Linux on Clover Trail: Much Ado About Nothing

clovers_feat

The tech world has been buzzing recently about comments made from Intel’s Architecture Group Executive Vice President David Perlmutter at the Intel Developer Forum, where he said that the company’s new “Clover Trail” chip “cannot run Linux”. Moreover, Perlmutter claimed the Clover Trail to be a specifically Windows 8 chip, and that Intel had worked closely with Microsoft to implement the new power saving features exclusive to the Clover Trail.

Clover Trail

Clover Trail is the latest in Intel’s long line of Atom processors, this time designed specifically for tablets. While the Atom chips have always accomplished some impressive feats of power conservation, the new 32nm Clover Trail processor is able to drive energy consumption low enough to be viable for tablet usage.

To accomplish this, the Clover Trail features new power management features which essentially allow it to turn off the processor cores on demand. This trick naturally requires support in the operating system, which is exactly what Intel has been working so closely with Microsoft to develop; and what they aren’t so keen to release to the open source community.

Intel’s reasoning for not sharing this Clover Trail magic with the public is anyone’s guess, but likely one of them is the fact that Intel would like to keep these features under wraps for as long as possible to keep competitors from implementing it. There may even be an exclusivity agreement in place with Microsoft, further compounding the issue.

Linux Support

Whether it’s possible for Linux to run on the new Clover Trail chip is not really an issue for debate, because certainly it can (and will). The Clover Trail, despite its new power saving features, is still just a standard x86 processor; Linux will certainly boot and run on this processor out of the box.

The real issue is those new power management features. Intel has made it clear they do not intend (at least for the time being) to directly support Linux, so the kernel developers will be on their own to implement the new functionality. Depending on how much documentation Intel releases about these features, this task could either be a relatively simple matter or a massive reverse engineering undertaking. At this point, there is simply no telling.

Concerns and Conjecture

At this point” is really the key takeaway from this whole issue. The Clover Trail chip hasn’t even been released yet, so the fact that the Linux kernel is not currently equipped to handle its new features isn’t news: it’s common sense. Making noise about this issue before the chip has even hit the market is pointless, as there simply isn’t enough information to go on. The issue is how long it takes after release to get Clover Trail working 100% under Linux without Intel’s support.

Which brings us to the second point, support from Intel. The Linux developers and the community, from kernel all the way up to userland applications, are no strangers to operating without support from hardware manufacturers. If anything, that’s always been the norm. While the recent increased involvement from hardware manufacturers has been a blessing, it’s by no means required. If there is sufficient demand, the kernel developers will get it supported.

Demand is the final point in this topic. Is there demand for this functionality in the mainline Linux kernel? With Clover Trail being exclusively targeted for Windows 8 tablets, is there really an opportunity for Linux to even run on this hardware? Will the hardware be locked down and controlled in such a way that getting Linux to work with the advanced power management functionality of the Clover Trail be the least of our problems?

One thing is for sure, until devices with Clover Trail start shipping and Linux developers get their hands on them, everything is conjecture.


Tom Nardi

Tom is a Network Engineer with focus on GNU/Linux and open source software. He is a frequent submitter to "2600", and maintains a personal site of his projects and areas of research at: www.digifail.com .

Related posts

  • Shaun Hunter

    It’s Certainly an about face for Intel. I wonder how much this cost Microsoft? HP paid Intel more than half a billion to keep Itanium going until 2017, any guesses?

    AMD CPU’s already switch off cores without new instructions what’s the benefit to this? I’m at the point where I think that if a company doesn’t support FOSS it shouldn’t support them. FOSS communities have enough power to change the tune of even the biggest players. Look at the recent changes at Nvidia for example.

  • Charlie Whitman

    Intel’s Linux support for their Atom SoCs appears to be concentrating on the upcoming Valley View series of processors. These are a more logical candidate for full Linux support because they don’t feature PowerVR GPUs, like Clover Trail does. That is, PowerVR refuses to support any open source drivers for their GPUs, so the SoCs that feature these GPUs have problems with Linux support already.

    Valley View will feature Intel’s own next generation low power GPU, rather than one from PowerVR. Intel is planning on full Linux support for this series of SoC from end to end. This may end up being the best Linux support for any SoC ever made (the GPU is often a sticking point), so I’m not going to write Intel off entirely just yet.

    We see, however, with this deal, and the similar deal Microsoft made with AMD on their Hondo APUs for exclusive Windows 8 support at launch, that Microsoft hopes to continue creating stumbling blocks for competing operating systems. AMD appears to indicate that the deal for exclusive support for their chips is only temporary, and that Linux support may be given in the future. They also assured people that Linux would run on the chips, but might not have support for all the special features of the series (at least for a while).

Top