I was just wondering why everyone seems to be so eager to upgrade to 2.6. I'm running 2.4X on slackware and everything seems great...I'm hoping someone will point out a few tasty features that will compell me to switch.
Yes I could go and look at some revisions page but those usually go: "added support for virtual COMM wrapper persistence by updating KCOMMLIBB to 2.28898238" so?
what does that MEAN? is it a good thing?
anyway if anyone can give me some practical diffs, much appr.

regs
psycode.

i personally think we should be eager to upgrade, because otherwise we will be stuck where we are.

then one day a will come feature that requires kernel upgrade, but since upgrade wont be popular there won't be any manuals, tools or people to help with the upgrade. leaving us stuck. if we keep this up kernel upgrade then one day it will be like typing "apt-get upgrade" (without complications).

If the 2.4 kernel is working for you, then there is absolutely no need to move to 2.6. The only reason I moved to 2.6 is that it is required for the wireless card drivers I use. A lot of people work under the "newer must be better" fantasy, but if you are simply looking for a good, solid, stable system to work on (as opposed to tinkering on) then 2.4 is a good choice.

Yep, no reason to upgrade unless you need support for newer hardware. Netfilter is quite nice if you need routing capabilities.

I upgraded because of the need to support most of the hardware on my laptop, specially the wireless networking, cpufreq capabilities and the 64-bit processor

Then I liked the boot-up speed and lil' useful things like udev which creates the devices on /dev as you need them and that the kernel loads and unloads modules on demand

Give it a try, you can always go back if you don't like the results; and anyway, you might learn something new

Because it's fun!

I use udev and hal/dbus to automount cd's which requires a 2.6 kernel

There are always security fixes.

er, um, let's see:

- improved networking
- performance optimizations throughout the kernel
- greater hardware support
- automount
- security fixes
- support for more filesystems
- break past the 2GB file size barrier

Other than those reasons, upgrading to 2.6.x is pointless.

I guess you can go use 2.2 if it works.

2.4.x still gets security fixes

Please, lets not raise a security bugaboo without some evidence. The 2.4 version is being actively maintained and is every bit as secure as 2.6. If anyone has any information to the contrary, please post a link.

If you decide to upgrade due to udev support it's not correctly working in slackware yet, but it is worth to try hal and dbus.

Wow, first off I'd like to say thanx to all those that responded...
As for an update: i'm going to stay with 2.4x, heres why..

- improved networking

I run a multiOS home network smb, a proxy, windows boxes, the whole trip, I've made it as complex as I could for learning purposes, for example I am not slacking in the security department which people sometimes do in favor of simplicity. I have not yet come accross anything that required me to upgrade. I'm not saying these situations don't exist just that in a fairly standard small lan/appache/shh/smb/proxy server enviroment I can honestly say I have not found anything that needed 2.6 capabilities.

- performance optimizations throughout the kernel

ok here here you got me, I really have not studied much kernel code ( a little sockets coding but nothing much else ) so I really can't comment.

- greater hardware support

heres the kicker, when I tried the "test" kernel on slackwares' cd, the modules
supported LESS of my hardware (for instance my soundcard) than the standard install. Results may vary, but there it is..

- automount

Look, I dont want to argue with the masters but I think that this is distro specific. For example I ran RedHat 9 with 2.4 and it was fully automount capable. Besides, I like typing "mount -t xxx /dev/hdc /mnt/cdrom" don't ask me why

- security fixes

I think other users have already made the comment that security fixes are still released for 2.4x.

- support for more filesystems

Really? name a filesystem that I can't mount in 2.4x (and that includes ntfs,mass storage devices, etc) yes the support may not be native but that is a technical footnote not an obsticle.

- break past the 2GB file size barrier

if I had harddrive space for 2G files I'd be running RedHat

So guys there it is, for now. I'm still going to upgrade (so that I can do the slapt-get upgrade thing one day) just not now.

regs
psycode

I think this is a point which gets confused. The term "automount" gets thrown about pretty regularly for several different processes. There was (is?) an older system of automounting that used (if I recall correctly) autofs. Some distros used it by default (Mandrake?) and it could be added to any distro manually with patches. Lately, though, I think that most people who refer to "automount" are talking about the Project Utopia stack (HAL/DBUS/UDEV). Since UDEV is used with the 2.6 kernels, you need a 2.6 kernel to get automounting with the Project Utopia stack.

The really nice thing (IMHO) about having the Project Utopia stack working properly is not added function. I know how to mount a device. The benefit is the convenience. Other people who use my system don't have to know anything. Plug in a USB drive, just works. Plug in an iPod, just works. Throw in a CD (data or audio), just works. Hook up a digital camera, just works.

It is still possible to use a 2.4 series kernel with autofs to get the old form of automounting. This is most likely how the RedHat 9 system that you used before was configured. The interesting point is that autofs never seemed to really catch on (I don't think it was ever mainline kernel stuff). UDEV is definately mainline kernel stuff in 2.6.

On to the main topic of this thread, why upgrade to 2.6. There are a bunch of reason to upgrade or not to upgrade to a 2.6 series kernel. The one that struck with me hasn't been mentioned yet, so I thought I'd throw it out there.

When I first started with Slackware, there was no option to use a 2.6 kernel during installation. Still today, the easiest method for installing Slackware is to use a 2.4 series kernel. The first thing after installation was to configure various things and learn my way around Linux in general. Before long, I was feeling comfortable. I was using a lot of different sources for learning. The only thing I had not touched on at all was compiling a kernel.

Now here is something interesting to think about. I suggest that anyone who works with Linux will at some point benefit by knowing how to compile a kernel. It's not too difficult a skill to learn, but EVERYONE makes mistakes while they are learning how to do it. Now mistakes can really hose your (already working) system. Anything which reduces the chances that you will render your installation un-usable is a good thing. One good idea when learning kernel compiling is to not recompile the same kernel that is already working. /lib/modules can quickly become a big fat mess. You already have a working 2.4 series kernel. Why mess with that while still learning? Jump to a 2.6 series kernel. Since the 2.4 kernel on your system is already working (and should continue to work if you compile new kernels correctly), I think compiling a 2.6 kernel is a great learning tool. Keep compiling until you get it to work. After that, re-compile until you get it to break. Eventually, you will be able to configure a kernel at will. This is a handy skill to have. Learning this way, you can always still use your system with the 2.4 kernel that has been working.