Slashdot Mirror
Linux 2.4.22 Stable Kernel Released
An anonymous reader writes "Marcelo Tosatti has officially released another stable 2.4 Linux kernel. 2.4.22 was released early this morning and includes a lengthy list of fixes. It follows the last stable kernel in this tree, 2.4.21, by a little over two months."
For those of you who use bittorrent, try:
linux-2.4.22.tar.bz2
patch-2.4.22.bz2
Choose yer poison: Prophets or Profits
I've been running the 2.6 test for a while and the latest O1int patches from con have really made this a pleasure for desktop users. 2.6-test3-mm2 had massive skips when playing xmms and untaring, like 2.4 also has on my machine. But 2.6-test4-mm1 completely fixed this. I have not had a skip yet. Please not that this is from normal usage.
I don't have detailed info on OS design, but probably you'll want to take a look at the OSKit and Exokern websites, as they do go into some depth on design issues.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Well the last one took over 6 months. Remember that most people won't install and test kernels unless it's "final"
I use:
emilio@ellugar:~$ uname -a
Linux ellugar 2.6.0-test3 #9 Wed Aug 20 15:21:40 CEST 2003 i686 GNU/Linux
I can say 2.6 feels better (mainly from better disk I/O scheduling), but the process scheduler starves sometimes.
Anyway, worth trying it, I think it's almost ready.
Is it possible you made a mistake installing the new one?
If not, you would be performing a public service (in a small way) by trying to track down the problem and reporting it (e.g. on linux-kernel). This is a stable release, and dropped support for any hardware or configuration is a bug.
Well, you heard wrong. The NT kernel is not good. It's faculty for determining timeslices and aggregating CPU time is convoluted at best. It is all designed to make user level timeslice requests (ie GUI events) occur more quickly. The solution provided by NT kernel is a timeslice lookup matrix with more elements than I can currenly remember to cound (something like 3 levels with 11 sublevels for each).
When compared to the Linux or even Solaris CPU timeslice allocationg, it is horribly overcomplicated and slow and clunky.
This short guide walks you through the steps used to compile and install a new 2.4 kernel.
Sorry, this is an incomplete list for the 2.6 series, while the kernel released is 2.4.22. For a almost complete list of features going into 2.6 you can see Linux Kernel 2.6 Status
No he is not, he is the boss of the 2.6 releases.
Here's what you can do to use vanilla kernels in RedHat. It may not be the easiest route, but it works like a charm every time.
1) Enter your regular redhat kernel source directory. Everything should be preconfigured, unless you changed things;
2) make menuconfig, then Save Configuration to Alternate File. Pick a name you'll remember;
3) Exit menuconfig, enter vanilla kernel directory;
4) make menuconfig, Load an Alternate Configuration File. Enter the name of the previous config, with full path.
5) Profit! You might want to check any new options just for fun (such as built-in crypto for this kernel release), but it isn't necessary.
Your new kernel should now run just fine under RedHat, with a minimum of fuss. It'll be a long compile, though, because even the kitchen sink will be included. Still, it's more work for the computer and less work for you.
It won't. Really, if a kernel works for you, and you have no interest in upgrading it if a new release only contains bugfixes, don't. It would be different if there were serious security issues, but it's just a bugfixing-release (with speedtouch usb and crypto support added). Repeat after me: 'My Life Will Go On With An Older Kernel', there ya go...
This sig is intentionally left blank
Building a Custom Kernel.
enjoy!
i tried several times to compile a kernel on redhat until i realized that they have changed things a bit, and you need to follow their docs.
Schrodinger's cat is either dead or really pissed off...
Check out this great writeup by Con Kolivas on the subject. We've got a lot of really great minds looking to improve interactivity in the kernel. Great stuff, and no doubt more to come.
Look out honey, 'cause I'm using technology; Ain't got time to make no apology
Since RedHat adds a bunch of patches you'll need to start with the Linus source that the RedHat kernel is based upon, then add the patches from that kernel to bring it up to the new.
It's not exactly easy.
There will be hundreds of files that change from the stock kernel to the RedHat kernel. Then there are the hundreds more that change from the stock release to the next stock release.
Your best option is to use a kernel from the RedHat beta releases then recompile it for the current. These will *usually* work, but there are no guarantees. It's pretty simple to rebuild a kernel RPM:
rpmbuild --rebuild --target=athlon kernel-2.4.22-20.9.src.rpm
Someone mod this down ... it is a troll that has been posted before. These are some 2.5.X patches that will be in 2.6, nothing that is in 2.4.22. Read the real changelog that was linked from the article. Moderators ... wake up!
Do you even know anything about perl? -- AC Replying to Tom Christiansen post.
I see that nobody bothered to answer your question. Offtopic posts seem to multiplying as fast as spam these days on Slashdot.
8 59-1/boo ks/design-44bsd/
- HO WTO/index.html
BTW, have you heard of Google???
BSD:
http://freebsd.unixtech.be/doc/en_US.ISO8
Linux:
http://www.kernelhacking.org/docs/kernelhacking
http://cs.uml.edu/~cgould/
I'm running the current 2.4.x release that Gentoo has and I tried with the preemptive kernel options, but it only seemed to break a lot of modules. Even deleting the modules folder and rebuilding caused many of the modules to fail on startup saying that it could not find the preempt function (something like that)make clean / make mrproper didn't help either so I'll wait to 2.6 is gold before messing with it again.
Unfortunately, for most of the world, releasing a new kernel doesn't mean much until a distro releases it in a release.
Sure it does. You take your old config do a 'make oldconfig' and it'll prompt you for the new features. Then you just compile and install as usual. It's easy.
Give me Classic Slashdot or give me death!
Changelog is your friend.
Summary of changes from v2.4.22-pre5 to v2.4.22-pre6
snip
o fix a race in the plugin api for ac97
o example ac97 plugin codec
Then, a bit earlier (lower down in the Changelog) and also from Alan:
o update AC97 codec core
o switch cards to new ac97_audio
o switch i810 to generalised digital out, new ac97
o ac97 updates
o update trident, fix printks, new ac97
o Update via audio - fix problems esd, mpg321
o update to new ac97_codec
o update ac97 codec headers
Mielipiteet omiani - Opinions personal, facts suspect.
Is the ptrace race-condition vulnerability fixed?
.21 fixed that.
:-)
Still running 2.4.20? Because
It's certainly more important than some whitespace convention in 2.5, or
updates to JFS (which may or may not contain copyrighted code.)
Of course JFS contains copyrighted code... somebody had to write it. I think it was IBM.
Forgiven.
... 2.6. The general idea is to backport security, stability and driver patches and leave the architectural changes alone.
The changes will need to be dupicated from one series to the other.
A change to the 2.4 series may be submitted to Linus for inclusion in 2.6, and added if he deems it worthy. The process also works the same way in the other direction. For instance, if Linus accepts a patch to the 2.6 series, then the 2.4 maintainer may choose to include a "backported" patch for the 2.4 series. Alan Cox (I believe) has been the 2.2 maintainer for a while now. (Remember the beloved stable 2.2 series?) If he feels a 2.4 series patch is essential he may backport it to 2.2.
The decision to backport is not automatic, since applying all the patches from 2.6 back to 2.4 would yield
-Jon