Please create an account to participate in the Slashdot moderation system


Forgot your password?
Linux Software

Linus says 2.6 kernel will be out by June 2003 187

Xpilot writes "C|Net reports that Linus Torvalds predicts 2.6 will be out by June next year during a talk on his Geek Cruise. Linus called the next release '2.6', but knowing him that may be just a working title;)" Update: 10/26 17:29 GMT by T : An anonymous reader adds "Rob Landley has published the latest list of features being considered for inclusion" in the new kernel; ... "the long and impressive list is available in more or less human readable form on Linux and Main."
This discussion has been archived. No new comments can be posted.

Linus says 2.6 kernel will be out by June 2003

Comments Filter:
  • by cOdEgUru ( 181536 ) on Saturday October 26, 2002 @01:01PM (#4537077) Homepage Journal
    Or did he just have one too many Margaritas on the Cruise :)
  • by lar3ry ( 10905 ) on Saturday October 26, 2002 @01:02PM (#4537081)
    This will be March, 2004 in "Linux Years."
  • I thought... (Score:3, Insightful)

    by Anonymous Coward on Saturday October 26, 2002 @01:02PM (#4537083)
    ...that Linus was going to call it linux-3.0. Can somebody please stick with an official version number?
    • Either they called it 3.0 and the release date was June 2004, or it was 2.6 and released June 2003...

    • by Anonymous Coward
      He should call it Linux 8.0
      2.6 sounds old, we need to catch up to MSN and AOL
    • by cscx ( 541332 ) on Saturday October 26, 2002 @01:10PM (#4537119) Homepage
      Nah. As with most open-source projects, there is this sickly fear of ones-place-rollover in the version number. That's why you have so many programs with a version number like 0.9999.
      • by dattaway ( 3088 ) on Saturday October 26, 2002 @02:25PM (#4537504) Homepage Journal
        To prevent this dreaded war upon version numbers, a good formula would be something like:


        As your revisions increment, you will be closer to the famed 1.0 release, but never quite there. The press can always ask, "ARE WE THERE YET?" and always be told, "IN A FEW MINUTES!"
      • Nah. As with most open-source projects, there is this sickly fear of ones-place-rollover in the version number. That's why you have so many programs with a version number like 0.9999.

        Actually, that sort of version number is what you would expect if requirements don't change and you are _honest_. Basically you aproximate the perfect version (1.0 final) asimptotically, but never quite make it.
      • It isn't that there's a "fear of ...rollover". It's that open source types that aren't marketing their code have the luxury of making the version numbers actually mean something. Apps can change major version numbers when the file format changes. Libraries when compatibility-breaking ABI changes take place.

        If you have a marketing department, *they* want to jack the major version numbers constantly so that it looks like one "must" upgrade, or because it makes the changes look better.

        Frankly, I'd prefer 2.6 over 3.0. The kernel's performance has been improved, but there's been no rearchitecting. I consider it a bit of a mark of pride.

        Also, people complaining in many of these posts about the number of devel releases before a stable -- be sure that you aren't the *same* people complaining about lack of QA on the stable branch, as this is what it's intended to fix.
    • Well, it will be called 3.0 and it won't be out by June. The re-numbering will be a ceremonial excuse to attone for the feature creep and the resulting lateness in shipping.

      ...well, probably. I mean, I don't have any insider information; I just pay attention to kernel release schedules. I really like the idea of calling it 2.6 because it lets contributers know that it's not supposed to be a miraculous improvement--so it might ship sooner that way.

    • Linux goes to 3.x when it breaks compatability with 2.x.
      • by mbrubeck ( 73587 ) on Saturday October 26, 2002 @04:10PM (#4537957) Homepage
        Linux goes to 3.x when it breaks compatability with 2.x.

        Nope. In this lkml thread [], Linus says:

        We've never had that as any criteria for major numbers in the kernel. Binary compatibility has _never_ been broken as a release policy, only as a "that code is old, and we've given people 5 years to migrate to the new system calls, the old ones are TOAST".

        The only policy for major numbers has always been "major capability changes". 1.0 was "networking is stable and generally usable" (by the standards of that time), while 2.0 was "SMP and true multi-architecture support". My planned point for 3.0 was NuMA support, but while we actually have some of that, the hardware just isn't relevant enough to matter.

    • by mbrubeck ( 73587 )
      Kernel Traffic has a good summary [] of the 2.6 vs 3.0 discussion. In one post, Linus writes:
      I see no real reason to call it 3.0.

      The order-of-magnitude threading improvements might just come closest to being a "new thing", but yeah, I still consider it 2.6.x. We don't have new architectures or other really fundamental stuff. In many ways the jump from 2.2 -> 2.4 was bigger than the 2.4 -> 2.6 thing will be, I suspect.

      But hey, it's just a number. I don't feel that strongly either way. I think version number inflation (can anybody say "distribution makers"?) is a bit silly, and the way the kernel numbering works there is no reason to bump the major number for regular releases.

  • 2.5.xx (Score:5, Funny)

    by NWT ( 540003 ) <tom@s y n t> on Saturday October 26, 2002 @01:03PM (#4537090) Homepage
    If they continue like that, we'll soon have 2.5.100 ... chicks dig fancy kernel numbers.
  • by GreyWolf3000 ( 468618 ) on Saturday October 26, 2002 @01:07PM (#4537106) Journal
    After the feature freeze, if as many people as possible test out the new features and provide bug reports to work on, maybe the impending issues can be fleshed out sooner.

    Unless, of course, Linus decides that there must be a set time between when the features are frozen and when the firse betas hit the servers.

    I'm getting fairly excited about this, even though I don't plan on using any of these new features. Does that mean I read /. too much? ;)

    • I regularly report bugs for certain pieces of software, but the kernel is too big of a beast for me. It is considered uncouth to report bugs which have already been reported, but you have to be kidding me if you think I have the time or perseverance to trudge through megabytes of mailing lists.

      Another problem is that of information gathering. With something like Gaim or XMMS I can accumulate all I need in a few minutes and fire off a bug report, but proper kernel debugging requires time consuming dumps and backtraces. However, since the kernel now officially supports a fairly modern compiler (GCC 2.95.3), one no longer has to downgrade to the stone age to properly debug.

      The 2.5 branch has been infinitely less stable for me compared to 2.3. Out of the twenty or so point releases I've tried, only three have actually booted. All have panicked when I tried to actually do something beyond log in at a prompt. My hardware is far from exotic (and is rock solid under 2.4, just to quell those accusations), so I assume the developers are aware of such showstoppers.

      Now I'm not insinuating the kernel is a crappy piece of software or whatever. In fact, I'm fairly convinced my problems are the fault of Via weirdness, but it's hard to test something which won't even boot properly, and I've run out of patience trying 2.5 builds.

      I guess you could say I'm lazy, but I'd rather do nothing at all than fill lists with halfway done bug reports, and I'm not dedicated enough to delve completely into 2.5's issues.
  • by FearUncertaintyDoubt ( 578295 ) on Saturday October 26, 2002 @01:08PM (#4537110)
    How about "Linux XP" -- eXtended Procrastination
  • 2.6?! (Score:5, Funny)

    by m0i ( 192134 ) on Saturday October 26, 2002 @01:08PM (#4537111) Homepage
    For most users, Linux is around 8.0 anyway :-) Don't ask'em the difference between linux and the packaging around it a.k.a distribution..
    • Re:2.6?! (Score:4, Insightful)

      by garcia ( 6573 ) on Saturday October 26, 2002 @01:15PM (#4537143)
      this really isn't funny. It's an interesting point. Do most people really care that Linux is at version 2.6 or 3.0? No, not really.

      They want to push for 3.0 as a marketing tool, yet most companies that would even consider deploying Linux wouldn't be concerned w/the kernel version #'s. They are going to be concerned w/the distribution version #'s.

      After all, their support is probably going to come from the distribution manu, not IRC or a mailing list.
      • actualy, I would say using 3.0 insted of 2.6 would be for corperate reasons. Look how MS has used the big jumps in version numbering (no .1 or .001 patching). And look to see how much money corperate america spends on them.

        Also, Sun did this, they went from 2.6 to 2.7 and called it 7, and now 8, etc.

        Personaly, what I'd like to see is 1.2.11 (Version.Features.BugFixes) Meaning its version 1 with 2 updates for featers and 11 fixes of bugs. Adding to the Version number would usualy be when breaking previous protocalls/apis/etc or a complete rewrite.
      • Re:2.6?! (Score:3, Funny)

        by asteinberg ( 521580 )
        The thing I'm dreading about Linux hitting 3.0 is all the inevitable jokes posted to Slashdot about how they can't wait for Linux 3.1 or better still 3.11 for Workgroups. It seems to happen every time something closes in on that dreaded number (most recently with Debian and KDE).
      • After all, their support is probably going to come from the distribution manu, not IRC or a mailing list.

        One would like to think. I can already see help support requests going to lkml...
  • When it's ready... (Score:4, Interesting)

    by NewbieSpaz ( 172080 ) <nofx_punkguy@li n u x m> on Saturday October 26, 2002 @01:15PM (#4537139) Homepage
    What ever happened to the saying "When it's ready"? Or is that just a Redhat/Debian specific philos.?
    • by bogie ( 31020 ) on Saturday October 26, 2002 @01:44PM (#4537305) Journal
      I think he feels that way as well. Its just that he has learned that the corporate world likes release dates and the one he "gave" is in all likelyhood a resonable estimate. I see nothing in the GNU/Linux philosophy that states that you can't try to set a schedule and stick to it. Just because Redhat says "when its done" doesn't mean there isn't a giant whiteboard at their headquarters saying a certain date is "D-day".

      Also this isn't some sort of sign of selling out, but I do think if anyone is to guesstimate when a release is likely, they talked to the right person. He is after all the final authority when it comes to releasing kernels.
    • by Gubble ( 619690 )
      Giving people dates to catch is a Good Thing. The assumption is that it will be ready then, if everyone tries to make it ready at that time. I don't think if the kernel is not more-or-less ready, he will release it anyway.
    • by rgmoore ( 133276 ) <> on Saturday October 26, 2002 @02:28PM (#4537522) Homepage

      AFAIK nothing happened to the "when it's ready" philosophy. But saying that it will actually be delieverd when it's ready doesn't preclude trying to estimate when that will be or trying to encourage people to get changes committed instead of procrastinating by giving reasonably hard deadlines. And that's what this is: it's an estimate of when things will be done and a target for developers to tell them when he wants them to be finished. It's certainly not a drop-dead, will be released by this date type of deadline.

      My impression from what I've read is that Linus is pretty happy with the features that have been implimented in the latest version, and that he thinks that most of the things that can actually be included within a reasonable time frame either have been put in already or can be put in by the end of the month. After that he plans a feature freeze, where no new things are added but existing features can still undergo changes, and then a code freeze, where no changes are allowed except for serious bug fixes. I don't see why it's unreasonable for him to give rough estimates of how long those things will take.

    • Random, useless fact: I type in startx entirely with my left hand.

      People get turned on by the weirdest things.
      • Way to karma-whore, my friend ;)
      • by rgmoore ( 133276 )

        It is a vaguely interesting thing. A coworker once mentioned a word that she typed entirely with one hand and speculated about what the longest such word was. Being a geek, I thought that the obvious solution was to write a perl script to find out. I found that the longest words that could be typed entirely with one hand (the left) were 12 letters long; the only one that might actually be used in ordinary usage was stewardesses. ISTR that the longest word typed entirely with the right hand was only 9 letters, but I don't remember any examples.

  • I thougt it was agrreed to mane the new kernel 3.1. 3.1 is better, as everyone knows. As a user, I am disappointed the new kernel will have to wait that long, but who's goung to argue? (That's a nice tie, Linus. Yes Mr. T, I'll wait another 6 months)
  • Huh? (Score:4, Interesting)

    by StarHeart ( 27290 ) on Saturday October 26, 2002 @01:18PM (#4537165)
    I have been hearing a feature freeze for early November. Can it really take 7-8 months to go from feature freeze to a final version? Or is Linus actually planning to make 2.6.0 what 2.X.18+ quality?
    • Re:Huh? (Score:4, Interesting)

      by be-fan ( 61476 ) on Saturday October 26, 2002 @02:13PM (#4537454)
      Yeah, it takes a long time to get something release quality. 2.5.44 is already quite usable...if it compiles at all that is. It's that seemingly last little bit (getting all the old drivers updated to new interfaces, polishing up code, last minute bug fixes) that take a long time. Also, there are certain features (like Reiser4) that'll probably sneak in slightly after the feature freeze, because they don't really touch core code. From Kernel Traffic:

      Elsewhere, someone said they'd love to test these heading-toward-stable kernels, but didn't want to risk trashing their filesystem. They asked how likely that would be, and Linus replied:

      "Personal opinion (and only that): not much chance for a filesystem trashing. There's more chance of something just not _working_ than of disk corruption. Ie you may find that some driver you need doesn't compile because it hasn't been updated to the new world order yet, for example.
      And people still report problems booting, for example, whatever the reason. So make sure you have a working choice in your lilo configuration or whatever. But from what we've seen lately, there really aren't reports of corrupted disks or anything like that that I've seen. Which is obviously not to say that it couldn't happen, but it's not a very likely occurrence.
      That said, I can't set other peoples risk bars for them."
  • by jarrettwold2002 ( 601633 ) on Saturday October 26, 2002 @01:19PM (#4537167)
    I don't know, if RMS keeps going on about Bitkeeper it very well may have to be settled with wrestling.
    I would pay cash to see that. Both of them are perfectly suited for a celebrity deathmatch episode as well. While RMS ponders the definition of free, I can see Linus inviting the crowd to 'share' the 'open' beating of RMS. :)
    • As for Celebrity Deathmatch, Lets make it Linus and Bill Gates as not to confuse the layman watching. Then, At least 20 percent of the viewers would know why this pair was fighting.
  • minor vs. major (Score:3, Interesting)

    by m0i ( 192134 ) on Saturday October 26, 2002 @01:19PM (#4537176) Homepage
    Btw, what would be the killer new stuff in the current devel kernel granting it a major version number upgrade to 3.0 instead of the regular minor to 2.6? They must have a good reason to do so, me thinks.
    • Re:minor vs. major (Score:4, Interesting)

      by iabervon ( 1971 ) on Saturday October 26, 2002 @03:07PM (#4537697) Homepage Journal
      There isn't a killer new feature this time. It's really the large number of major features since 2.X started. They didn't change all that much since 2.4, but there's very little that hasn't changed since 2.0, and it doesn't make sense to never change the major number just because you improve things at a steady rate.

      Or you could say that the number of minor version increases exponentially with respect to the major number, and, since the major number changed after 1.2, it should clearly change after 2.(2^2).

  • Docked in Jamaica. (Score:5, Interesting)

    by Forge ( 2456 ) <> on Saturday October 26, 2002 @01:21PM (#4537189) Homepage Journal
    The cruise docked in Jamaica and everybody had a ball.

    We were told that just a few of the speakers would be presenting in Jamaica so 3 of us drove down to the pier to colect them.


    we neaded all 3 cars plus 2 busses to haul them to "the Ruins". We sat ESR and Linux on a panel with 4 other senior geaks and asked them some lame questions for an hour or so.

    All the baby Linuses were there and Tove is realy cool. everybody seams to think the Coffee here is great (exact words: "The best I have ever tasted") so we will try to have a few bags ready for the next deligation.

    PS: No the Geak Cruise dosn't normaly hold talks on land for the locals. However JaLUG asked nicely :).

    Kevin Forge.
    Jamaica Linux Users Group. JaLUG
    Founding member.
  • by ScottBob ( 244972 ) on Saturday October 26, 2002 @01:33PM (#4537249)
    something to the effect of "Now you can hold your breath until 2.6 is released"?
  • by Xpilot ( 117961 ) on Saturday October 26, 2002 @01:46PM (#4537312) Homepage
    In this interview [] with Robert Love in July, he predicted 18 months before 2.6 gets released(that would make the release early in 2004).

    I'm more inclined to go with Robert Love's estimate considering 2.4's late release.

    Offtopic : Hey, my story submission got accepted! :) Now that's a first.
    • by sagei ( 131421 ) <rlove&rlove,org> on Saturday October 26, 2002 @04:04PM (#4537921) Homepage
      In this interview with Robert Love in July, he predicted 18 months before 2.6 gets released(that would make the release early in 2004).

      I'm more inclined to go with Robert Love's estimate considering 2.4's late release.

      I think I may need to revise that (although I did go on to say a year from then, which would be summer of 2003)... the kernel is remarkably stable at this point and if we can stick to the freeze and get enough testers, I really believe we can have a code freeze in early 2003 and a release not too long thereafter. Five or six months from now seems very doable.

      And I really encourage testers. We need you. Part of the VM debacle in 2.4 was we just did not understand the corner-cases because there were not enough testers testing on enough different machines on enough different workloads. We need to know where catatrophic VM failures are, where areas of high latency exist, and in general where the bad behavior is. This kernel is remarkably better in all aspects than 2.4... it is very smooth. But it needs testers to ease us into a stable release.
  • by Anonymous Coward on Saturday October 26, 2002 @01:58PM (#4537362)
    Ok, I can understand shoveling features into the kernel for fun, but I think that things are getting way beyond what is necessary for the real beach-head -- the desktop.

    I think it is time for a fork. DTLinux and SVLinux. DT for the desktop, SV for servers. I mean really, does Oscar Office Worker really need to hot swap processors? Come ON!.

    This is getting way out of hand, and resources that could be foucssed on the battle for the destkop (BFD (haha)) are being wasted on some sort of kernal probe thing that sounds painfull.

    Seriously, don't you think this kernel feature thing needs to stop!.

    -- ac ah home

    • Well the whole point about the kernel is that you only compile in what you need. If you want a stripped down version, you can easily build it that way.
    • by Hrunting ( 2191 ) on Saturday October 26, 2002 @02:27PM (#4537519) Homepage
      All right, I'll bite.

      That's why you run through the configuration utility before you compile the kernel. You don't need to branch the kernel source to limit features in the kernel. You just don't select them when you compile the kernel. Voila. Your kernel does not have those features. Do you think when Oscar Office Worker got that copy of Windows 2000 Workstation and Mitch MIS Admin got that copy of Windows 2000 Server, they came from different source repositories? I doubt it.

      With that said, the kernel source is getting gigantic, and it would be nice if they released source bundles geared towards those who might be compiling in more desktop-oriented features and those who might be compiling for a server.
    • Since this is easily possible by just turning off (or on) certain options in the kernel configuration file, perhaps a better idea is to make some sort of "sock configurations." So you could do make dep-desktop, and it would use a different configuration file designed specifically for desktops. The same could be done for servers, or whatever else the kernel would be used for.

      It certainly seems a bit more sane than actually forking the kernel.
      • Since this is easily possible by just turning off (or on) certain options in the kernel configuration file, perhaps a better idea is to make some sort of "sock configurations."

        So then if the kernel doesn't compile, at least it can keep your feet warm.
    • Yeah, that's right! Instead of "wasting effort" putting both desktop and server features into the same kernel, we should fork Linux into two different projects! That'll really save a lot of effort! Sheesh. No wonder you're an AC.
    • That's just silly. A lot of work this time around has gone into things like preemption and the O(N) scheduler that improve Linux's responsiveness as a desktop system.
  • by Anonymous Coward on Saturday October 26, 2002 @02:57PM (#4537649)

    That list is just the list of features that are not yet merged and thus need an imminent decision before the feature freeze next Thursday. It's also not especially long or impressive, since these are minor features and a much greater number of patches of that kind are already in. Of the stuff on that list, probably only IPSEC and one of the LVM replacements (needed since LVM1 has been removed) will impact most users, though the crash dumps would also be nice.

    The significant changes in 2.6 will be the new block layer and attendant performance/scalability improvements, the new NPTL thread support, ALSA, and the XFS and JFS merges. See Guillaume Boissiere's list [] for more.

  • by Billly Gates ( 198444 ) on Saturday October 26, 2002 @03:45PM (#4537833) Journal
    Just kidding.

  • The next version of the heart of the Linux operating system is expected by June, project founder and leader Linus Torvalds predicted on Thursday.

    RMS is blowing steam out his ears. Heart of the Linux OS?! AHHHHHHHHHH!!

    "It's OK, Richard, just have a seat and breeeeathe it out. Let it out. That's right... Gooood."
  • by drinkypoo ( 153816 ) <> on Saturday October 26, 2002 @04:43PM (#4538086) Homepage Journal

    7) High resolution timers (George Anzinger, etc.) Home page: Sourceforge download page for this patch: Descriptions of each patch: 557676007653&w=2 557677207693&w=2 558349714128&w=2 Linus had concerns with this one (possibly resolved?): Oct/3463.html

    A must for embedded systems.

    13) MMU-less processor support (Greg Ungerer) Announcement with lots of links: Oct/7027.html

    Makes Linux dramatically more useful (without funky patching) for (again) embedded systems, especially given the coldfire 683xx support.

    11) Kexec, luanch new linux kernel from Linux (Eric W. Biederman) Announcement with links: Oct/6584.html And this thread is just too brazen not to include: Oct/7952.html

    What can I say about this? Another must for embedded systems, and really nice for an enterprise-wide context.

    20) Initramfs (Al Viro) Way back when, Al said: 30/0110.html I THINK this is the most recent patch: And Linus recently made happy noises about the idea: Oct/1110.html

    Need I tell you why this is handy?

    I'll settle for just the above features but the LVM patches seem like they'd be insanely handy, the console rewrite seems like a very good idea, and the non-high-resolution POSIX timers are a good idea, too. Anything POSIX should be a priority since (hopefully) it makes code more willing to compile on more platforms. Provided people actually use the calls correctly.

  • It's perhaps more an aesthetic issue than anything else (after all, it's his, and if Linus decides to call the next Big Release 3.0 or 2.83986 or "Peggy" or anything else, I can't stop him), but I'm happy that so far he seems to be holding out, calling the next one 2.6 however equivocally.

    It goes with the idea of "underpromise and overdeliver" which seems like a smart one to stick to, in software particularly. A series of pleasant, quietly presented surprises is much better than the sour taste of Not Quite What Was Promised. Outside of a minority (those who in particular care about Free software, and in particular the almighty GNU/Linux operating system in some form or another) within a minority (those people who give > a tinker's cuss about computers / computing at all), no one will care about the version number -- but since context matters, so do those people, however few.


  • Harry Potter? (Score:1, Offtopic)

    by MyHair ( 589485 )
    So, Linux 2.6 will come out about the same time as the new Harry Potter book? Maybe it will be a joint release.
  • but what is Alan Cox up to these days?

    Have to say, Linux kernel releases have never really lived up to their word... I seem to remember Linus talking about "release often, release early", but that actually turned into a pile of crap.... we waiting eons for 2.4...

    My advice: "Don't hold you breath for 2.6!!!!"

    • Look, there's a lot of good arguments to *not* release on a more frequent cycle.

      First, stable releases suck for a lot of developers. A lot of people do this in their spare time, and all of a sudden they have a bunch of deadlines.

      Second, feature freezes reduce devel speed, since a lot of developers (who *could* be doing work) have to cool their heels and wait for everyone to stablilize their code.

      Third, there's context switching time. It's a lot of work to release a new stable kernel, and you have to put out this big chunk of work, pretty everything up...and that's time that could be more productively spend working on features to go into the next release.

      What specific features in the new kernel do you need? Tens of thousands of threads? Linus *still* says it's a dumb idea to have more threads than processors, so unless you have a machine with 10K processors, it's not a big deal. Sure, it makes for sexy benchmarks, but they're pointless for real world apps.

      Latency? This is nice, but it's mostly helpful for hard realtime systems. From a user standpoint, the time latency is an issue (assuming you're doing the HZ-redefine, etc stuff that you can do with 2.4) pretty much exclusively when the disk is saturated with requests and part of an app needs to be swapped in or a file read.

      ALSA? I already use it, as do quite a few distros. I know that at least SuSE uses it out of box, and it's pretty easy to build.

      Most of the changes seem to be pretty small (though cute). New driver work, input changes, the ability to use the PC speaker as a microphone...

Beware of Programmers who carry screwdrivers. -- Leonard Brandwein