SUSE’s chief technology officer said he believes the practice of “backporting” features from the 2.6 Linux kernel into older versions is a “bad thing” because it interferes with standardization of the open source operating system. During a keynote address at the Real World Linux Conference here Thursday, Juergen Geck also chided industry players to avoid practices that could further fragment open source standardization efforts.
SuSE CTO Takes Issue With ‘Backporting’
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Uhhh…. Straight from the v9.0 press:
Linux kernel
* Optimized Linux kernel 2.4.21 with many backported features from the next kernel generation 2.6, including an improved scheduler, ACPI, and ALSA (Advanced Linux Sound Architechture)
http://www.suse.com/us/private/products/suse_linux/i386/new_feature…
Is the CTO a bit Jealous? The thing is he actually wants to base Suse off of Red Hat’s kernel. Red Hat still focusing on the 2.4 kernel will have none of that until they are ready.
Should the linux kernel move to a time-based release schedule of some sort? It would keep the user-base closer to the development tree, thereby reducing the amount of backporting that goes on (among other advantages). Whether or not SuSE is being hypocritical, the current situation does seem problematic.
-Tom
So RedHat’s commercial distro is going to use 2.4 for a while longer. Fedora is already incorporating kernel 2.6, so it’ll only be a matter of time before this gets propagated to RedHat Enterprise Edition.
Kernel 2.6 is still relatively new, so it makes sense that RedHat sticks with a kernel that’s already got a proven track record. What’s this got to do with SuSE or ‘standardization of the open source operating system’?
It really sounds like SuSE is envious of RedHat.
what difference does it make if redhat backports? stuff gets backported all the time outside of redhat as well, and redhat releases any and all changes to the kernel it makes. all redhat does is take a stock kernel and apply their own patchset. (at least in terms of what doesnt get pulled into the official 2.4 line, as redhat definately puts more into kernel development than suse)
i fail to see a problem, especially when as VSW points out, suse is guilty of using a kernel with backported features as well.
You know, I like both distro’s to be honest. I use Slack at work for server stuff, but at home and on my laptop I’m constantly changeing between suse, fedora, mandrake… I like them all. I don’t care what the y do to be honest, backport all you want. It’s open source, right?? Personally, I think there’s a lot to be said for sticking as vanilla as possible, but to see suse jab at someone else when they’re doing the same thing seems kind silly.
Are we down to the ‘2 big kids on the block’ type of mentality know??
SuSe is just upset they can not list all these features in thier PR release of the new box.
They rushed to be “the first distro with 2.6!” But they can’t list many of the features since
Red Hat either has had them since 9 or basically wrote them. Linux is definatly a mainstream
distro, fud is not just comming from users anymore, It’s from CEO’s and VP’s.
Don’t know where “linux is a mainstreem distro” came from, Just got home from work and having a beer, bah. I ment linux is def main stream for those would-be replys.
I think the difference there is that was 2.6 even released when SuSE 9.0 came out? They backported from a non-final version of the kernel but now that it’s final, it should be used as is.
That would be my interpretation of it anyway.
Either way, I agree. It seems to show a lack of trust in the process to not go with the latest and greatest to me. But I’m not a linux-phile, just linux-curious.
Nobody sane wants Red Hat’s kernel.
yes, SuSE backported _before_ the 2.6 was out.
Backporting because you can’t handle a 2.6 kernel but need to rush out a distro, Red Hat fashion, is not a good plan, because you are missing one important ingredient – the 2.6 kernel with its enormously improved performance on, for example, Apache. Important to a lot of us.
“Backporting because you can’t handle a 2.6 kernel but need to rush out a distro, Red Hat fashion, is not a good plan, because you are missing one important ingredient – the 2.6 kernel with its enormously improved performance on, for example, Apache. Important to a lot of us. ”
However, in terms of stability, 2.4 is tried and true. 2.6 simply doesn’t have that kind of field exposure yet. I’m not arguing that 2.6 is not stable, only that it’s better to go with the thoroughly tested solution rather than the new sexy one.
If medium-sized and larger businesses want to handle more load, they’ll buy new hardware, not upgrade their kernels. That’s just the way things work in the real world. Ergo, 2.6’s performance improvements, while certainly nice, are not really any kind of deal-breaker when it comes to sticking with 2.4.
-Erwos
True – it depends on your needs. But I don’t actually see that backporting is particularly good for stability and ‘bug-free-ness’ either. Could be better, could be worse, depends. It seems an unnecessary duplication of effort.
Remember, some of us have been running the 2.6 pre from SuSE for a good while now, (dual-boot if you’re sane) and SuSE have the experience of actual customer debugging
However, in terms of stability, 2.4 is tried and true
I find this argument really amusing due to the fact that the code beong backported is no more “tried and true” than the same code in 2.6. How then could the resulting patched 2.4 kernel be called “tried and true” by someone with both a clue, and a straight face?