided to play 2.0 first to see what changed first. And there were many changes! And new achievements :D

Took me some time but once I finished I bought “Space Age” DLC and started playing again. Without rush or pressure to get complex factory.

After 260 hours I finished but did not had a feeling that I won anything. In normal game you build a factory to send satellite to the space. In DLC you have to make a useless ship and send it “where no man has been before”.

But that does not give anything. Going to the “shattered planet” gives 12th science pack but it is needed only to speed up research.

Mods

But basic game is just a piece of game. The most important piece of course but there are so many mods for Factorio…

Some are simple ‘quality of life’ ones, some are ‘why this is not in game’ (like Rate Calculator).

And there are complex ones like ‘Krastorio 2’, ‘Space Exploration’ or ‘Warptorio’ which change game, add lot of buildings, technologies, change winning requirements etc.

You can enable countless mods to make your game completely different. Easier, harder, more complex etc.

Multiplayer

One of things I like is multiplayer. I had own server where I played countless hours with friends. We did ‘Krastorio 2’, we did huge megabase.

We rebuilt half of megabase after visit of biters — it was ‘rebuild or start from scratch’ moment and we decided that starting from scratch is boring. Turned out that rebuild took only few hours once we got all reactors running so electricity was not a problem.

Community

One of great things around Factorio is community. Modders, youtubers, speedrunners (I learnt many tricks watching speedruns), people on forums etc.

I remember how one day Factorio was misbehaving on my system. Everything else was working fine, problem was only visible when I played the game. Reported a bug, got contacted by someone from development team and we joined multiplayer game. Some minutes later I got “Please check stability of your system. Run memtest or something”. And few hours later memtest86 shown some issue with memory config.

Final words

Money spent on Factorio was the best spend cash when it comes to my game related expenses.

And there are still four achievements to get :P

In 2012 I wrote “What interest me in ARM world” post. Listed hardware I played with and some words about what I was looking for.

Decided that it is time to write 2025 one. About AArch64.

Hardware I used

When it comes to AArch64 world I skipped most of Seriously Bad Computers and concentrated on servers.

Servers

During all those years I have used several servers. From X-Gene to NVidia Grace. Some for longer, some for shorter tasks. Most of those systems were in Linaro or Red Hat labs and I used them remotely.

At home I had AppliedMicro Mustang. It served me as a desktop in 2014 (for a few days) and as main development system for a few years.

Seriously Bad Computers

Of course there were SBCs too. Pine64, RockPro64, Espressobin, Honeycomb, NanoPC-T6. Even Raspberry/Pi 3 — which I bought on announcement day to check how bad it was.

I do not use any of them to run some service at home.

Used Espressobin as my router for a month or two. Honeycomb was my main development box (until Macbook replaced it). Rest of them I bought to check how SBC market looks like.

Macbook

I use Apple Macbook Pro (14”, M1 Pro, 2021) as my work laptop (Red Hat provided). It runs Fedora 41 Asahi Remix and serves me for almost 3 years. During that time I ran MacOS maybe four times. Most of time I use it as headless machine over SSH+Wayland connection. Nice, fast machine for my Linux/AArch64 work.

Still has several not supported things. No Thunderbolt, no microphone etc. I do not blame Asahi team for it — they have done great work. I learnt to live with those because system performance pays for it.

Early adopter? Not anymore

After 20+ years of working with Arm hardware I decided to stop being early adopter. Sure, it is fun to be one of the first users of any platform but I learnt to value my time.

When I pay, I want hardware that works. Let it boot Debian ‘testing’ or Fedora ‘rawhide’ from generic install image. Being able to run *BSD would be a nice bonus showing that platform was well tested.

Hardware which boots after pressing power button. And shut downs completely on poweroff command — including all peripherals and expansion cards.

I do not care much will it use ACPI or DeviceTree to describe hardware to the operating system. As long as it is done in sane way. Source code for firmware is optional when it works properly.

On a new Arm system I would run BSA ACS (or ask someone with hardware). To check did vendor even looked at BSA specification before doing hardware.

Future?

Who knows, maybe one day I will use Arm system for home services.

Instead of yet another x86-64 thin terminal. Which works. Out of the box.

During last years Arm released some specifications in an effort to help organise standardise their market. We got Server Base System Architecture (SBSA), then Base System Architecture (BSA) and finally PC Base System Architecture (PC-BSA) one. Both SBSA and PC-BSA extend BSA (SBSA 7.0 was rebased on top of BSA).

Did these documents changed the market? That’s a discussion for another time.

The checklist

Each of mentioned specifications comes with a compliance checklist referencing document sections such as B_PE_11 which states:

Each PE must implement a minimum of six breakpoints, two of which must be able to match virtual address, contextID or VMID.

To visualize these checklists I created the Arm BSA/PC-BSA/SBSA checklist page where this data is presented as a table. How it is generated will be explained below.

Running ACS

Following the checklist manually is a difficult task so Arm released Architecture Compliance Suites (ACS in short) for BSA and SBSA specifications which run tests and tell whether your hardware is compliant. PC-BSA so far does not have own ACS yet.

NOTE: I ran compliance suites only on UEFI+ACPI systems so do not know how BSA ACS behaves on DeviceTree based ones.

At the start, ACPI tables are parsed and hardware details are checked, such as GIC, SMMU, PCIe etc. You then get a summary of the information:

 Creating Platform Information Tables
 PE_INFO: Number of PE detected       :    4
 GIC INFO: GIC version                :    v3
 GIC_INFO: Number of GICD             :    1
 GIC_INFO: Number of GICR RD          :    1
 GIC_INFO: Number of ITS              :    1
 TIMER_INFO: System Counter frequency :    1000 MHz
 TIMER_INFO: Number of system timers  :    0
 WATCHDOG_INFO: Number of Watchdogs   :    1
 PCIE_INFO: Number of ECAM regions    :    1
 PCIE_INFO: Number of BDFs found      :    4
 PCIE_INFO: Number of RCiEP           :    2
 PCIE_INFO: Number of RCEC            :    0
 PCIE_INFO: Number of EP              :    1
 PCIE_INFO: Number of RP              :    1
 PCIE_INFO: Number of iEP_EP          :    0
 PCIE_INFO: Number of iEP_RP          :    0
 PCIE_INFO: Number of UP of switch    :    0
 PCIE_INFO: Number of DP of switch    :    0
 PCIE_INFO: Number of PCI/PCIe Bridge :    0
 PCIE_INFO: Number of PCIe/PCI Bridge :    0
 SMMU_INFO: Number of SMMU CTRL       :    1
 SMMU_INFO: SMMU index 00 version     :    v3.1
 Peripheral: Num of USB controllers   :    1
 Peripheral: Num of SATA controllers  :    1
 Peripheral: Num of UART controllers  :    1

The format may differ a bit between both BSA ACS and SBSA ACS but data is there.

Then tests run in groups (PE, Memory map, GIC, SMMU, PCIe etc.). Each test can return one of 3 values: PASS, FAIL or SKIPPED (usually when the requirements to run the test are not met):

  28 : Check Fine Grain Trap Support         
       Failed on PE -    0
       S_L7PE_01
       Checkpoint --  1                           : Result:  FAIL
  29 : Check for ECV support                 
       Failed on PE -    0
       S_L7PE_02
       Checkpoint --  1                           : Result:  FAIL
  30 : Check for AMU Support                 
       Failed on PE -    0
       S_L7PE_03
       Checkpoint --  1                           : Result:  FAIL
  31 : Checks ASIMD Int8 matrix multiplc          : Result:  PASS
  32 : Check for BFLOAT16 extension               : Result:  PASS
  33 : Check PAuth2, FPAC & FPACCOMBINE           : Result:  PASS
  34 : Check for SVE Int8 matrix multiplc         : Result:  PASS
  35 : Check for data gathering hint              : Result:  PASS
  36 : Check WFE Fine tune delay feature     
       Recommened WFE fine-tuning delay feature not implemented
       S_L7PE_09
       Checkpoint --  2                           : Result:  SKIPPED

As you can see when test does not pass you get a tag (like S_L7PE_01) pointing to relevant specification.

Verbose mode

Each ACS can be run in verbose mode by adding “-v 1” to the command line. Amount of detail is increased to the level that logging output is highly recommended. You can compare my older logs:

• BSA normal
• BSA verbose
• SBSA level 7 normal
• SBSA level 7 verbose

Back to table stuff

In my sbsa-ref-status repository I have scripts that gather and parse data for QEMU’s SBSA Reference Platform (sbsa-ref in short). The result is set of YAML files (status-bsa.yml and status-sbsa.yml) that contain information how the tests went:

31:
  level: '7'
  status:
    cortex-a57: FAIL
    cortex-a72: FAIL
    max: PASS
    neoverse-n1: FAIL
    neoverse-n2: PASS
    neoverse-v1: PASS
  tags: S_L7PE_04
  title: Checks ASIMD Int8 matrix multiplc

As you see test for S_L7PE_04 passed on some cpu core models and failed on old ones. This pattern continues for other tests and tags. Several entries have only SKIPPED values because the hardware lacked something required to run them.

Those scripts should work with results from other hardware.

Test coverage

Some tags do not have coverage in the ACS. Some tests check for things that are not present in checklists present in the specifications. In these cases it is important to look into ACS documentation:

• BSA ACS testcase checklist
• SBSA ACS testcase checklist

Both of these expand on the checklists from the specifications with additional information. Which SBSA level test was written for, is it tested in the UEFI or Linux environment, is additional exerciser card required etc.

I used both to expand status-(s)bsa.yml files to ensure all tested entries are listed.

Generation of checklist page

To generate a page with checklist table I use another YAML file: xbsa-checklist.yml. This file maps tags from the specifications into groups and subgroups and keeps information on whether tag required for BSA v1/v2, PC-BSA or SBSA levels. I wrote it by hand and it needs to be updated with every specification update.

Next generate-xbsa.py needs to be run which generates HTML page with the table.

Caveats

Changes to ACS may alter the test numbers or tags used. I reported several issues against both BSA ACS and SBSA ACS about it and they were handled. At this moment all tests have tags assigned.

If the generated page lists “ACS only tests” entries it means one of status-*.yml file needs to be updated because some unhandled tag was used. Or ACS change had error in tag name.

Tests may be renamed — in such cases status files will get new ones. When test numbers change (which is rare) then manual checking may be required.

PC-BSA ACS?

According to Arm developers (BSA ACS issue #395) there will be PC-BSA ACS in first week of December. Once it is released, a new parser will need to be written (like one for BSA ACS or for SBSA ACS) and the page generator updated to use this information.

Future?

This page is one of projects I plan to abandon in 2025. It was a useful tool for checking what needed to be done for the SBSA Reference Platform, either on virtual hardware (QEMU) or in firmware (TF-A + EDK2).

None of Arm hardware I use at home is SBSA compliant. Running BSA ACS on some of those causes them to hang. I do not expect this situation to change in the coming months.

The current page will remain online but I do not plan to invest time in updating it.

Again this year, Arm offered to host us for a mini-debconf in Cambridge. Roughly 60 people turned up on 10-13 October to the Arm campus, where they made us really welcome. They even had some Debian-themed treats made to spoil us!

Hacking together

For the first two days, we had a "mini-debcamp" with disparate group of people working on all sorts of things: Arm support, live images, browser stuff, package uploads, etc. And (as is traditional) lots of people doing last-minute work to prepare slides for their talks.

Sessions and talks

Saturday and Sunday were two days devoted to more traditional conference sessions. Our talks covered a typical range of Debian subjects: a DPL "Bits" talk, an update from the Release Team, live images. We also had some wider topics: handling your own data, what to look for in the upcoming Post-Quantum Crypto world, and even me talking about the ups and downs of Secure Boot. Plus a random set of lightning talks too! :-)

Video team awesomeness

Lots of volunteers from the DebConf video team were on hand too (both on-site and remotely!), so our talks were both streamed live and recorded for posterity - see the links from the individual talk pages in the wiki, or http://meetings-archive.debian.net/pub/debian-meetings/2024/MiniDebConf-Cambridge/ for the full set if you'd like to see more.

A great time for all

Again, the mini-conf went well and feedback from attendees was very positive. Thanks to all our helpers, and of course to our sponsor: Arm for providing the venue and infrastructure for the event, and all the food and drink too!

Photo credits: Andy Simpkins, Mark Brown, Jonathan Wiltshire. Thanks!

It's been a while since I've posted about arm64 hardware. The last machine I spent my own money on was a SolidRun Macchiatobin, about 7 years ago. It's a small (mini-ITX) board with a 4-core arm64 SoC (4 * Cortex-A72) on it, along with things like a DIMM socket for memory, lots of networking, 3 SATA disk interfaces.

The Macchiatobin was a nice machine compared to many earlier systems, but it took quite a bit of effort to get it working to my liking. I replaced the on-board U-Boot firmware binary with an EDK2 build, and that helped. After a few iterations we got a new build including graphical output on a PCIe graphics card. Now it worked much more like a "normal" x86 computer.

I still have that machine running at home, and it's been a reasonably reliable little build machine for arm development and testing. It's starting to show its age, though - the onboard USB ports no longer work, and so it's no longer useful for doing things like installation testing. :-/

So...

I was involved in a conversation in the #debian-arm IRC channel a few weeks ago, and diederik suggested the Radxa Rock 5 ITX. It's another mini-ITX board, this time using a Rockchip RK3588 CPU. Things have moved on - the CPU is now an 8-core big.LITTLE config: 4*Cortex A76 and 4*Cortex A55. The board has NVMe on-board, 4*SATA, built-in Mali graphics from the CPU, soldered-on memory. Just about everything you need on an SBC for a small low-power desktop, a NAS or whatever. And for about half the price I paid for the Macchiatobin. I hit "buy" on one of the listed websites. :-)

A few days ago, the new board landed. I picked the version with 24GB of RAM and bought the matching heatsink and fan. I set it up in an existing case borrowed from another old machine and tried the Radxa "Debian" build. All looked OK, but I clearly wasn't going to stay with that. Onwards to running a native Debian setup!

I installed an EDK2 build from https://github.com/edk2-porting/edk2-rk3588 onto the onboard SPI flash, then rebooted with a Debian 12.7 (Bookworm) arm64 installer image on a USB stick. How much trouble could this be?

I was shocked! It Just Worked (TM)

I'm running a standard Debian arm64 system. The graphical installer ran just fine. I installed onto the NVMe, adding an Xfce desktop for some simple tests. Everything Just Worked. After many years of fighting with a range of different arm machines (from simple SBCs to desktops and servers), this was without doubt the most straightforward setup I've ever done. Wow!

It's possible to go and spend a lot of money on an Ampere machine, and I've seen them work well too. But for a hobbyist user (or even a smaller business), the Rock 5 ITX is a lovely option. Total cost to me for the board with shipping fees, import duty, etc. was just over £240. That's great value, and I can wholeheartedly recommend this board!

The two things that are missing compared to the Macchiatobin? This is soldered-on memory (but hey, 24G is plenty for me!) It also doesn't have a PCIe slot, but it has sufficient onboard network, video and storage interfaces that I think it will cover most people's needs.

Where's the catch? It seems these are very popular right now, so it can be difficult to find these machines in stock online.

FTAOD, I should also point out: I bought this machine entirely with my own money, for my own use for development and testing. I've had no contact with the Radxa or Rockchip folks at all here, I'm just so happy with this machine that I've felt the need to shout about it! :-)

Here's some pictures...

It (was) that time of year again - last weekend we hosted a bunch of nice people at our place in Cambridge for the annual Debian UK OMGWTFBBQ!

Lots of friends, lots of good food and drink. Of course lots of geeky discussions about Debian, networking, random computer languages and... screws? And of course some card games to keep us laughing into each night!

Many thanks to a number of awesome friendly people for again sponsoring the important refreshments for the weekend. It's hungry/thirsty work celebrating like this!

• Codethink
• Mythic Beasts