Slashdot Mirror
Linux 2.4.0-prerelease is Released
Mark Bobak writes "2.4.0 should be ready soon. 2.4.0-test12 has been superceded
by a new release, and it's not a testxx. It's 2.4.0-prerelease.
Can't wait for the real thing....I can almost taste it! ;-)
Available, of course, at kernel.org" And happy new years to everyone else. Hopefully holiday traffic will be light enough that we can all share the bandwidth to the mirrors. I shall download while I watch the Iron Chef marathon :)
Do you not realize at ALL that Linux is a fragmented (in a good way ;) project, and what the kernel developers do is simply develop the kernel. They don't do it instead of doing other stuff for Linux, they do it because that's their niche. If they release a new kernel, it has nothing to do with other people not releasing better development tools!
A deep unwavering belief is a sure sign you're missing something...
In any emulation layer, there is a performance hit. And if you are incurring the performance hit on higher performance hardware, then people get pissed. I mean nobody really cares if their zip drive is taking a 20% performance hit, since its slow anyway. But when their expensive 12x10x32 Plextor CD-RW drive is taking that hit, then people get mad.
A deep unwavering belief is a sure sign you're missing something...
C++ is great ...
but C++ compilers are crap.
Even tiny problems like the red hat 7.0 gcc would be enormous if the kernel was written in C++. C compilers are relatively simple to implement and this makes them more reliable.
(gcc 3.0 will cure cancer of course. But it's not finished yet.)
Ok. I didn't make 2.4.0 in 2000. Tough. I tried, but we had some last-minute stuff that needed fixing (ie the dirty page lists etc), and the best I can do is make a prerelease.
There's a 2.4.0-prerelease out there, and this is basically it. I want people to test it for a while, and I want to give other architectures the chance to catch up with some of the changes, but read my lips: no more recounts. There is no "prerelease1", to become "prerelease2" and so on.
One thing other architectures will want to catch up with is the changes to handle 2GHz+ machines, which due to overflow issues caused "loops_per_sec" to become "loops_per_jiffy". And some architectures have not had much chance to synchronize with me due to other fires to put out.
Give it your worst. After you recover from being hung-over, of course.
Linus
Yea, but WinNT hasn't been claiming to be a consumer operating system (Mandrake, Redhat, ahem) And they've had AGP support for a long time. Lastly, Win95 had USB support a long time ago, and the BSDs got it a while ago as well. Linux really IS late to this game.
A deep unwavering belief is a sure sign you're missing something...
Its not the language that's the problem. (BTW kernel designers tend to have fits everytime someone mentions C++ going near a kernel;) The problem is that there is no paper standard on what the module interfaces are. That's strictly a Linus design issue, not a problem with the C language.
A deep unwavering belief is a sure sign you're missing something...
The whole "without deadlines" crap is bullshit. Linux has deadlines. They might not be on paper, but if Linus took 3 years to release 2.4, everyone would know that some deadline must have been busted. Linux 2.4 was supposed to come out last year. While that wasn't an official paper deadline, missing the mark by 12 months DOES count as "overdue."
A deep unwavering belief is a sure sign you're missing something...
I patched to the prerelease (tho test12 gave me no problems) because I wanted to help Linus out (however infinitesimally) by making sure everything ran ok on my system before the official release, not because I expected my kernel to start doing anything worthwhile -- kernels don't really have to "do" anything; they just have to stay out the way and not crash.
How can anything be easier than clicking 'Windows Update' checking off the latest service pack, and then waiting for your computer to reboot? In general, Windows upgrades (when they worked, and to be fair, it didn't work for a lot of people) have simply consisted of downloading the relavent .exe, double clicking, and rebooting.
A deep unwavering belief is a sure sign you're missing something...
>Then you have the fact (that unless you use RedHat or Mandrake, which bring their own set of problems) you cannot upgrade the kernel (realistically) without recompiling!
You forgot Debian.
Debian doesn't require you to recompile to upgrade. ( But of course Debian has it's own set of problems )
But in general, I would say upgrading the kernel in Debian is much easier than when I had to do it in Windows 2000.
Nope, I didn't miss a thing. Actually, the test releases have been quite stable for me. As far as I've been concerned, my system has basically been running feature complete with the 2.4.0-test series and reiserfs patches. I guess you also missed the humor of my original post.
In the past 6 months, I've had one kernel panic under Linux ( with test kernels), 100+ crashes with win98/WinME and about 10 with W2K.
If this kernel has no show stoppers, it will be the fork for stable 2.4 and 2.5 devel. Thanks Linus, Alan and crew. I enjoy the fruits of your labor daily. I also endure the stench of the MS crowds labor daily during work.
Kindness is the language which the deaf can hear and the blind can see. - Mark Twain
Try linuxhq.com.
I am not a lawyer.
I remember Linus saying that, after the release of 2.2, future kernel releases would be more frequent. How long has 2.4 been in devel, compared to 2.2?
I'm not trying to flame the kernel developers for late releases, it's just an intrigue. Take your time, guys, because we run servers that depend on your code!
Anyone have the kernelnotes.org on one of their /. sidebars? They're severely behind over there. The versions need to be updated with fury. They're only at 2.2.16 for new stable, which is now at 2.2.18. And they only have 2.3.99 for new devel, which is at 2.4.0-prerelease as you can see in this /. post.
-kidlinux.
what's next ?
- ke rnel ?
2.4-alpha ?
2.4-beta ?
2.4-pre-test ?
2.4-devel ?
2.4-let's-invent-some-other-tag-for-a-year-late
what's wrong with 2.3.x, and when it's ready call it 2.4.0.
Can I just take a 2.2 kernel based recent distribution, grab the prerelease source slap it in /usr/src/linux and recompile?
Or are there issues like we had from 1.2 --> 2.0?
It will also serve as an advocacy and education site to improve the quality of Free Software in general.
If you want to participate, please contact me at crawford@goingware.com or subscribe to the mailing list - instructions are on the site.
Michael D. Crawford
GoingWare Inc
-- Could you use my software consulting serv
But distros are used widely by inexperienced users and this is code that is not extensively tested. Many people get their first and only experienced with Linux from some CD they buy off a store shelf for $29.
Yes, it has been worked on for years but it has not been used in production that much.
It would be better if most of the initial users of 2.4.0 were users who were competent to download and build their own kernels, until the bugs that will inevitably crop up in widespread use are found.
Then let the distros ship with 2.4.1.
Michael D. Crawford
GoingWare Inc
-- Could you use my software consulting serv
2.4-turbo, 2.4-ultra, super-2.4, super-2.4-turbo, oh well i'm happy. i think i play too much street fighter 2 tho.
What do you mean? Win98 supports USB just fine? As for Win95, my USB speakers aren't complaining about it.
A deep unwavering belief is a sure sign you're missing something...
USB stuff isn't "odd hardware." And BeOS was never a factor here (although it DOES support every bit of USB hardware I have, a Perfection 610 scanner, a Sidewinder Precision Pro, and a pair of USB speakers), I was comparing Linux to Windows and BSD.
A deep unwavering belief is a sure sign you're missing something...
I don't steal other people's "intellectual property". I buy all my music, on CD. I buy all my books, on paper. All the software I run is free software, mostly GPL. Fuck you and your assumptions.
I can see why. If you had posted under your own name, that baseless, ill-conceived rant would come back to haunt you.
Everyone knows who I am and that it's me telling you (whoever you are) to go fuck yourself.
--
"Where, where is the town? Now, it's nothing but flowers!"
Where can I find a list of changes in 2.4 vs 2.2
This page should get you started on the significant differences...
Hey, have mercy on us left-coasters! Label posts like that with *SPOILER* ! ;-)
Otherwise, I believe 2.4 drops pretty much straight into woody.
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
Just a clarification, I was speaking for me only as an interested observer rather than somebody who actively works on that aspect of Debian.
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
Officially, I don't know, I don't monitor the devel lists.
However, it's reasonably easy to manually drop in any kernel you want. I'm currently running 2.4.0-test11 on my woody box, and it's running great!
Here's how:
This is explained much more verbosely in the documentation that comes with kernel-package, but I had few problems with the process, and I'm no expert. The process is very smooth. I like it.
What version is currently in Woody?
2.2.17 and 2.2.18pre21, it seems.
--
Wow, before leaving for this year's party, I just finished my dist-upgrade, leaving me with such wonders as a new init, XFree86 4.0.2, and some other new, good things. Just this morning, I also managed to get DRI working with my i810 chipset.
Now, at 2:25am, as Octave compiles (it appears dist-upgrade broke Octave, plus a computing-intensive application ought to be optimized for my system), I download 2.4.0-prerelease. I can't think of a better way to usher in 2001 then with a new mathematical package, a new compiler (minor updates to gcc 2.95.2 packages, I guess), a new windowing system, a new direct rendering infrastructure, and now, to top it all, a new kernel.
If you ask me, life is good. I find it hard to imagine that in the future, such things as having the latest kernel and windowing system will not mattter to me. But now, in my youth, I am content.
Happy New Year. I will continue enjoying the year until spring semester starts up, but that is two weeks away...
I see my kernel is finished downloading, and Octave is about wrapped up... That means time to go.
A new year calls for a new signature.
I don't know how much of a choice anybody has at this point. Linux is a macrokernel, and probably always will be. Sure, if everyone wanted to, it could be turned into a microkernel (and has, in the case of projects like MKLinux and L4/Linux) but it would end up as a single-server microkernel, which isn't really the pinnacle of the design. Traditionally, these single servers have been pretty dissapointing, both in performance (the fastest still aren't as fast as traditional *NIXs) and features (the single server limits a lot of the fault tolerence and distribution capabilities inherent in a microkernel.) Even if it was done, Linux is so far along that the project would slow kernel development by at least 6 months and in the end, I don't think the benifets are there. Also, modularity really doesn't suffer at the hands of a macrokernel. Linux's inability to transparantly use hardware is not the fault of the kernel architecture, but the actual implementation. Even Microsoft makes loadable drivers work pretty well with a monolithic kernel (Win95) and BeOS's ability to automagically detect and configure hardware probably has little to do with its being a microkernel.
A deep unwavering belief is a sure sign you're missing something...
Win2K better.. Haha, that's a good one! The ONLY thing better about Win2000 is the ability to run some more games..
>>>>>>>>>>>>>>
Right. The fact that it is as stable as Linux, that DirectX 8 whips the hell out of anything Linux has, and the fact that the hardware support and OpenGL perfromance of Win2K still beat Linux aren't really important, are they? Not to mention the advantages Win2K has for notebooks (better power management) and all the features that the GUI has that GNOME/KDE are still lacking (they're there, but immature, and not yet pervasively used.)
There is STILL a lot of settings that you can't change without having to reboot the computer. That's crap.
>>>>>>
Who cares? You only configure your computer once in awhile!
For a "production" machine, you shouldn't have to do that.
>>>>>>
Therin lies the rub. You're thinking from a server standpoint.
With Linux, the ONLY thing you have to restart for is a kernel rebuild, which is very rarely necessary.
>>>>>>>>>>
Except to add support for stuff that your distro maker didn't think was important to put in. For stable distros like Slack, that means ALSA, ReiserFS, firewalls, etc. Then you have the fact (that unless you use RedHat or Mandrake, which bring their own set of problems) you cannot upgrade the kernel (realistically) without recompiling!
A deep unwavering belief is a sure sign you're missing something...
Vaporware that really exists? What will "industry analysts" complain about now?? :)
Kindness is the language which the deaf can hear and the blind can see. - Mark Twain
steveha
lf(1): it's like ls(1) but sorts filenames by extension, tersely
I'm rather torped that the 2.4 kernel was classified as 'vaporware' in Wired Mag's 2000 vaporware winners. It would be different if hundreds, if not thousands were actually using the test versions. They also applied this same logic to Mac OS X despite the fact that there are similiar numbers using the public beta.
The next Slashdot story will be ready soon, but subscribers can beat the rush and slashdot the links early!
Correct Translation: The whole concept of "intellectual property" is damaging to humanity as a whole and should be done away with.
--
"Where, where is the town? Now, it's nothing but flowers!"
Linus said there probably wouldn't be a 2.4.0-prerelease2 -- so I personally am looking forward to running 2.4.0-prerelease2-test13-pre7-ac2. ;)
This message is provided under the terms outlined at http://www.bero.org/terms.html
There are proxy servers (like Junkbuster) that will run on Windows. They will allow you to block cookies very easily (and IMO much more flexibly than Mozilla will allow). Of course then again you could just put Mozilla on W2K :)
Better yet are shell scripts that filter out your cookies after they have been saved to disk. That way you can use a cookie-requiring site, and come back the next day with no tracking info.
I had to upgrade to ppp-2.4.0 (from Rawhide, I think I got it) before my modem would work at all, and there still seems to be a problem (with easy but annoying workaround) with Red Hat's ppp-watch setup. Oh, well, I'm only stuck on a modem for another couple weeks.
Of course, if you've upgraded to Red Hat 7 (like I have), you've got 50 other updates you'll want to download, so don't feel too special.
Mirrors? Are you kidding me? Kernel.org has a huge amount of donated bandwidth, I don't think they're going to suffer from the "Slashdot effect" anytime soon. In fact, on their homepage, they say that their current bandwidth utilization is 25.52Mbits, hardly even a third of their available bandwidth.
Friends don't let friends use multiple inheritance.
When will Debian's Woody tree get the new Linux-2.4.0-prerelease kernel? What version is currently in Woody?
cpeterso
Wow! You can see the future!
Maybe you can tell us when we'll be getting our flying cars and personal robots.
--
"Where, where is the town? Now, it's nothing but flowers!"
In the current Linux kernel, there are few constraints on what part of the kernel a module can depend on or what data structures it can modify. And any fault in a kernel module can cause the whole kernel to crash.
Both of those issues make it really difficult for people to add new functionality to the kernel, and I think are significantly responsible for the long release cycles.
A microkernel isn't the only approach to achieving this kind of modularity--using a language with a little bit more error checking and support for interfaces would do the same thing without the overhead of a microkernel design. And adding runtime support for C++, natively-compiled Java, a JIT, Modula-2, Oberon, Modula-3, or any of a number of a number of other languages would be feasible. But my impression is that that would not stand a chance of making it into the distribution even as an optional module.
Can be found on Linux Today here. It includes an explanation of what's changed since 2.4.0-test12, what to expect for the immediate future, and a Happy New Year!
Interested in open source engine management for your Subaru?