Intel Open Source Continues to Lead in Linux Kernel Contributions

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
According to a report released by the Linux Foundation today, Intel tops the list of companies contributing to the Linux Kernel from Sept 2013 through Jan 2015. Of the 239 companies that made submissions to the open source project, Intel made 10.5% of the total contributions. Additionally, Intel has been the most active in introducing new developers to the project. Linux powers 98% of the world’s super computers, most of the servers powering the Internet, the majority of financial trades worldwide and tens of millions of Android mobile phones and consumer devices. The Linux Kernel is the foundation for software running on Linux systems and Intel’s contributions are part of the work done by the Intel Open Source Technology Center in the Software and Services Group.
 
If Intel wants to make sure that the current and future super computers and servers run on their processors, they better make sure the operating system code being used is written to take full advantage of those processors.
 
I trust AMD will get to their contributions right after they finish that next driver. :)
 
From what I have seen, at least Intel's contributions seem useful and help the platform, for servers, etc.

Unlike most of the contributions aiming at the ARM side of Linux -- where the move to systemd and Device Tree are making it MUCH more complicated, much more difficult, and much less useful as a true embedded platform. Yes, as Linux said, the plethora of board files for ARM was an issue and some solution needed to be found -- unfortunately, rather than coming up with a new solution, it finally just gave the Power PC/Open Firmware camp the impetus to ram Device Tree into place as a solution (unfortunately, from an embedded standpoint, Device Tree is a nightmare and is like trying to make a round peg fit in a square hole). I am currently trying to migrate a custom ARM board from 3.0 to 3.10 and creating all the appropriate dts files, but some of the things we currently do with hardware on our platform may not even be possible under Device Tree without resorting to undoing a bunch of mainline kernel mods and reverting some drivers back to the way they used to be -- which is a nightmare of it's own. Someone really needs to come up with an order of operations for Device Tree -- make it a true "Device Script" that can do things in sequential order (which is often necessary for bringing up hardware with timing dependencies) and add compatibility with and the ability to pull from the Linux header files (which violates the "OS independent" philosophy behind the current scheme -- but the drivers it is configuring are already kernel dependent and header dependent anyway, so having actual headers shared with the kernel would beat the heck out of a 1001 "magic numbers" that require digging into /include/linux to figure out what they mean anyway!).

Based on our own experiences, we are really wondering if migrating away from ARM and back to Intel (Atom, etc) for the next project might not be the best option. The Atom's beat the heck out of the ARM chips in performance pretty much across the board -- and the power management is MUCH better and the chips themselves run a LOT cooler, even though they are way higher in clock frequency. The ARM parts may be cheaper, but I'm not sure it's worth the difference in engineering headaches (and trying to keep some of the quad core Arm7 parts cool without active cooling is NOT a trivial task).

I for one am glad to see Intel helping the Linux kernel out so much -- it is a worthwhile investment on their part and is good for everyone.
 
Nice, this is some really good news.
x86 is hear to stay, and the more efficiency Linux has with it, the better off we all will be.
 
Back
Top