Linux 2.4 Kernel May Finally Go End-Of-Life

[nanonyme]

It's unfortunate that no one is interested in back porting hardware drivers. I think we're lucky to get them past two kernel releases.

Linux is becoming the rich man's motor carriage.

It's mostly that everyone's accepted that it's the job of distros to backport drivers from upstream releases to the previous one. If your distro doesn't do that and you think it should, you might want to start looking at another distro.

[Apopas]

Heh well Popular Science did list these jobs as the worst in science:

http://www.usatoday.com/tech/science...obs18_ST_N.htm

Hehe come on the "Oceanic-snot diver" is much more interesting job. Just the diving part is enough to convince me

[gilboa]

... Ugh. 2.4.
Anyone that ever lived the transition from 2.2 to 2.4 knows how amazing it was. (E.g. switching from RedHat Linux 6.x to 7.1-2).
It was the first time I had fully functioning machine, USB, network, GPU (via nVidia's proprietary drivers) - the works.
Coincidentally, RedHat Linux 7.1 marked the last time I had Windows (2000) on my primary machine.

Nostology aside, people who still use 2.4 (mostly embedded systems) don't really need additional features - heck, in many cases they don't even need security patches. (The last I used 2.4 we had an extremely optimized kernel with our own patch-set that was tailored for our machine and was capable of full boot within 4-5 seconds - most likely its still being used...). For these embedded markets, having a maintained kernel is a bonus - nothing else.

- Gilboa

[squirrl]

Time will tell

Boot 2.4.31 and look at how much ram it uses.

This was the most efficient series. It started life with a very interesting memory manager by Rik Van Riel. RedHat even kept their own version. http://kerneltrap.org/node/26

It's unfortunate that no one is interested in back porting hardware drivers. I think we're lucky to get them past two kernel releases.

Linux is becoming the rich man's motor carriage.

[droidhacker]

candidates might include quantum computing, or major revisions to the layout of the system if we stick with silicon.

You aren't trying to suggest that one couldn't engineer a silicon-based quantum computer, are you? You probably mean if we stick with BINARY. Being binary or quantum isn't necessarily linked to the particular material used to implement it.

[DeiF]

Hey, I used Linux 2.4 for at least a year or two when I was just starting out with Linux. It was the tried-and-true kernel at that time; ordinary users were discouraged from trying the brand-spanking-new-and-often-broken 2.6. I'm sure some smart ass will come by and mention how they tried Linus' initial release of the kernel the microsecond he released it, but at least I can say that I have used two different release series of the kernel while each respective series was the prominent recommended kernel. That said, I don't really remember the times before 2.6.0 was even released... that was right on the periphery of when I first heard about Linux...

He, me too. I started with SuSE 8.0 that included the 2.4.18 kernel (I had to google it), although I had SuSE 6 somewhere from a computers convention that I never used.
It's great to see how much things have evolved.

[allquixotic]

Hey, I used Linux 2.4 for at least a year or two when I was just starting out with Linux. It was the tried-and-true kernel at that time; ordinary users were discouraged from trying the brand-spanking-new-and-often-broken 2.6. I'm sure some smart ass will come by and mention how they tried Linus' initial release of the kernel the microsecond he released it, but at least I can say that I have used two different release series of the kernel while each respective series was the prominent recommended kernel. That said, I don't really remember the times before 2.6.0 was even released... that was right on the periphery of when I first heard about Linux...

I remember 2.4 as being pretty darn good, in fact. Except for advanced hardware support like wireless and graphics, it was quite good at making efficient use of CPU, RAM, DMA, 100 Mbps networking, process/thread management, etc. There have definitely been improvements in 2.6 and micro revisions thereof, but most of the groundwork for the kernel we enjoy today was already there in 2.4.

I think 2.6 is here to stay, though, until some major fundamental change in the hardware ecosystem forces a major re-thinking of concepts. I'm not sure what such a change would look like, but candidates might include quantum computing, or major revisions to the layout of the system if we stick with silicon. It would have to be more revolutionary than eliminating the FSB, though, since 2.6 gently evolved to support the Nehalem architecture and its rather significant departures from the past 10-15 years of computer architecture.

[RealNC]

OpenVMS 8.4 was released recently, it has full IPv6 support and runs the latest version of apache. I'm testing with it now.

Sounds like lots of fun.

/sarcasm off

[frantaylor]

Kinda sounds like a "I'll maintain it if I feel like it or of somebody else really wants me to". This way if nobody is interested after a long time he doesn't have to feel obligated to get it up to snuff in an "officially supported" way.

Probably a ton of cheap embedded crap made in $AsianCountry shipping with kernels old even by 2.4 standards which never get updated by either end-users or manufacturers. They probably won't even notice that it's not being maintained.



I was gonna say maintainer of obscure 30-year-old GNU utilities such as ed or M4 - but the only thing worse than old and crufty is old, crufty, and huge.

OpenVMS 8.4 was released recently, it has full IPv6 support and runs the latest version of apache. I'm testing with it now.

[deanjo]

I provoke everyone to find a more mundane job than the maintenance of the 2.4 kernel...

(the maintenace of 2.0 does not count as an answer...)

Heh well Popular Science did list these jobs as the worst in science:

http://www.usatoday.com/tech/science/2010-03-22-worstjobs18_ST_N.htm