Slashdot Mirror
Alan Cox says 2.4 Kernel in November
BoNeS writes "Alan Cox, revealed at the first ever UK Linux Expo that the 2.4
kernel of Linux could be available for developers everywhere as soon as November. Cox also revealed that in a couple of weeks he will be working "directly" for the most successful commercial Linux distributors by contracting programmers for their European operations."
I'm aware of that. It would have made the heading a little ungaingly, though.
IMPROVED multi processing AND USB?
Just doesn't work for me.
By the way, the SMP I'm referring to is the massive SMP that SGI is working on. I also expect so see improvements due to the threading improvements. As for USB, isn't it being completely rewritten compared to 2.2 because Linus didn't like the way it was done?
Aaah! I seem to have screwed up. What was in parentheses was also supposed to be indented and in .
The long awaited SCSI rewrite has not happened for Linux 2.4 although a major cleanup effort is underway. Wake me up when you Linux advocates can claim a decent SCSI subsystem.
Erm... I also was there, and wasn't particularly happy with any of the commercial aspect of it.
I did like what Bob Young had to say for himself and RedHat; but the lady I spoke with on the stall had to enlist the help of A Techie(TM) to answer the most brainlessly-phrased question, "so why RedHat as distinct from anything else?".
And the SuSE chap didn't so much answer his own question of "is linux for everyone?" but rather stood there and plugged SuSE sales figures at us. BORING!!
As a matter of fact, SuSE has gone down in my estimation as a result - I'm glad I moved to Debian when I did, and hope it stays as [Ff]ree as ever for years to come (Corel debacles notwithstanding!).
Anyway. I think it's remarkably sloppy and 'overwritten' reporting on the part of ZDnet.
~Tim
--
Rushing on down to the circle of the turn
He's probably referring to the fact that simply loading a new kernel under an old distribution isn't guaranteed to work.
Get 2.4.0 out the door, and make sure it doesn't break glibc and other core libraries. (In practice, they're so critical that they should be running against 2.4pre on the kernel developer's boxes.) Then rebuild the distributions on top of the new kernel and libraries and make sure they're stable.
I *used* to upgrade my tools whenever a new upstream version was released, but I stopped doing that a *long* time ago because you can run into serious problems with seemingly minor dependencies. This is especially true with kernels. I'll start running 2.3 as soon as we hit 2.4-pre status, to help shake out any final bugs, but I probably won't run 2.4 exclusively until around 2.4.4 or so, if the last few major releases are any guide.
coyote-san on soon
Yup... I'm pretty sure I do remember some comment from Linus a while ago about getting fewer development versions per stable, or something to that effect.
The question is more, though, whether the number of features per version is staying the same or not. I suspect it is, and the version numbers are just moving faster than the actual code.
Certainly 13 x 2.2.X + 18 x 2.3.X is not a vast number of things compared to 38 x 2.0.X + 130odd x 2.1.X...
~Tim
--
Rushing on down to the circle of the turn
Oh that's OK.
I was there for the Wednesday bit of it, and I can hardly say Alan was "swamped by adoring fans" and all that crap. It's such bullshit it's unbelievable.
ZDnet going in my 'never again' file...
~Tim
--
Rushing on down to the circle of the turn
So when did Linux become a server-only OS? And who the hell says you have to compile in the drivers? They're there for people who have a use for them, you obviously don't, but I'm sure a heap of desktop users are saying 'why do we need all this networking crap in here...' etc etc.
I thought that the space race was started between the Germans working for the USA and the Germans working for the USSR!
Nobody is going to say that you aren't good at using ideas. Just that you make seriously good use of ones that somebody else came up with.
I think that is what we keep accusing Microsoft of.
BTW - I have been to the US and liked it and am a keen follower of Star Trek. My favourite author was also a US citizen - Asimov.
I'll see your Constitution and raise you a Queen.
The full guess I gave ZD is - code freeze November, 2.4pre December, 2.4 march or so. I know Linus wants to get things moving rapidly on that. But only Linus (and I doubt even Linus) has a totally clear timetable 8)
As to the other stuff thats mostly pretty accurate. Currently I run building #3 which is mostly contracting for Red Hat. With Red Hat europe in place this no longer makes sense. Lest anyone is worried about that I can assure them that part of the paperwork we are putting in place is something both Red Hat as well as I wanted to be sure we had there - which guarantees appropriate degress of autonomy.
Hello to everyone I met both at the show.
Alan
Linux is very User-friendly, VERY!; its just very particular about its choice of friends :P
Best way to forecast the future is to create it. Cray toys won't help much in modelling it either.
That SlotB uni-processor motherboard was NOT madeby AMD. It was/is made by Alpha Processor Inc (API) It is called the UP1000.
It DOES use the AMD K7 core logic chipset.
mike (I work for API)
What's my Karma Mr. Burns? "Excellent"
Hmm, England would be a fairly poor country to do encryption work in. The most recent law they are thinking of passing is a good example of the kind of attitude they have towards it. They're almost as bad as the US.
From what I understand Canada would be a much better place to develop that sort of thing. I do believe that is one of the major reasons OpenBSD is developed there.
Hopefully this will have some features that can be used to fight some of the negative press Linux has been recieving.
I'm increasingly concerned about how the Unix/Linux services are so tightly coupled with the kernel. They really need to be moved into their own security level or run in userland. Linux/Unix bashers have a point in this criticism. Does anyone know if there is a major enhancement to how Unix/Linux handles services in the works?
On the plus side, the Borland announcement of C++ Builder and Delphi for Linux is extremely exciting -- instantly, thousands of fairly spohisticated custom/in house apps will suddenly run on Linux.
I'm still not sure how all well all those Inprise "Enterprise Objects" will come across. I assume Inprise hasn't found any big show stoppers between the way NT and Linux handle security/services. I never got that deep in it, but the NT services/hosts files looked just like the Unix stuff.
Linux is catching up to NT.
No, no, Linus! November 10! Or at least Nov 11 for the Rememberence (Veterans) Day value!
I dont think its too soon.
As soon as the developers and all other ones daring to use an "odd" kernel cannot break it anymore, they cannot do anything else but "release" it.
But a more fine grained release cycle might be advantageous. Bacically because number and diversity of people using linux has increased greatly.
I wonder if this also means a new http://rio.dhs.org/penguin.html
http://lwn.net/1999/1007/a/wwo_2.4.html
I have seen other versions of this; this is allegedly the "final" version.
--
Infuriate left and right
Glad to hear it.
Here ya go:
The Wonderful World of Linux 2.4
I wish we could get a rundown of all the expected features of kernel 2.4. I'd really like a little bit of FUD to hit Redmond. Can you imagine the fear, uncertainty, and doubt the Win2k developers feel knowing the Linux kernel will be released by November with as many features as the Win2k kernel, but with a much smaller memory footprint? They gotta be shakin' in their loafers over there! I'm going to wait to upgrade my distro until 2.4 comes out because I can't wait to kick its tires and see how it works. I just hope Creative releases a driver for their SBLive!
I would guess that he meant that the kernel would be released, but it might take a few weeks for any distro to carry it.
From his position, I don't know whether I am a developer or not. I build applications for end-users, but the kernel is basically still terra-incognita (here there be dragons!). So it won't be available to me until someone like Red Hat, Mandrake, SUSE, Caldera, or Debian puts out a distro.
(I put Debian last because they tend to be very conservative about kernel versions, and I would probably ignore SUSE, as I disapprove of proprietary installers [Caldera said they were opening theirs, I haven't checked back recently though].)
I think we've pushed this "anyone can grow up to be president" thing too far.
Um, being a whiny AC (and standing on a noisy show floor), I didn't actually read the article, but what does it mean "available to developers"? If it's labeled 2.4, it should be a stable kernel for general use, no? :(
BTW, the iMac's keyboard really sucks!
"Same thing we do every night, Pinky. Release a patch for Linux, and use it to Take Over The World!"
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)
This might just be the ultimate argument against the by now well known MS FUD: MS has no roadmap to guide it's development. Yes, MS has such a roadmap although it seems they've fallen off the cliff a few times. The proliferation of all the different Win9X flavors combined with the (8 ?) different version of Win2000 sure don't sound like much of a roadmap to me.
...
So what if we have no roadmap! Again Linux is showing that decentralized development is able to advance an OS (or any other piece of code for that matter) faster than any company could. Add to this the increasing technology 'donations' by companies like SGI and Siemens, and you have a turbocharged development process that makes MSs progress from WinNT4 to NT5 seem pretty glacial.
On a related note: does anybody know if Ultra-DMA 66 will be supported by the new kernels? Right now I have to use my UDMA66 disk as a UDMA33 disk as Linux 2.2 will not recognize UDMA 66 disks
UDMA66 is supported in Andre Hedrick's IDE patches (most of the basics of which are in 2.3.x). Backports to 2.2.x of these patches can be found at your local kernel.org mirror (ftp.xx.kernel.org) under /pub/linux/kernel/people/hedrick/
--
Jeremy Katz
It sounds more like a expansion of Building Number 3.
I guess the redhat IPO allowed him to expand things a bit more. Having decent help is a start and security is always a good thing, so using these developers in better linux encryption is a possibility.
Although I have the latest 2.3.x source tree bar one, I have not played with it as much as I did with 2.1.x, mainly due to a lack of time.
One of the major improvements seems to be the changes with the filesystem caching mechanism. 2.2.x and earlier used two separate caches, one for reads and the other for writes. 2.3.x amalgamates them together into one single cache. Is the performance any better??
The other changes appear to be mainly for hardware I do not have (still using a humble Pentium 100).
I will be applying 2.3.19 this weekend, recompile PPP to use the new kernel, and switch exclusivly to 2.3.x. Heck, most of my system is running seat of the pants releases as it is, one more will not make too much difference.
Point(s) well taken.* Though I reserve the right to hypocritically hold writing professionals (granted, it's debatable whether that applies to ZDNet) to higher standards than I hold myself or acquaintances.
If I were to submit written words that would potentially be read by millions, I'd definitely want everything to be proper.
Hell, I usually hit the "Preview" button here at least three times before submitting.
*Yes, yes. This is another example of me being lazy with the language, and I should have written, "Your points are well taken," or somesuch piffle.
--
Okay, I got Linux installed. So where's the free beer everyone keeps talking about??
Last I heard Debian was debating a code-freeze for sometime in November. Do they wait for the 2.4 kernel or continue with the freeze and have the most stable 2.2 release? Will the 2.4 kernel break lots of stuff like the 2.2 did?
Just wondering aloud,
F.O. Dobbs
No, I'm not talking about the 2.4 or Alan Cox news. That's all great. Great for Linux, great for Alan. Hopefully great for the UK Linux scene.
Ahem . . .
[rant]
Did anyone else play "Count the Typos" in that article? WTF is up with that? Did ZDNet UK give the editors a holiday or something? Here's one particularly offensive bit of text:
Cox was swamped by adoring follows of the Linux hoping perhaps that a little of his coding expertise might rub of on them.
That barely makes sense! It reads like a Babelfish translation! And there were plenty more where that came from.
It's not just this article, either. More and more, the quality of grammar/spelling/editing on web-based news sites are going to crap. A lot of these places also do print versions, which I never really see. Do the typos make it in there, too?
This is offtopic, to be sure, but I just had to vent my spleen. Are my perceptions skewed, or is this journalistic atrocity becoming more common?
*muttermuttermutter....*
[/rant]
--
Okay, I got Linux installed. So where's the free beer everyone keeps talking about??
A good holiday. We won't get to use it for 2.4, but I'd like to see some really hoopy OSS thing come out then, unless one already has and I've missed it.
> Cox was swamped by adoring follows of the Linux hoping perhaps that a little of his coding expertise might rub of on them.
:)
This sentence now reads:
[...], Cox was swamped by adoring Linux followers hoping perhaps that a little of his coding expertise might rub off on them.
Which is a litle better atleast.
Maybe ZDNet reads Slashdot
Some people in this discussion are asking what's new in 2.4. Well, Joe Pranevich seems to be the de facto source for whats new and exciting in kernel releases, so see his The Wonderful World of Linux 2.4 for more info.
Here's info straight from the Debian web site:
Major Changes For "potato":
The major changes we know will be part of potato are the following:
Linux kernel 2.2 based (for architectures with Linux kernels at all).
GNU libc 2.1 based (upwardly binary compatible with GNU libc 2.0)
The egcs version of gcc, also known as egcc, will be the default C compiler.
There is also a good chance that the potato release will include the powerpc and arm architectures.
cpeterso
I thought there where several calls for a 2.1 kernel freeze that didn't hold. Wouldn't a November release date for 2.4 require a 2.3 feature freeze occur around today or the next couple weeks?
Maybe a new version tag is needed, additionally to the the odd numbers, and the "pre" and "ac" tags. Like "ea" for "early adopter". Whatever. Anything above "beta".
;)
:)
The distributions would offer two kernels then: "ea" and "stable as hell".
And "stable as hell" is what "release" should mean.
I'm all for K.I.S.S.
If there was such a thing as a special EA release, why put it in a Linux distribution at all? It's the Kernel, not the OS, so a distro doesn't have to provide packages for all releases! Newbies should stick with what their favorite distro provides pre-packaged, then get into source packages once they want to learn more, and finally try the unstable kernel versions when they feel ready to tackle some serious stuff. To be ready to use the "beta" kernerls, one should have read the appropriate FAQ's and HOWTO's, which means you know about the version numbering scheme and risks. Why change the simple rules and make them more complicated for those who shouldn't mess with the Kernel anyway? If one gets confused by the Kernel version numbers and release phases, do NOT go any further, stick to what your distribution provides! Leave the Kernel development to the developers and get tested packages from your distro's site. You're free to play with the unstable releases, but make sure you know the rules first, it never was intended to be for everyone. Everybody can get involved, but they must learn, or rely on their vendor.
It's that easy.
-- Eavy (: Linux Is Not UniX
Gone are the times when only unix hackers used new Linux releases.
This means: A kernel officially released as "Release" should be VERY stable. Stable enough for anyone to put it on his most important servers, without a second thought.
Remember the problems the 2.2 series went through. Even 2.2.12 is not yet completely there. 2.2.13 or 14 might effectively be.
Maybe a new version tag is needed, additionally to the the odd numbers, and the "pre" and "ac" tags.
Like "ea" for "early adopter". Whatever. Anything above "beta".
The distributions would offer two kernels then: "ea" and "stable as hell".
And "stable as hell" is what "release" should mean.
PS: Potential deficiencies in the NT release versions are not really of interest here. Linux can do better. The people in control of the kernel dont need to care about public company quarter results.
As much as I hate to say it, I think they are rushing it out the door too damned soon. While I invite the changes that ARE happening, I don't think that there has been enough of a breakin time for the 2.3 kernel to goto 2.4 already. There are not NEARLY as many users of 2.3 as there was 2.1.
I truely hope they are not trying to get this out the door to give the appearence of rapid development. It's be wrong, and come back to bite the entire community in the rear..
-- I'm the root of all that's evil, but you can call me cookie..
This doc should answer your questions and more:
Wonderful World of Linux 2.4
LILO boot: linux init=/usr/bin/emacs
Well, based on the track record of the 2.1.x to 2.2.0 release, this doesn't look too promising. Linus promised October 1998 then, and the kernel came out in January 1999. Alan says November 1999 now, so I'm looking for late December 99 or sometime in January 2000.
Also, I hope the stability is a hell of a lot better. Something labeled the "stable kernel" should be just that - stable. 2.2.12 isn't even really stable yet, though 2.2.13 is promised to be good (and Alan Cox's pre-13 stuff seems to solve most of the major problems). IMHO 2.2.12 should've been 2.2.0, and the previous kernels should've been part of the development cycle - they were not stable kernels by any stretch of the imagination, so should not have been falsely called that.
If we continue to release things before they're ready, people will be scared off, and everybody will still be using 2.0.38 forever (many many people still do) - do we really want that?
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Umm, that's all fine and dandy, but when did Alan Cox get the kernel version blessing wand (tm)?
"for developers"? Umm, that's what the current 2.3 series is there for. Let's not start down the 2.4pre1 road please.
(btw, as always, apply the mass media filter.)
Apart from Alan that is. I thought the RedHat stand was good, actually having interesting people there but generally the Expo was small and not very well attended. The big corporate booths (esp. IBM and Intel) had little to do with Linux (IBM man said to me "Linux is not enterprise ready).
Nice looking slot B Alpha 21264 mobo maid entirely by AMD for launch in November, looking not disimilar to the Athlon mobos.
State of UK Linux doesnt seem that healthy if this is all thats going on (which of course it isnt...)
it would be a great present to my birthday. linus, please try to release it on the 28th of november :-)
Dammit, a roadmap is a document put out for marketing purposes only. The product delivered some time in the future usually has little or nothing to do with the roadmap. Remember how Cairo was going to be this big open object soup? .
Don't fall into the trap of playing on their turf. Linux needs goals for technical reasons, like Merced support, better SMP, Journaling. But it also is Open Source which means that you CAN'T roadmap it. People are morphing it all the time into amazing applications like TiVo. Each time this happens it adds to Linux 'stone soup' in a new and unpredictable way. YOU CANNOT PROGRAM OR PREDICT CREATIVITY. It just can't be done.
Microsoft FUD is FUD, and that's it. Not worth the photons needed to carry the image to your eyeballs.