ARM version of Linux Mint

[BUpp]

I would like some information about Linux Mint for ARM.
I know there is no official version of Linux Mint for ARM architecture. I often see on the Mint forums that it's because the ARM architecture has a wide variety of processors. But what should be adapted to the Linux Mint team for each processor.
WARNING!!! : when I say "for each processor", I am not talking about each architecture (ARM64, AMD, etc.) but I am talking about each processor (M1, Cortex A55, Ryzen, Intel Core i9, etc.)

I'm a developer familiar with Linux Mint although I don't know much about ARM processors.
I find it really unfortunate that Linux Mint is not compatible with ARM processors because ARM is becoming an increasingly used processor architecture (notably in Macs M1, M2) and I personally find that Linux Mint is the best open source OS . If the Linux Mint team doesn't have time to do this I volunteer to create the ARM version of Linux Mint.

Thank you for your answers.

[xenopeek]

Your topic isn't about programming so I've moved it here instead. This board is a support forum. For Linux Mint users to help each other with Linux Mint problems and questions. There have been several suggestions for an ARM version of Linux Mint in the past; https://duckduckgo.com/?q=site%3Aforums ... rm+version

Supporting ARM is a huge undertaking. Mostly not about programming but about building, packaging, maintaining and testing all packages and ISOs for ARM (and then which ARM?…). For similar reasons that Linux Mint doesn't have a KDE ISO, a LMDE ISO with MATE or Xfce, or a minimal install ISO I doubt this will happen soon because practically it means dropping another ISO to free up time with the various teams. It's not a one person job if it is to meet Linux Mint's quality process.

Anyway, there is already a request for this here: https://github.com/orgs/linuxmint/discussions/22. Don't make empty "me too" or "+1" comments there, that just wastes developer attention. Instead use the upvote arrow in the request to show your interest in an ARM edition.

[BUpp]

Thanks i understand

I have closed this topic

[duracell80]

x86 will remain the go to, however if we do get more ARM laptops, we might see 15 hour ... 27 hour laptops in the 2030's (and I hope I get to see this come true too). The mini PC's that are out running at less than 35 watts are incredible, if you're not creating content there's no reason an ARM OLED laptop can't work out. Imagine it from how we grew up with a light bulb that would consume 100 watts, to what we can do now? We're gonna need it too as we chug through to the second half of the century.

[Exirion]

Don't underestimate the relevance of ARM support. Even more so in the near future. I've been a Linux user for about 27 years; initially downloaded Slackware on floppies at university. In the past I used Red Hat, SUSE and Debian, but settled on Linux Mint like 15 years ago. Recently however, I switched back to Debian because it has ARM64 support. I can run it on my M1 Macbook Pro with nearly native performance using QEMU as a hypervisor. It's faster than anything I've used before. ARM64 on the desktop is gaining popularity at Intel's expense, so I would not stick with the Intel-only focus.

[duracell80]

It's coming, for laptops, maybe desktops in the mini PC factor for sure. And if there are still battery life issues in this ARM future (which is coming), something is going really wrong.

Cinnamon does run on Arm already so the future maybe now.

[tavvva]

Hopefully my first aarch64 laptop will arrive soon and I decided to rebuild Mint packages for aarch64 from Ubuntu and if they work, I'll share them or create the very first Mint/aarch64 repository. There will be no installer and the installation proces would be to install Ubuntu, then add the extra repo and install the Mint packages...

[tavvva]

Hopefully my first aarch64 laptop will arrive soon and I decided to rebuild Mint packages for aarch64 from Ubuntu and if they work, I'll share them or create the very first Mint/aarch64 repository. There will be no installer and the installation proces would be to install Ubuntu, then add the extra repo and install the Mint packages...

So far, I have rebuilt all the packages required to successfully install the mint-meta-xfce package and the Mint additions somehow work ... of course the updates do not work as I only managed to setup a simple local repository, but at least the visual look is there ... The Firefox package from Mint is missing as it is based on a binary build from upstream and there's no binary package for arm yet, but it's always possible to install the firefox package from the ubuntu repos.

I tested the installation on top of the Mantic and Jammy releases and all the dependencies were pulled by apt without problems.

[Hawaiihemd]

Awesome!
Please keep up the good work.

[bengtfalke]

I run Cinnamon/Debian on Orange Pi 5+ and raspberry Pi 5 on Armbian.

[tavvva]

I believe the biggest help from the Mint team would be to add aarch64 builder and start building all architecture dependent packages also for the arm64 platform once there is a successful build of the package on the x86 platform and then make the arm64 packages available via the official repositories so that everyone who wants to start testing them can easily do so. The arm64 support could be kept experimental with absolutely no guarantees until everything is slowly tweaked/fixed to run correctly on both architectures. The installer is not required in the first phase as there seem to be very few models of hardware with UEFI support (and ability to boot from ISO) available at the moment.

@xenopeek: any chance for the above?

[SMG]

...I'll share them or create the very first Mint/aarch64 repository...

Please see this post. Rebuilding Mint packages means you are creating something new and should be removing Linux Mint branding so people are aware it is not software from Linux Mint and support will be coming from you and not the Linux Mint team.

[tavvva]

...I'll share them or create the very first Mint/aarch64 repository...

Please see this post. Rebuilding Mint packages means you are creating something new and should be removing Linux Mint branding so people are aware it is not software from Linux Mint and support will be coming from you and not the Linux Mint team.

Well, this is very controversial topic ... the linux distributors often patch upstream software and then release it without removing info about upstream to keep the credits where they belong. It's probably a question for legal department what level of modification can still be considered negligible enough so that the result couldn't be called a derived work.
If the Mint team is so afraid of reputational damage, then it's also good to think about reputational damage caused by missing support for aarch64 that is wanted by a group of Mint fans.
I never wanted to create a new distribution ... I only wanted to help with the porting of Mint to aarch64 and also mentioned the aarch64 support could be marked experimental just to avoid any kind of reputational damage. I'm pretty sure it would be absolutely apparent to everyone who would like to try it and some kind of liability disclaimer would also be mentioned in the deployment procedure if it is ever required ...
I believe the above modifications with keeping the Mint artwork could be considered "fair use" according to the law ... I have no profit from the activity and my goal is to help those who are in need.
So far I've deployed Linux Mint to more than a hundred of computers including non-profit sector and schools, doing enlightenment whenever and wherever possible. So, I'm not an enemy

[xenopeek]

I believe the biggest help from the Mint team would be to add aarch64 builder and start building all architecture dependent packages also for the arm64 platform once there is a successful build of the package on the x86 platform and then make the arm64 packages available via the official repositories so that everyone who wants to start testing them can easily do so. The arm64 support could be kept experimental with absolutely no guarantees until everything is slowly tweaked/fixed to run correctly on both architectures. The installer is not required in the first phase as there seem to be very few models of hardware with UEFI support (and ability to boot from ISO) available at the moment.

What's the purpose of building packages if there's no way to install the OS?

I don't know much about ARM. I've had a Raspberry Pi or two in the past. Oh and my first computer 40 years ago was a Acorn Electron, from the company that went on to make ARM. But isn't a problem with ARM there's no platform standard so distros that support ARM end up having to build their packages and ISO for dozens of different platforms? Taking a look at Arch Linux ARM download page for example: https://archlinuxarm.org/about/downloads. Is that solved with Aarch64?

In any case if you want to get involved with Linux Mint development you should talk to the developers. The email address is on the website. You can also post an idea here https://github.com/orgs/linuxmint/discussions and describe what you could do to help.

[tavvva]

What's the purpose of building packages if there's no way to install the OS?

There IS a way. I've tried to explain that in my previous comments. The mint-meta-* packages pull the required Mint dependencies even when they're installed on top of Ubuntu. So, the alternative installation process is to install Ubuntu or Armbian and then migrate it to Mint by adding the Mint repository and installing the meta packages + other optional Mint packages according to the personal needs. I'm already running Mint on aarch64. It's incomplete, but it's already giving me the positive user experience I'm using Mint for.

I don't know much about ARM. I've had a Raspberry Pi or two in the past. Oh and my first computer 40 years ago was a Acorn Electron, from the company that went on to make ARM. But isn't a problem with ARM there's no platform standard so distros that support ARM end up having to build their packages and ISO for dozens of different platforms? Taking a look at Arch Linux ARM download page for example: https://archlinuxarm.org/about/downloads. Is that solved with Aarch64?

Yeah, it was a problem in the past and it had to be solved with uboot+dtb, but the most recent hardware is finally equipped with UEFI and Ubuntu offers a server version (Mantic), that has an ISO file that CAN be installed on these UEFI compliant pieces of hardware and it could be used as a starting point for the future Mint installer. And even when a lot of the HW doesn't support UEFI yet, I bet all the owners will be trying to hack their Armbian images to run Mint stuff on top of them.

In any case if you want to get involved with Linux Mint development you should talk to the developers. The email address is on the website. You can also post an idea here https://github.com/orgs/linuxmint/discussions and describe what you could do to help.

Well, I tried to reach any devels on the IRC channel, but it seems they don't come there. Emails and Github discussions seem to me too offline and I have a feeling it could take ages to discuss the topic there :] Anyways, I've tried to write some comments there too.

Btw. I've created my very first pull request for the mintwelcome application with added support for the aarch64 .... it's waiting for 'clefebvre' to review and eventually merge it.

[SMG]

Well, I tried to reach any devels on the IRC channel, but it seems they don't come there. Emails and Github discussions seem to me too offline and I have a feeling it could take ages to discuss the topic there :] Anyways, I've tried to write some comments there too.

This is a user forum where users help other users. This is not a place where development is determined or done.

You need to discuss your desires with the Linux Mint developers in order to get approval for your creations to be approved as part of the Linux Mint family and all the subsequent maintenance and support such an inclusion would create if you do not want to fork it and make your own distro.

It's probably a question for legal department what level of modification can still be considered negligible enough so that the result couldn't be called a derived work.

You seem to be missing the point that by putting Linux Mint's name on it then people using it will come to this forum (asking for help) and to the Linux Mint development team for support for it (ie. posting bugs) when, so far, it is not an approved Linux Mint project.

Right now, this is your project.

If the Mint team is so afraid of reputational damage, then it's also good to think about reputational damage caused by missing support for aarch64 that is wanted by a group of Mint fans.

Mint fans want a lot of different things and it doesn't cause "reputational damage".

People creating software and putting Mint's name on it when the Linux Mint development team did not have anything to do with it has much more potential for causing reputational damage. That is why that is not permitted.

I'm pretty sure it would be absolutely apparent to everyone who would like to try it...

I doubt that very much based on comments I've seen posted on this forum and across the internet. Your view is well-intentioned, but Mint users and critics encompass a wide variety of people.

I believe the above modifications with keeping the Mint artwork could be considered "fair use" according to the law ...

What law are you using? Irish/EU law?

So far I've deployed Linux Mint to more than a hundred of computers including non-profit sector and schools, doing enlightenment whenever and wherever possible. So, I'm not an enemy

You are someone who appears to be asking for an as yet undetermined current and future time commitment from a very small development team. Are you committing to supporting this on a long term basis or are you just volunteering to do the initial work expecting others to pick up the future work?

[tavvva]

You need to discuss your desires with the Linux Mint developers in order to get approval for your creations to be approved as part of the Linux Mint family and all the subsequent maintenance and support such an inclusion would create if you do not want to fork it and make your own distro.

I already tried it on github but got very few responses so far. Hopefully we'll get somewhere soon

You seem to be missing the point that by putting Linux Mint's name on it then people using it will come to this forum (asking for help) and to the Linux Mint development team for support for it (ie. posting bugs) when, so far, it is not an approved Linux Mint project.
Right now, this is your project.

I hope I'm not missing your point. I believe I just wasn't clear about the things I'm asking for and wanna do. Nobody want's you to announce Linux Mint supports aarch64. I'm just kindly asking whether it would be possible to setup a builder for aarch64 and start building the Mint packages QUIETLY so that the aarch64 enthusiasts could start doing alpha testing and create pull requests for their fixes. That way the Linux Mint would slowly become prepared for the day D, when the leaders decide to introduce and announce the aarch64 support. Prior that all the manuals would clearly state they're unofficial and thus completely unsupported, to prevent people asking for support.
Adding the aarch64 builder would hardly mean a constant increase of load to the development team as I believe it can easily be automated and triggered by successful x86_64 builds.

Mint fans want a lot of different things and it doesn't cause "reputational damage".

How do you know for sure?
Do you have exact stats coming from a survey targeted at the topics?

People creating software and putting Mint's name on it when the Linux Mint development team did not have anything to do with it has much more potential for causing reputational damage. That is why that is not permitted.

Well ... again ... I never wanted to do anything against your will ... otherwise I wouldn't write comments here and instead of that I would search for ways how to do it without you noticing. That is really not my goal.

What law are you using? Irish/EU law?

It has a very similar wording in many countries of the world ... a non-profit research is usually fine.
But like I wrote ... I have no intention to do something against your will. The Mint artwork is one of the reasons why people use Mint and therefore it makes no sense to redistribute the Mint stuff without the artwork.

You are someone who appears to be asking for an as yet undetermined current and future time commitment from a very small development team.

Unless you build all the packages manually, the alpha phase wouldn't take much commitment as it only requires to setup a builder and automate the builds to be triggered by a successful x86 build.

I'm just curious ... why is the team very small?

Are you committing to supporting this on a long term basis or are you just volunteering to do the initial work expecting others to pick up the future work?

It's too early to tell. Unless we get results of the alpha testing there's no chance to guess how much time and effort it would need. That's why I recommend to start with the aarch64 builder and do future decisions based on that research.

[xenopeek]

We've had a discussion about it in the team. Linux Mint isn't opposed to Aarch64 support but as with any addition there has to be a sufficient demand for it. Are there many people wanting to run Linux Mint on Aarch64 or is it super niche? Right now aside from Apple Mx laptops there are what, a handful of Aarch64 laptops? The team can likely add Aarch64 to the package builds but it would increase build times, file transfers, storage etc. If it becomes more common that the average Joe or Jane who'd use Linux Mint can walk out of an electronics store with a new Aarch64 laptop, maybe. No promises though.

With wayland, libadwaita, snap, etc. there are plenty big things for the team to work on this year.

I get that's not what you were hoping for and I'm sorry if it disappoints. Thank you for outlining your idea though.

There are other distros for Macs, Raspberry Pi's and such.

[tavvva]

We've had a discussion about it in the team. Linux Mint isn't opposed to Aarch64 support but as with any addition there has to be a sufficient demand for it. Are there many people wanting to run Linux Mint on Aarch64 or is it super niche?

That's difficult to answer without doing a proper survey, but we could at least do a best guess based on the available statistics:

1.) There are more than 100 million aarch64-based computers sold worldwide (RaspberryPi is on top with nearly 30 milion pcs)
2.) Linux Mint is currently second in the popularity rankings on distrowatch

Right now aside from Apple Mx laptops there are what, a handful of Aarch64 laptops? The team can likely add Aarch64 to the package builds but it would increase build times, file transfers, storage etc. If it becomes more common that the average Joe or Jane who'd use Linux Mint can walk out of an electronics store with a new Aarch64 laptop, maybe. No promises though.

The new Chromebooks come with Snapdragon, MediaTek and Qualcomm CPUs and I think the trend is starting right now.

With wayland, libadwaita, snap, etc. there are plenty big things for the team to work on this year.

There will never be a moment you'd feel idle enough to say "everything is done". Right?

I get that's not what you were hoping for and I'm sorry if it disappoints. Thank you for outlining your idea though.

There are other distros for Macs, Raspberry Pi's and such.

Nah I already have Mint on my Raspberry Pi and it is staying there

http://tavvva.net/files/linux/linux-mint-rpi4.jpg

[Hoser Rob]

...I'll share them or create the very first Mint/aarch64 repository...

Please see this post. Rebuilding Mint packages means you are creating something new and should be removing Linux Mint branding so people are aware it is not software from Linux Mint and support will be coming from you and not the Linux Mint team.

Well, this is very controversial topic ... the linux distributors often patch upstream software and then release it without removing info about upstream to keep the credits where they belong. It's probably a question for legal department what level of modification can still be considered negligible enough so that the result couldn't be called a derived work.
If the Mint team is so afraid of reputational damage, then it's also good to think about reputational damage caused by missing support for aarch64 that is wanted by a group of Mint fans....

If you do what you're describing (and you would indeed wuite obviously be the one responsible for supporting it), the only one in danger of reputational damage would be you.