Linux Kernel 2.2.26 -- 2.2 is not dead!

midianus writes "Marc-Christian Petersen released Linux Kernel version 2.2.26. The release includes several security fixes. After Alan Cox released the previous 2.2 kernel, 2.2.25 in March of 2003, he began a one year sabbatical to study for an MBA in August of the same year and handed over the Maintainership of 2.2 to Marc-Christian."

2.2.?

[Ann Coulter]

2.0.? is not dead either. Their 2.0.40 was released only two and a half weeks ago.

Re:2.2.?

[gustgr]

I was up to say that. After that new 2.0.40 release I have started beliving in life after death.

Re:2.2.?

[$calar]

And a 2.4.26 prerelease is out too!

Re:2.2.?

[Zork the Almighty]

How well do these earlier kernels perform on old hardware, say a 486 ? Are they lighter than the 2.4 and 2.6 kernels ? I've never really had a chance to try them.

Re:2.2.?

[alienmole]

Not quite a 486, but I've been running kernel 2.0.36 on a Pentium 90MHz with 64MB RAM, since 1998. It hosts mail (SMTP+IMAP) for a few small domains, CVS, and a mainly static web server. It also runs the Twiki wiki for a small workgroup. That's Perl-based, and it's the one thing which is a little slow, that we're planning to migrate to a faster box.

I've never tried running a newer kernel on a box that old, though, so I can't say how it would compare. 2.0.36 definitely uses less RAM, though - the vario

Confused

[kurosawdust]

I don't get it - what exactly does this guy do with the 2.2 kernel? Just bugfixes? How does he know he's not working hard to fix a bug that was fixed by someone who noticed it in 2.4 or 2.6? Not knocking the guy or anything, I'm just absolutely baffled as to what he does with the 2.2 kernel.

Re:Confused

[Haeleth]

If someone fixes a bug in 2.4 or 2.6 that was also present in 2.2, it makes sense to fix it in 2.2 as well. That's one of the jobs of the 2.2 maintainer.

Re:Confused

[Vlad_the_Inhaler]

Here is the Changelog in all it's glory. Mostly security fixes.

If you need a more secured, more stable, even faster, better SMP support, IDE LBA48 support, LFS support, IPSec support, HTB support, IPVS support etc. enabled kernel with many more features and important fixes, please use my 2.2-secure tree.
You may find more informations about it at http://www.wolk-project.de.

2.2.26

• CAN-2004-0077: behave safely in case of do_munmap() failures in mremap(2)
• CAN-2003-0984: /dev/rtc can leak parts of kernel memory to unprivileged users (2.4 backport)
• CAN-2003-0244: hashing exploits in network stack (David S. Miller)
• update_atime() performance improvement (2.4 backport) (Solar Designer)
• ability to swapoff after a device file might have been re-created
• MAINTAINERS correction for Kernel 2.2 and 2.2 fixes (me)
• fixed some typos (Solar Designer, me)

Sorry, had to edit that a bit to get it past the lameness filter. Most (all?) of the other patches here came from Solar Designer.

COOL!

[Sevn]

So this makes it into debian stable by say, December 2005?

Re:COOL!

[bhtooefr]

Actually, starting Debian Woody Netinst with bf24 gives you a 2.4 kernel. So, since it's an update to the DEFAULT kernel (linux, instead of bf24), it'll probably hit security.debian.org.

Re:COOL!

[sydb]

I know you're kidding but we need to fight this false image Debian has unfairly gained.

Debian has 2.6.2 images in unstable. And just recently got Firefox builds too.

So Debian is bang up to date, if you want it to be (by running unstable).

Re:COOL!

[damiam]

I use Debian and love it, but let's not be too misleading here. It took a couple weeks for Firefox to get into unstable, and kernel 2.6.0 was still the newest option even for a while after 2.6.2 had been released. Unstable is great, but it's not always up-to-date (it took a year for XFree86 4.3 to get in, although that's a whole other can of worms...).

Re:COOL!

[cbcbcb]

Not only did it take a year for XFree86 4.3 to get in, but now that it has, it DOESN'T EVEN WORK! I have no 3D and screen corruption, which is not present in the hand-compiled X build I was running previously.

Re:COOL!

[alexpage]

Unstable is up-to-date across all architectures. There's more to Linux than x86, and Debian is one of the distros which recognises it. Bleeding-edge x86-only packages are normally available in the experimental branch, if you need them that badly.

obligatory joke

[Tumbleweed]

The 2.2 kernel isn't dead...it just smells that way. :)

SCO?

[TekZen]

I remember something about SCO's claims not affecting any kernel before 2.4.

Quick! Avoid the SCO tax, downgrade to 2.2.26 today!

-Jackson [jaxn.org]

Just like...

[jutski]

...saying "Pope - He's not dead, yet!"

Number of minor releases

[reignbow]

Is there a reason that 2.2 has only one tiny release more than the 2.4 kernel? I mean, if you look at the most recent tiny versions:

2.0.40
2.2.26
2.4.25
2.6.3

I believe that the releases of 2.x.0 were about equally spaced, so why isn't there 2.2.34 or something like that coming out? Fewer bugs? A more laid-back maintainer? Not as much movement in the OSS world?

Re:Number of minor releases

[bersl2]

The organization of releases is a subjective matter, at the discretion of the maintainer.

Without specific context, the number of revisions is meaningless.

Re:Number of minor releases

[ocelotbob]

Really, kernel releases in the Linux world are more related to how fast the kernel gets to the point of stability than anything else. If you look at the history of the 2.4 kernel, you see that there were a lot of early problems with areas like the memory manager which caused a lot of releases in an attempt to get the problems taken care of. 2.4 really didn't stabilize in most people's eyes until 2.4.15 or so. Plus, once a new stable kernel comes out, the old stable enters a sort of maintenance mode. Few, if

Re:Number of minor releases

[oobar]

Ummm, coincidence maybe? There is no relation between the releases.