Forgot your password?
Close
typodupeerror
Linux Hardware

Qualcomm Announces Snapdragon X2 Series Processors Will Support Linux (qualcomm.com) 74

Posted by EditorDavid from the in-the-chips dept.
Snapdragon X2 Series processors feature a neural processing unit (NPU) delivering 80 trillion operations per second and allowing advanced AI features to run locally. And Snapdragon X2 Series "is expanding to Linux," Qualcomm announced today, calling it one of their most-requested capabilities: Qualcomm Technologies is a top contributor to Linux development at a kernel level, and now we're embracing support for Snapdragon X2 Series as a platform directly. We're upstreaming core drivers for Snapdragon X2 Series — including the Hexagon NPU and Adreno GPU — to open the door to developers and partners. We are starting with support for two operating systems based on Linux...

- Debian: We're kicking off with Debian by the end of this year, one of the most influential Linux distributions and the foundation behind many of the distros people use every day.
- Ubuntu: Qualcomm Technologies has partnered with Canonical to bring Ubuntu, the world's most widely used Linux distro, to the platform with Snapdragon X2 Series certification targeted for the first half of 2027.

...and this is just the beginning. Our partners HP, ASUS and HUMAIN are planning Linux support in the first of 2027, so that their devices deliver the incredible experience users expect with Snapdragon X2 Series in a new operating system.
Qualcomm's developer blog called it "a significant step forward" in Qualcomm's commitment to a developer-first approach, and "to the open-source community."

"For developers, the important part is simple: more core hardware features are being reviewed and merged, so that laptops with Linux on Snapdragon X2 Series will be easier to build, test, and debug..." The enablement work completed so far has been validated on a Debian 13-based ("Trixie") user space and a custom kernel, so developers should treat this as the current reference environment while support continues to mature... Linux support for laptops with Snapdragon X2 Series is rolling out in stages, starting with the core pieces developers need before they can do production-level work on their device... For developers who want to try it today, the call to action is straightforward: start with Snapdragon X2 Series hardware, review the latest recipes to build Qualcomm Linux Debian Images, build available upstream sources the Debian OS image and test the peripherals that matter to you, such as graphics, AI inference, device I/O, or basic application bring-up. Early testing helps identify the gaps that matter most before support becomes broader and ready for production-level workloads...

You do not have to wait for everything to be fully finished before getting hands-on. If you are comfortable working from upstream sources you can start evaluating Snapdragon X2 hardware today... Check out the step-by-step instructions, including the full build flow and deployment guidance for Snapdragon X2 Series Linux software.

Put simply, this effort is less about supporting specific Linux distributions and more about empowering the developers who make Linux available on new hardware. This Developer Preview targets distribution maintainers, toolchain developers, kernel contributors, and hardware enablement engineers. It provides the upstream building blocks, including kernel patches, drivers, and reference device trees, needed to enable Snapdragon X2 Series support in their own projects and distributions. For end users looking for a turnkey "install and go" experience, that will come later as distributions adopt what lands upstream. We're encouraging the community to build on this work and help shape what comes next.
The blog post notes that in the initial enablement stage, "Qualcomm Linux supports systemd-boot as the Universal extensible firmware interface (UEFI) boot manager to load and boot the Linux kernel."

Hands-on demos were given at the Snapdragon Summit in Maui of Linux running on Snapdragon X2 Series hardware, which the developer's blog calls "a practical look at what works today and where Snapdragon X2 Series Linux support is headed."

Qualcomm Announces Snapdragon X2 Series Processors Will Support Linux More | Reply

Qualcomm Announces Snapdragon X2 Series Processors Will Support Linux

Comments Filter:

  • no way does this mean, no it cant be (Score:5, Funny)

    by ZERO1ZERO ( 948669 ) on Thursday September 24, 2026 @06:05AM (#66349470)
    â" 2027 will be the Year Of The Linux Desktop!

    • 2027 will be the Year Of The Linux Laptop!

      FTFY

      • It's some Greek mythology level curse that Linux made it to everywhere except the Desktop.

    • Re: (Score:1)

      by CEC-P ( 10248912 )
      After Chromebooks absolutely kicked MS in the nuts in about 2 years and Steam made Linux work on a popular mobile compute platform, I'm inclined to think it may actually be for once.

    • â" 2027 will be the Year Of The Linux Desktop!

      (GenMe) The fuck, is a 'desktop'..

    • That's driving up RAM prices for the next 10 years then yeah we could have the year of the Linux desktop. Not from Qualcomm but from steam.

      Microsoft completely shit the bed on Windows 11 and steamos can run a hell of a lot of games with it only being held back by anti-cheat software that could easily be ported over and has been by some companies.

      That would bring millions and millions of people to Linux for the better performance on 16 GB machines.

      The trouble is video cards. Valve made the steam

      • Re:Honestly if we could break up the Monopoly (Score:4, Interesting)

        by sabbede ( 2678435 ) on Thursday September 24, 2026 @01:11PM (#66349863)
        There are new players entering the market, and new capacity is being built by the big 3. Who can't try and drive anyone out with a price cut until they've cleared the backorders and produce batches that aren't already sold.

        There are a couple of hurdles I wasn't aware of. Like Japan restricting etching gas exports and EU environmental policy that's forcing fabs there to retool.

        https://markwideresearch.com/new-ram-market

        • This trend is cyclical, and has been since the 90s. Glut in the RAM market leads to price erosion, which lead to a slowdown in manufacturing. In the meantime, those low prices tend to spur new designs that use plenty of it, spiking the demand, then creating a shortage which leads to a ramp up in manufacturing, which then leads to another glut

          Unless and until the AI bubble bursts, don't expect RAM prices to come down anytime soon. Also, by now, it's well known that this is a volatile market, making it u

      • 16GB machines? What archaeological dig did you find that thing in?

        The RAM shortages are courtesy of our addiction to AIs.

      • That's driving up RAM prices

        There is a brand new entrant into the industry and they are selling RAM at a higher price than Samsung per unit. So clearly the only thing driving up RAM prices is the shortage of RAM and the insane backlog in orders.

        Also there is no monopoly. One company doesn't control the market. The word you're looking for is oligopoly.

        Microsoft completely shit the bed on Windows 11

        No one outside of Slashdot cares. Windows 11 has a large market share. SteamOS while great is still a pittance of the actual market. The idea that millions of people are going to switch t

    • Based on online reviews, and the glowing endorsement [youtube.com] from Linus at LTT, if it wasn't for the AI driven RAMpocalypse I'd have already purchased the new Steam Machine.

      I don't need the bestest and fastest rig because I've still got a massive backlog of titles from the seemingly endless Steam, GOG, and HumbleBundle sales left to play.

      What I do need is a small, quiet, daily driver that can handle my media library, take care of my emails, and slice the various .STL files for my FDM and Resin 3D printers. A daily

  • Yeah, we'll see (Score:3)

    by stevenm86 ( 780116 ) on Thursday September 24, 2026 @06:48AM (#66349489)
    We already tried upstream-first development, back in 2010. It was kind of a disaster. The downstream drivers are much more capable than the versions that had been upstreamed, due to the upstream kernel lacking support for the degree of integration between subsystems that Qualcomm's design required. The upstream drivers were functional but compromised in a few ways. When it next came time to do a kernel upgrade, the downstream drivers conflicted with their upstream counterparts. Unofficially, having been in charge of at least one kernel upgrade was sort of a prerequisite for becoming a sub-team lead, and this is partly why. Maybe this time things will be different, between device tree, PSCI, and generally better maturity of ARM support. But, we'll see

  • The blog post notes that in the initial enablement stage, "Qualcomm Linux supports systemd-boot as the Universal extensible firmware interface (UEFI) boot manager to load and boot the Linux kernel."

    When we said we wanted ARM platforms to support a standard for booting, we didn't mean like that.

    • It'll be interesting to see if it gets more enthusiastic support from Qualcomm than we saw from the various Chinese arm variants that took YEARS to get anything more than the basics working without a lot of mystery kernel blobs on an older kernel. It'll also need to arrive at a sane price point.

      I've had an Orange Pi 5 plus for almost 3 years and it only JUST now works without a lot of hand holding on current Debian/Ubuntu and I think the npu is still not quite there yet.

      • I'm always a little curious what the thinking/strategy is with the Chinese outfits. I don't necessarily expect much from the OrangePi guys; they ship a surprising number of different SoCs, including multiple architectures and multiple vendors, for their size, so are likely to be mostly at the mercy of the SoC vendors; but is the situation for someone like Allwinner such that people who want linux to work on their stuff could be a resource for them and they are just not taking advantage of that very well; or

        • It can only be the latter, because if it was the former then the SoC vendors wouldn't have to be dragged kicking and screaming to mainlining. As far as I can tell it's universal.

          • It's unfortunate. The Orange Pi 5 Plus on paper is nice. When I bought all the bits for about US$300, that got me 8 fairly fast RK3588 ARM cores, 16GB RAM, dual 2.5GB ethernet, decent video, decent npu (for then), native nvme, 256GB emmc storage, 1TB nvme ssd and a small case with active cooling. It would be a LOT more expensive to roll that together today, but it's working ok as a headless home media server. I suspect it would also be fine for a desktop device for many people (no gaming, but it plays 4

            • For the same money I got a AMD Zen3 MiniPC with 8C16T, 32GB RAM, and 1TB nVME. An equivalent machine is about $600 now, whee! It also has dual 2.5GbE and WiFi 6. Same no serious gaming (it has a fairly pathetic number of GPU cores) but it's really quite passable... and no ARM problems. Power consumption is a little higher, but not much, TDP of the processor is only 15W without overclocking.

    • Re: (Score:1)

      by wed128 ( 722152 )

      When we said we wanted ARM platforms to support a standard for booting, we didn't mean like that.

      I think the real headline here is "we boot with UEFI" and not the specific EFI binary they chose. If systemd-boot runs without a ton of patches, it's likely that grub-efi and the EFI kernel stub will run too. As far as "arm supports a standard for booting", this is good news.

      • That's the sort of curious thing about this announcement: at least in theory(I assume things are uglier in practice) the X1 parts that they shipped for windows-on-ARM/Copilot PCs already use UEFI/ACPI in line with what MS wants for booting Windows.

        That obviously doesn't mean that there's a working GPU driver; or that naive power management doesn't cause the system to run at 90% of maximum power and hammer the battery flat; but the theory was that the last gen was already doing the "big kid ARM that has e

    • Re:oh good (Score:4)

      by ArchieBunker ( 132337 ) on Thursday September 24, 2026 @09:23AM (#66349575)

      What always annoyed me about these little ARM and RISCV boards is if they can't boot for some reason you get nothing but a blank screen or dead terminal. Come on we're not limited to kilobytes of ROM space here. Add some debug info or even a heartbeat indicator. Give me some signs of life.

    • Why do some people hate on systemd?

      • Re:AI and systemd? How could I be less interested? (Score:4, Interesting)

        by buzz_mccool ( 549976 ) on Thursday September 24, 2026 @02:15PM (#66349917)

        > Why do some people hate on systemd?

        Perhaps because major bugs don't get fixed?

        https://github.com/systemd/systemd/issues/2913 [github.com]

        • Okay, but isn't that also fairly common?

          And did the kernel change noted in there ever get made?

      • Re: AI and systemd? How could I be less interested (Score:4, Interesting)

        by fluffernutter ( 1411889 ) on Thursday September 24, 2026 @03:54PM (#66349977)
        i don't like systemd because it makes very simple things, like running a process on a timer, very complicated. I'm sure it is very powerful but for the average admin who creates one timed process per year it is very complicated. also why the hell did they make the default command output put through a pager that cuts lines off on the right of the screen? that's almost never what i want. clearly the person who did that doesn't come from a terminal command line background.
        • That's the most helpful response I've had, thank you.

          Unless you're talking about cron jobs, I don't think I've had cause to do that myself. I'm trying to remember if I've seen the no-wrapping issue you mentioned, but most of the terminal interfaces I use handle wrapping. Or I resize the window. Obviously. that would be a pain, but if I'm in a terminal I can't resize I probably have bigger things to worry about. Like a dying drive, or I'm in a frustrating-for-other-reasons HyperV remote desktop window

          • well i frequently copy and paste large sections of text from/to a terminal scroll area regularly because copy paste is way faster than making a file and transferring it. So say you display 300 lines of log, you can't cut and paste because the ends of the lines are off the screen. Most Linux logs are just in regular text files because Linux commands are designed to pipe stdout and manipulate it however you want. like there are hundreds of not thousands of commands to do that. that's the way Linux works.
            • I do that too, I'm just trying to think of a case where cating a log file didn't just dump the entire thing to the screen, word-wrapped and non-stop. The only time I've had any trouble like is when I've had a linux VM running in hyper-v and I'm connected directly instead of through ssh. In which case copy and paste doesn't work right at all for anything.

      • Re: (Score:2)

        by gweihir ( 88907 )

        Why do assholes like you stipulate an emotional stance?

        Systemd is simply bad software, so it has no place on my systems.

        • I will gladly admit to having emotions about systemd.

          They are based in the fact that it's low quality software, but they are about it being pushed on me from every direction.

          In a sane world, a new large project from the creator of pulseaudio would be laughed out of the room immediately, because pulseaudio has always been and remains shit.

          • If it's bad, why do distros keep adopting it?

            This is a sincere question. I don't understand the controversy.

            • I didn't look into any of them at all except Debian, where it was really weird. They rushed a decision without consulting the usual breadth of stakeholders. In their case the excuse was GNOME 3 but I'm not aware that anyone was especially clamoring for that.

            • Re: (Score:2)

              by gweihir ( 88907 )

              Incompetence. There has, unfortunately, been an influx of low-skill, low-insight Windows people into Linux for a while. For the Debian decision. the tech board had been compromised by Red Hat people (former, mostly), that made a political decision instead of a technological one.

              • What made it a political decision instead of a pragmatic one? It's easy to blame incompetence and politics when a decision doesn't go your way, but that doesn't mean it's the case.

          • Re: (Score:2)

            by gweihir ( 88907 )

            I do not have an emotional stance about systemd. It is simply one more piece of bad software made by an incompetent team. There is a ton of that around. I do have an emotional stance about the people that did run a marketing campaign based on dishonesty and emotional manipulation for systemd. These people and their "useful idiot" type enablers are scum and have no place anywhere near quality technology.

        • Forgive me my colloquialism.
          What's bad about it? What's good about it (there must be good, otherwise it wouldn't be so widely adopted)?
          I don't have a dog in this fight (forgive me my idiom), and don't really know what the fight is about.

          • Re: (Score:2)

            by gweihir ( 88907 )

            Do some minimal research. You do know how to use a search engine?

            • Will a search engine provide me with your opinion on the matter? I'm interested in the thoughts of the people talking about it here, not what some unconnected person put in a blog post ten years ago.
          • It does too many things. The latest being age-verification, in response to laws by some US states and Brazil. Too much for an init system

            • Re: (Score:2)

              by gweihir ( 88907 )

              Yes. A crass violation of KISS and the philosophy that makes Unix superior: https://en.wikipedia.org/wiki/... [wikipedia.org]

              • Does "have nothing monolithic" make any more sense than, "make everything monolithic"? What if extra complexity and combining functions makes more sense than keeping everything separate for a particular situation?
            • I'm surprised it took so long before someone brought that up. "Violates the 'do one thing and do it well' philosophy", was what I expected to hear most (though it was the only argument I knew of).

  • So annoying when companies just jump to supporting Ubuntu and often block tools that would work fine on Debian (looking at NVIDIA and its AI related tooling especially). Every time I am forced to use Ubuntu they've made it worse than the last time. I just love the scary "you might not be protected from threats without a subscription" advertising when running apt update.

  • Why did choose to support Linux on the new processor only?
    • SOP at Qualcomm. It has been literal years of hammering on them by Google and basically every Android handset OEM to even keep parts that ship with android support in support for more than a few years(I think the current timeline, at least for 'snapdragon 8 elite' flagship parts, is a princely 8 years from release; and that is with the 'project treble' stuff that attempts to reduce the number of changes that might potentially need to be supported). Especially with the, um, raging success, of Copilot+ AI PCs

      • How many Arm based Windows laptops have sold from anybody? Any Wintel user who bought them could only have lived w/ them if s/he had no legacy Wintel-only software that needed to run on it. But if they needed software that was not cloud-only and whose ISOs they had in Wintel, they'd have been SOL on it here

  • Since I live In Venezuela, Qualcomm does not want me to read the blog post.
    So, I have a question for those who have read it, as TFS does not say:

    The Qualcomm integration work is fully FOSS drivers? Or is a bunch of closed source drivers and binary BLOBs.

    I do not care one way or the other, as I am more pragmatic than ideological in that regard. But I do know that with Closed source Linux drivers:
    1.) In the tail end of the life of the hardware, support becomes spotty, as newer kernels stop supporting the clos

  • Linux will support X2!

    • In that case, Linux could have supported the first generation of Snapdragon based Windows laptops. It doesn't, since the firmware is very different from the UEFI firmware on Wintel laptops

  • Yep, https://www.qualcomm.com/devel... [qualcomm.com]

    Fool me once, shame on...shame on you. Fool me ... you can't get fooled again...

  • Why is this new? (Score:3)

    by schweini ( 607711 ) on Thursday September 24, 2026 @03:03PM (#66349949)
    Why is this something new? Isn't it just an ARM chip, and Linux has been running on ARM for ages? And I think Android still uses the Linux kernel, no?

    • Why is this something new? Isn't it just an ARM chip, and Linux has been running on ARM for ages? And I think Android still uses the Linux kernel, no?

      Is all the other bibs and bobs that hang of the chip. Mem controller, GPU, NPU, USB controller, networking, wireless, the works

    • I don't fully grok it, but basically: x86 has ACPI and other mechanisms. They're abstraction layers that tell the OS kernel how to interact with low-level hardware without the kernel having to have specific files/support for every chipset and processor.

      In ARM, there isn't that level of abstraction, which is why it feels like every ARM variant needs specific support built-in to the kernel for an OS to boot on them.

      In ARM world, someone has to put in the work to define "device trees" for each individual SoC,

  • Wasted operations (Score:3)

    by Khyber ( 864651 ) <techkitsune@gmail.com> on Thursday September 24, 2026 @03:46PM (#66349971) Homepage Journal

    80 trillion operations per second wasted on AI garbage when that could have gone to x86 emulation and made a smartphone on the level of desktop systems.

    Instead they choose to spy on you and steal your information.

    • First these ARE already running desktop Windows, I don't think any are sold with something else.

      Second, here are in question all the low level bits and bobs specific to this platform, you can't meaningfully run the low level drivers, stuff related to sleep and wakeup and so on in emulation. Even if you could even more you're lacking the code to handle Qualcomm hardware on the x86 side so it wouldn't help at all.

    • Why bother about x86 emulation, when iCore Ultra and Ryzen is gonna beat you all day long?

  • Be careful with Qualcomm (Score:3)

    by hwstar ( 35834 ) on Thursday September 24, 2026 @04:00PM (#66349989)

    Given what they did to Arduino, I view them more like the "Mr. Little" advertising character from Western Exterminator in California where the open source community is the rat with the knife and fork, and Mr. Little is the guy trying to have a heart to heart talk but has a sledgehammer behind his back.

    https://www.westernexterminator.com/blog/company/do-you-know-mr-little

    Hopefully this time might be different, and maybe Qualcomm learned something from the Arduino fiasco, but I'd be wary in the meantime.

  • The RAM monopoly isn't the scariest part (Score:4)

    by hwstar ( 35834 ) on Thursday September 24, 2026 @04:14PM (#66350017)

    Its the fact that consumers won't be able to afford embedded devices which use DRAM and SSD. There's a design lifetime of 3-7 years on most electronics items which use embedded processors. Once these devices fail, and consumers look at the replacement cost, they might not buy a replacement because it costs too much, and may be forced to go without. Think Phones, Tablets, Routers, Ethernet Switches, and Personal Computers.

    This RAM/SSD shortage needs to end soon or we could be in real trouble.

      If it continues, we might see "Unconnected Masses" in this case where a large segment of society can no longer afford to remain connected to the Internet, and may be forced to use Library internet access provided the Library can afford the additional expense.

    Obligatory political statement:

    Of course this plays right into what the current Republican administration is trying to do in the United States, cleave society into 2 classes, the haves and have-nots.
    i.e. Business owners == Capital, Labor == The rest of us. Capital will be able to afford an Internet connection, I'm not so sure about the rest of us.

    End of political statement.

    • Telcos and shareholders of Telcos may not like it if a substantial part of the population can't afford a decent phone with net connectivity.

      Same for ISPs, if a substantial part of the population can't afford decent internet connected / IoT / laptops / PCs / gaming consoles / smart TVs, etc.

  • Cool! Hopefully, at some point soon, the BSDs - NetBSD, FreeBSD and OpenBSD - will add support for it

Slashdot Top Deals

If you think the system is working, ask someone who's waiting for a prompt.

Close