Remember when Qualcomm promised Linux would be a first-tier platform alongside Windows for its Snapdragon X Elite, almost a year ago now? Well, the Snapdragon X laptop have been out in the market for a while running Windows, but Linux support is still a complete crapshoot, despite the lofty promises by Qualcomm. Tuxedo, a European Linux OEM who promised to ship a Snapdragon X laptop running Linux, has posted an update on its progress, and it’s not looking good.
While Tuxedo did reach a major milestone last week by sending the laptop’s device tree to the LKML, that’s where the good news ends.
The next step is to support additional components of the ARM notebook within the device tree. This includes all USB functionalities, including USB4, external monitor connectivity via HDMI, and audio features, such as the headset jack. Additionally, driver testing is on the agenda. Unfortunately, a planned collaboration with Qualcomm, the manufacturer of the Snapdragon X Elite, did not materialize. However, we are in contact with the ARM specialists at Linaro and have sent test devices to them. We hope to receive valuable feedback from their developers and the community in the near future.
↫ Tuxedo’s website
This seems to indicate that Qualcomm isn’t as interested in Linux support after all, which may be because the Snapdragon X machines haven’t exactly taken over the laptop market as Microsoft and Qualcomm had hoped. The market for these things is probably not large enough for Qualcomm to justify investing in Linux support, especially when Windows on ARM is apparently not up to snuff yet either.
In case you are unaware of why device trees are such a big thing in ARM land, it’s because ARM devices do not have a nice ACPI table for operating systems to read system information from. Whereas x86 devices have their hardware components laid out in a nice ACPI table in UEFI, ARM devices do not, meaning that the Linux kernel needs to know specifically which device you’re using so it can load the correct device tree. On x86, this isn’t necessary, as the Linux kernel can just read the ACPI table, which works 99% of the time to get it to boot, even if specific components might not be supported (yet). On ARM, without a device tree, the Linux kernel doesn’t know what to do.
That’s one of the major reasons why it’s so hard for ARM to take off in the same way x86 once did. It’s just not designed to be infinitely intercompatible and interoperable as we’ve come to expect from the x86 world, and I don’t think anybody has any vested interest in changing that. I had hoped Microsoft might throw its weight around here, but it seems that’s not happening either.
The ARM desktop/laptop revolution seems mostly confined to Apple for now.
Not sure what the confusion is about. Qualcom lied. Did anyone whose followed their history with Linux really expect anything else? I had a lot of interest in getting one of these when the Linux support appeared (if it ever did) but I held off on actually purchasing one until then. Wouldn’t be shocked to hear that many others also held themselves back.
bornagainenguin,
We need better linux support on ARM. These problems make ARM too damn difficult to work with compared to x86. I really was hoping for the best after Qualcom’s announcement – a linux system for ARM that works as well as it does on x86 without hacking.
>”a linux system for ARM that works as well as it does on x86 without hacking”
What could possibly lead you to hope for that? History (and landfills) are littered with the dead skeletons of all the different cheap ARM systems that have come and gone, including from the Apple faithful who have slavishly (and predictably) upgraded from M1 through M4. Meanwhile, refurbished 15-year-old corporate fleet Thinkpads are still best sellers. Why should any of that be Linux’s problem?
andyprough,
Qualcomm aren’t a fly by night operation. They are a significant player and they are around for the long term. They could genuinely make the difference and it’s worth hearing them out but it comes down to whether they follow through.
It’s not clear to me who you are talking about when you say “linux”. Maybe it’s not a problem for users who throw their hands up and keep buying new hardware, but for those who want to independently maintain and update ARM hardware long term this is absolutely a problem. I’d go even further and say this cycle of unsupported hardware getting thrown away is a big problem for the planet as a whole. Every year that we don’t solve this is a missed opportunity IMHO.
The Linux kernel is not “the planet as a whole”, and it should not try to fix the planet’s problems.
It seems as though every company who has created abandoned tech products the past 20 years has tried to shove off long-term support onto, “well, those Linux developers will figure out something”. But that work doesn’t get done in a vacuum – kernel support is normally paid for by the corporation who makes the product. In this case, this should be entirely Qualcomm’s problem. If Qualcomm wants the Linux kernel to support their destined-for-the-landfill-laptops, then Qualcomm needs to pay engineers to write driver code and submit it for uptake into the kernel.
andyprough,
We’re in disagreement then. A very significant amount of ewaste comes from linux devices that consumers can’t reuse without manufacturer support. We don’t have this problem with x86 because with that architecture we are much less dependent on the manufacturers. On ARM, we’re dependent on the manufacturers for software, sadly.
I also blame the manufacturers. To be fair there is plenty of blame to go around. but we’ve been playing this blame game for decades and we can all see that passing the buck is a guaranteed non-solution. The linux community doesn’t get a moral exemption when it fails to provide adequate after market support. We have been stalling for decades now I say it’s about damn time we in the linux community get the ball rolling and finally help users replace their devices on their own terms instead of leaving it to the manufactures which doesn’t work!!! I concede that many people just don’t give a damn about ecological issues – it’s a “someone else problem” – but we all pay the cost for this attitude.
No. LOL
Qualcomm is a crappy company, and I guess, the general public confused Android support with real Linux support. LOL
I’ve gotten close to purchasing an Arm Thinkpad, but that says more about my confidence in Lenovo then in Qualcomm. I’m not sold on these little vendors and Qualcomm chips. Weirdly, I’d be more confident in Mediatek based laptops.
Do you want to say that Qualcomm X Elite work with Android? That’s new to me.
Thom Holwerda,
Apple is a bigger target, which leads to more attention, but I don’t think apple hardware is any better at standards. It’s the same proprietary crap everyone else is pushing. Hypothetically apple could introduce interoperability standards and encourage other manufacturers to join them. This could be fantastic for users & FOSS. But this is not in apple’s DNA. Interoperability is a non-goal for them.
I think you are misreading the sentence: it’s not about the fact that nobody wants to support Linux on Qualcomm X Elite, but about the fact that laptops with Qualcomm X Elite are not selling – not even with Windows.
I wonder what would happen when Android finally arrive on laptops… probably the end of the next year.
Would people use them or would they stay with x86 even then?
zde,
Selling more == bigger target. My point is that the most popular devices will get more attention. This is consistent with your point as well. Technically we should solve this with industry standards such that everything is compatible regardless of market size. Alas this outcome has been elusive with ARM and we may never get there
Sorry Thom, that simply not true. ARM has worked with the industry and created a program called ARM SystemReady, which
https://www.arm.com/architecture/system-architectures/systemready-compliance-program/systemready-band
There’s no lack of standard, there’s just lack of willingness from vendors like Qualcomm to do the things right, unlike ppl like Ampere, that create great workstation where you can install whichever OS you like:
https://www.ipi.wiki/products/ampere-altra-developer-platform?variant=41995037868194
The server space has great support and works like x86, from what I’ve heard.
I believe I called that on Microsoft. They went a long with Samsung trying to get in to this space, but they don’t have the right incentive here. They were always going to do the minimum, and nothing more. Microsoft doesn’t even seem particularly interested in Windows at this point. It’s all Azure and AI, and trying to find other forms of cloud rent. Windows just doesn’t get them that, and their switch to Chromium within Edge tells you everything you need to know about their current belief in holding the platform. They just aren’t interested if they can’t extract rent.
CaptainN-,
I agree about MS trying to rope people into subscriptions. However I wouldn’t say windows is unimportant to MS, it remains important but the business around it is changing. “Cloud” services is the big one. Windows is already being sold as a cloud service, It’s kind of dumb if you ask me, but I already know IT departments that are deploying it. And microsoft is trying to capitalize on windows users in other ways too including ads and data. There is the microsoft app store. They are forcing users to have MS accounts. Most customers don’t really want these changes, but microsoft are playing the long game and as long as the changes are gradual people tend to become accustomed and then accept it as the new normal.
I think it’s clear that microsoft aren’t trying to make windows better for users anymore; nowadays we’re more likely to see “upgrades” designed to squeeze consumers harder,.
There’s not a single cloud service that requires a fleet of Windows devices. That’s why they aren’t interested in it – that’s why they don’t even have a dedicated team on Windows any more. They have a hardware team (Surface) just to kind of keep the appearance of a classic capital oriented company, and maybe as a hedge – but that’s not their play. They don’t need Windows for basically any of the things they actually make rent on (cloud capital) – and I’d argue, they could save a lot by using community commodity software platforms. In fact, in Azure – they do.
CaptainN-,
Saying that cloud services can be used by other platforms doesn’t mean that microsoft aren’t interested in windows anymore. I’m not a fan of what they are doing, but it’s pretty clear to me that microsoft are still interested in exploiting the windows user base. If anything I think windows would be better for users if microsoft just let windows be without trying turn it into a service.
ARM is just not and never will be an “open” or Linux-friendly ISA. We should be glad that x86 hasn’t been dethroned yet which means the availability of hardware for Linux (and even more niche systems like BSD, Illumos, OpenVMS, Haiku, ArcaOS, AROS, …) will remain secure in the future. The duopoly of AMD and Intel doesn’t really matter unless you want to design your own hardware.
There are some actual open ISAs such as RISC-V, POWER and SPARC but they are all quite irrelevant and I’m not sure if they actual solve the problem of “one generic OS image will run on any of these CPUs”.
The market for linux/arm is actually probably larger than windows/arm… Virtually all software is available natively as opposed to running under emulation, there are a lot more drivers for third party peripherals and at least some level of demand from people who are managing or developing for arm based cloud servers.
Running windows on arm is a poor experience, most applications are still x86 only and require emulation which degrades performance and counteracts some of the battery life benefits. Not all applications work under emulation – some fail entirely, or have strange bugs that aren’t present when running natively. Most windows developers don’t bother to target arm at all.
For apple the transition is pretty much done, and because the future of macos is arm only developers were not able to ignore it and just continue pumping out x86 versions.
“Unfortunately, a planned collaboration with Qualcomm, the manufacturer of the Snapdragon X Elite, did not materialize.” This sounds like a death sentence to me. Despite the huge effort QCom made to enable the SoC for Linux, it always gets derailed by secrecy (or something).
I am going to sim against the tide here a little bit.
This is what Ubuntu had to say about X Elite laptops on Ubuntu 24.10 back in November:
Devices we have seen successful installations on:
Acer Swift 14 AI SF14-11 660
ASUS Vivobook S 15 2.9k
Dell XPS 13 9345 (32GB and 64GB) 2.4k
HP Omnibook X 14 945
Lenovo ThinkPad T14s Gen 6 (32Gb and 64GB) 3.6k
Lenovo Yoga Slim 7x 3.4k
Microsoft Surface Laptop 7 2.7k
Obviously Ubuntu 25.04 will be better..
The reason that this list is so long is because Qualcomm really did put the work in, getting X Elite support into the mainline kernel starting in kernel 6.8 and up. Compare this to Apple Silicon where, despite amazing work, Ashahi Linux only support devices from 2 generations ago.
And as for Tuxedo, I did not get from their statement that it does not look good. They started late but I did not see any show-stoppers. They are not going to have to reverse-engineer the GPU or secure boot or anything to do with the SoC really. Why would I not expect them to succeed especially as they are Linux first.
And passing customers off to partners is standard procedure for Qualcomm when working with all but the largest OEMs. So no alarm bells there.
Not that I am happy about the situation. Qualcomm’s interest does seem to have dropped for sure. But is that their fault?
I think it is more a Microsoft problem as Windows on ARM has still failed to get traction. It is not because the hardware is lacking really. Or, at least, that is not my view.
The real problem here is that Desktop Linux itself cannot make a platofrm successful. As things stand today, it is really only viable to run Desktop Linux on hardware that was designed for and made commercially successful a different desktop operating system. Asahi Linux is a viable project because Apple Silicon Macs sell like crazy. Amazing hardware is available at good prices and there are sure to be many more generations on it. On the x86-64 side, the same is true because of Windows.
Windows needs to be successful on X Elite first. As it stands, you can run Linux on this hardware. The core SoC support is in the Linux kernel. Ubuntu and Fedora have done a great job investing in the platform. But every new device lags behind on even getting a device tree supported. Becauase nobody is buying them. And Qualcomm see this.
I only hope we see future generations of X Elite chips. If not, I guess Linux will need to skip straight to RISC-V.. That is, once the hardware suppliers get it there.
Hmm, well, on those Snapdragon X devices. Linaro is AFAIK one of the core developers of the driver stack for the Snapdragon X architecture. So Tuxedo might “just” have found the right person to talk to on their intentions, which I find a bit odd. Qualcomm is very restrictive when it comes to information about their SoCs and in-house linux-y software stack. They usually have complete support of the hardware they bake into the SoCs (naturally), Since this is a $billions operation this is not very surprising, it is core of their business. The laptop-class chips are a new-ish LOB with also large customers, but for laptops its actually only one: Microsoft. They have a complete driver stack (from Linaro?) for their OS, and they deviate from the architecture that Qualcomm is using for their firmware. The Windows hypervisor (running on EL2) replaces the HYP firmware which is normally there and should function as a type-1 hypervisor. This requires quite extensive driver work (which the Windows stack has) to make everything dependent on it work again. Also, the “ACPI” is sort of shallow, most of the implementation to make it work on Windows is in the PEP driver. And all of these things are security by obscurity, secret information. Almost nothing of this is known.
To run Linux on these SoCs there is the platform support driver stack, and a (selective) dtsi with the nodes somebody deems necessary to run the SoC on Linux. It is not complete either. USB4 / Thunderbolt is one of the parts that is not supported. Or, for the Snapdragon X and X Plus (x1p4*), the GPU node. What is published is mainly dependent on what their reference designs need to work, and what the contracts say. Big-money concerns.
Now to the supported laptops you listed: Lenovo is a partner and made a token effort in supporting Linux, all the others didn’t (or at least not publicly). That there are so many different models supported is mainly due to fact that they are quite similar to the reference designs, making their device trees a good starting point for individual ones, and that capable Linux devs got their hands on the hardware. I speculated on getting a Snapdragon Dev Kit for Windows, and bought the HP Omnibook X14 on ebay to have something x1e to work on when the Dev Kit didn’t ship. So I tried with it and got mostly lucky in retrospect. Omnibook 14 has fairly extensive support (of what is known) by adding tidbits I find here and there. Same with the ThinkBook 16 (x1p42100 device) – conscious decision to buy one when Purwa support was only rumours, lots of work after the first bits materialized. Now quite good support, except for GPU (nothing from Linaro yet). This led to supporting the Asus Vivobook (x1p42100) too, and will expand to probably all the others popping up, since the Purwa SoC is drop-in compatible with the Hamoa SoC.
That’s why I’m talking about secrecy. It hampers all of this a great deal, and the “base” support from Qualcomm means nothing (from an end-user POV who wants to use a distro) when the vendors itself don’t ship device trees. Which they don’t.