Linux optimize

Linux Kernel 3.7 flexes its ARM muscles

With the Linux 3.7 kernel some major changes are in order. In particular, the support for ARM technology has arrived. Jack Wallen pontificates on what this could mean for the open source platform on multiple levels.

I get a lot of questions regarding the status of Linux on tablets, mobile devices, and other platforms running ARM processors. Most end users and consumers simply want to know if/when Linux would ever find its way onto the "system on a chip" architecture.

The reason we haven't seen Linux on these devices is simple: supporting the various ARM processor derivatives has become almost impossible to manage. The developers of the Linux kernel were looking at over 70,000 lines of new code for each new kernel release for ARM technology, versus roughly 5,000 lines of new code for x86 technology.

That all changes with 3.7. The upcoming Linux kernel release will finally, and officially, support the following ARM processors:

  • Calxeda's Highbank ARM server processor
  • The Versatile Express ARM-based developer board
  • Marvell's ARMADA SoC processors
  • Altera's ARM processors
  • Picohip's picoXcell chip for femtocells

That short list pretty much covers everything from high-end data center-level servers all the way down to mobile devices.

It's officially "make or break" time for Linux on tablets and smartphones. Why? Not only does this release further (and more efficiently) support ARM, it also has far more support to HID (Human Interface Designs) -- especially multitouch interfaces on capacitive touchscreens. All is ripe for Linux on mobile.

Up until this point, Linux had the perfect excuse for not finding its way onto tablets. And let's face it, if Linux is to succeed in the mobile space, tablets are the perfect platform. But should they not be able to make the leap now, the chances will grow continually slimmer of this ever happening -- fast!

Canonical is perfectly primed to bring Ubuntu (and Unity) to tablets and now the proverbial floodgates have been opened -- thanks to the developments in the kernel -- for this to happen. And, after becoming quite the fan of Unity, it is clear that the interface is perfectly suited to the tablet (more so than any other interface I have used).

But this isn't just about mobile devices. Server architecture could easily make a mass migration from x86 to ARM. With a shift to low-power chips and the focus moving to parallelism and frequent use of large amounts of data, the concern is on power. ARM solves that concern and, with the ability to easily support Linux now, those servers could be low cost on both power consumption as well as platform. Data centers are now looking at a cost effectiveness that hadn't been an option prior to kernel 3.7.

It's amazing what a difference a release can make. Although most desktop/laptop Linux users aren't nearly as affected by 3.7, as a whole, the Linux landscape should see a significant shift once this kernel is out in the wild and in use. The big test will be how Linux then manages to gain footage on the mobile space. Should that happen, everything in the PC industry could shift.

Although I am a fan of Android, I will happily drop my current tablets for a Ubuntu and Unity-driven device. I have enough faith in the interface that I fully believe it could really take the tablet stage and demand the spotlight from Android and iOS. More power and more flexibility in an elegant multitouch interface. What more do you want?

So, to the developers of the Linux 3.7 kernel -- a huge cheer should ring out for you. This is an important development that could bring huge change across the nerd-scape for Linux, mobile, and server technology.

About

Jack Wallen is an award-winning writer for TechRepublic and Linux.com. He’s an avid promoter of open source and the voice of The Android Expert. For more news about Jack Wallen, visit his website getjackd.net.

6 comments
alastair
alastair

It seems the author of this article doesn't realise that Andriod is Linux and Google has already included Andriod tweaks back into the main line Linux sources. In fact the big difference between Andriod and a Linux distro is that Andriod has the Dalvik virtual machine that runs java Andriod apps whilst a Linux distro would use a desktop manager such as Gnome 3 and is less efficient in memory usage. Thus the kernel is only a small part of the software required to run an OS environment.

alzie
alzie

Id Love to see it too, but ya gotta wonder. Android and iOS have it pretty well sown up. Standards are very hard to buck. Of course, android is a linux, and if this helps with OS fragmentation, all the better. Updates could happen sooner to more folks than now. Less grumbling!

aroc
aroc

Where's the Tegra/Snapdragon/TI support? I would love to be able to convert my Archos 80 to a Linux environment, and have real control over it without sketchy rooting based on exploits, and multiwindowing, and "real" applications I already know. Actually, I would think this could be do-able with a userland type layer above the Linux kernel and drivers that already come with the vendor-specific Android. Could this be achieved by just changing the libraries and other parts like the /usr/bin Busybox executables (that most Androids seem to have)? Some variation on that? Seems like the linux base is there, and we could just change the pieces on top of that to get a "distro" of choice. Maybe just focus on tablets, and not the phones with all their cellular provider lock-down phobias. I guess that is somewhat along the lines of what Cyanogenmod does, but that still seems to have the Android orientation of one app at a time on the screen, and using all that sketchy PlayStore app ecosystem with no worthwhile security, instead of a Linux desktop focus running mainstream, vetted repository apps.. I can dream...

gulli1
gulli1

What about Raspberry Pi? will it be supported in the kernel now? Thanks.

blatanville
blatanville

My office is across the hall from Seneca College's Centre for Development of Open Technology (CDOT) in Toronto, Ontario. There is a team of devs here working on the Fedora ARM translation project, and they've succeeded quite well. I've been informed that one of the primary reasons Fedora Linux is being ported to the ARMitechture (can I get a CC-SA for that? :) is to enable ARM-based Server devices to replace the much larger and hungrier Intel/compatible server processor offerings. One of the images used to sell the project is: Imagine a room the size of an American Football field covered in server racks, UPSs, and cooling equipment. Now imagine that all being replaced by three or four units the size of a household refrigerator, doing the same job, consuming a fraction of the power and requiring a fraction of the cooling. THAT's a goal worth the effort, if you ask me!