Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Linux Software

Linux 2.2.1 113

Posted by CmdrTaco from the brown-paper-bag dept.
Myrdraal writes "As the subject says, 2.2.1 is released. It just has bugfixes for the showstoppers in 2.2.0. I'll have a changelist on Cutting Edge Kernels in about 10-15 minutes." Linus announced it as the "Brown Paper Bag" Release. Cute.
This discussion has been archived. No new comments can be posted.

Linux 2.2.1 More

Linux 2.2.1

Comments Filter:
  • Change the poll cmdrtaco! :)

    ---
  • You could always try using the full mirror names... that way you can find one you like that gets the kernel quickly and keep going back there. I've had good luck with ftp4.us.kernel.org. ftp2 also seems to get them early, but only as .bz2, and I haven't got enough RAM in this machine to decompress those in anything like a reasonable amount of time, so I don't go back to that one any more.

    This also solves the ncftp weirdness you get when two hosts with the same alias don't have the same directory structure.
  • Unfortunately it looks like your patch broke. Look for *.rej files around in your source dirs, and if that fails, just download the whole tarball (from a mirror!). After that use patches. Unfortunately sometimes things go wrong.
  • Posted by oNZeNeMo:

    Well, it's a good thing to have a bugfix released a few days after the kernel. At least the kernel people keep up with problems whereas MS will sit for a few months on bugs in their OSs. I can't complain about the frequent updates. How often has Win95 been updated, once every 6 months or so? It's not like MS OSs have been particularly stable. In fact, most of my devel kernels have been far more stable than any windows machine I've put together.
  • Posted by Myrdraal:

    You need to delete arch/i386/lib/checksum.c and arch/i386/lib/checksum.o. You also may want to upgrade patch which should have deleted checksum.c for you.
    -Myrdraal

  • 2.2.0 (Score:1)

    by gavinhall ( 33 )
    Posted by Drizuid:

    There is a bug in 2.2.0 and 2.2.0ac1 (under certain circumstances) 2.2.1 is not affected by this bug. The bug would allow local users to reboot your machine by attempting to run ./core or something, in 2.2.0ac1 it has to have filepermissions 600 to work, get more info on it, @ http://www.genocide2600.com/~tattooman

    I know I'll be patching asap!
  • Can Linus take the bag off his head now?
  • And I thought, gee what if I kicked over the rock here and discovered 2.2.1.

    Well there is already a cd in the oven with 2.2.0, so it will have to do.

    Its a happy upgrade cycle, because its free.

    no worries...



    ^~~^~^^~~^~^~^~^^~^^~^~^~~^^^~^^~~^~~~^~~^~
    ABORTED effort:
    Close all that you have.

  • See http://roadrunner.swansea.linux.or g.uk/clue.html [linux.org.uk]

    If you have a arch/i386/lib/checksum.c remove it. That became arch independant in the 2.2.x series IIRC

  • NcFTP keeps track of the last *real* directory you were in. If you cd to "/pub/linux/kernel", that might just be a symlink to someplace like "/.1/mirrors/kernel", which is what NcFTP ends up saving.

    When you reconnect using one of the mirror hostnames, it's actually picking an IP out of the list randomly, which means the "real" directory structure might differ.

    Mirrors aren't always updated the instant something is released. It could take 5 minutes or an hour or a day. You're not likely to get any sort of decent download rate since everybody and their grandmother is trying to download the latest full kernel (as opposed to, say, the PATCH)... Might as well wait until things quiet down.
  • I've never seen an NT kernel patched the day after it was released. Too bad really, 'cause it needed one.

    -Derek

    P.S. It took me a day to recover from the blue screens of service pack 4!!
  • I'm not sure whether to go with 2.2.1 or 2.2.0ac2 - the ac patches are bigger.
  • The patch for 2.2.0-ac2 is bigger and meatier than the one for 2.2.1.


    Maybe I'll stick with -ac2, unless there's something critical in 2.2.1 that's not in the ac patch.


  • do a little testing before a mojor release?

    I guess their mojor isn't working....hoho...uhh..mojo....mojor....umm...I'll get back to work now...

    dylan_-


    --

  • I'm sorry, it seems we only have Penguin Salad. Fiddlesticks.
  • You don't get it. When your servers are susceptible to DoS attacks you need a fix NOW not months from now. This has nothing to do with being lazy, it has everything to do with people like linus staying up to make sure YOUR isp does not get hacked into.

  • Hm... (Score:1)

    by Rendus ( 2430 )
    What exactly were the problems? 2.2.0 works fine for me, so I'm not touching 2.2.1... I know there was that sound compiling problem... What else?

  • No, it takes Microsoft months to admit there are bugs, more months to investigate them, and then they decide it's more profitable to add cute animations instead of fixing them, and then they charge for it and call it Win 98.
  • Just quit bothering with it and wait until Red Hat, Debian, SuSE, Caldera, etc. release a distribution with it on the CD. Otherwise, quit yer complaining and keep compiling them...
  • Well... I'm stalking around on ftp.us.kernel.org as we speak...


    -rw-r--r-- 1 root root 8654848 Jan 28 21:55 .in.linux-2.2.1.tar.bz2.


    Almost there... I'm sweating... waiting... I can feel the bits flying my way already... anxiously awaiting my next kernel compile...
  • If I'm always compiling

    a) I never have to do anything else, like work

    b) I don't accumulate uptimes, so I don't feel obligated to do things like try to move my computer while running by keeping it on the UPS while riding the elevator down to the car, then plugging it in to the lighter....
  • What was fixed in 2.2.1?
  • I think you've missed the point of OSS completely. Let me fill you in.

    1) The main developer or team put together a release. They hammer the crap out of it as best they can, but they are relatively few, and catch only some of the bugs.

    2) The rest of the community downloads it and uses it. We rarely do testing as such, but there are bazillions of us, so a lot more bugs get flushed out.

    3) The bugs get reported, fixed, and 'round again we go.

    It's very simple and very effective, and we're damn lucky to have primary developers that respond so quickly to bug reports and issue fixed releases with an efficiency unparalleled in industry.

    And you're complaining that the process works so well. Sigh ...
  • 30 seconds.
    ;-)
    C-ya
  • When was the last time you saw a buggy TESTED NT kernel release? When was the last time you saw a NT kernel release? When was the last time bugs were fixed in the NT kernel the next day after they were discovered?

    Open your mind or go away.
  • looks like the sound might be a little easyer to configure for soundblaster now though i havn't tried it yet
  • It's a fairly small patch to fix the problem; check the BUGTRAQ archives (it was posted yesterday).


    After compiling 2.2.0 last night, I'd like to warn of another little caveat.


    If you use Solaris x86 in a partition before your linux partitions, be prepared to reboot to single user mode (or a previous kernel) to fix your system. I'd guess this only applies if you compile in ufs support and x86 partition table support. What you get is some extra partitions, so I have:

    • hda1: Win95 FAT32
    • hda2: Solaris x86 partition
    • hda3: linux
    • hda4: (not used)
    • hda5: Solaris x86 s0
    • hda6: Solaris x86 s1
    • ...
    • hda11: Solaris x86 s7
    • hda12: Linux /
    • hda13 onwards: other linux partitions.

    Obviously, hda5 used to be / for linux; now it isn't so linux wouldn't boot. Took a little work (especially as / was mounted ro..) but it's booting now and working fairly well. Now to get sound working again...
  • Hmm, after watching my friend's computer with NT4 SP3 bluescreen twice, fail to run executables until it was rebooted three times, lock up the whole system five times, and all in the same day, I'm very happy I'm running Linux.
    So you can take your NT super kernel and shove it where the Sun Microsystems doesn't shine.
    I've seen too many blue screens in my lifetime, and I'm only 19.
    And at least have the guts to not be an "Anonymous Coward"
  • For the love of god...
  • We "upgraded" to VFP 6.0 last September, where I work, and yesterday a 16MB service patch (spc2 - there wasn't an spc1) was made available for downloading. It only took 5 MONTHS to attempt to fix some bugs, but the major bugs are still chewing holes in the apps. It also seems that it wasn't enough that M$ forced one to isntall their latest version of Internet Explorer (to read the HTML docs) when VFP 6.0 was initially installed, this service pack installs M$'s version of the Java VM. Just more pushing and shoving of the sheep into the M$ corral for more shearing.
  • Let's see. SP2 out the door one day... _HUGE_ problems found (eg. bye-bye hard drive). It was pulled months ago and they still haven't rereleased it.
  • The original poster has a point. We are the last link in the kernel development cycle. If we don't test and report bugs, they are not going to be caught as quickly. As with any product, OSS or not, *.*.0 is at least a bit iffy. If you can't afford to take any chances, wait a while. If you can then by all means *test* and *report*. Note to the original poster: You don't have to be part of some elite crowd to have a valid opinion. If you run into any problems compiling 2.2.1 let me know; I'd be glad to help. Instructions are in your RH5.2 manual if you have the boxed set.

    -Steve Bergman
    steve@netplus.net
  • When was the last time you saw Microsoft do anything in a day?

    Sure, it takes work to ride the edge, but nobody is forcing you to. This is exactly how OSS improves so just stay at 2.0.x or whatever until things simmer down.

    Massive peer review is good, I'm glad these things get fixed.

  • I use SP3 on my work laptop and it hasn't blue-screened, but it does screw up every few days and require a reboot, and apps frequently lock up and require re-installing (e.g. Outlook yesterday). Even as a workstation OS, NT still has a long way to go - I don't subscribe to the view that it blue-screens all the time, though that can happen on some hardware, but it is not really stable yet.

    Needless to say, my home systems, running Linux, have yet to crash, except when a disk controller blew up...
  • Well try to get a braincell and work out that you don't need to download absolutely every kernel. Get one that works for your hardware and *use* it.
  • Someone is going to have to change 2.2.0 to 2.2.* in the poll now.
  • Fixing for instance a DoS hole doesn't take months of testing. What takes Microsoft so long is that they seem to consider it unworthy to release service-packs with a size less than 10 MB's, and thus they have to gather bugs for a couple of months.

    We kernel developers at least realise that patches are needed as soon as possible. A server just can't stay down a month because it's vulnerable to some exploit that there is an unreleased fix for.
  • Ok, so I'm an idiot and i tried it.

    It will reboot before you get your finger off the key. Ugly......thank you 'who ever the hell wrote it' for fsck
  • OMG!!! 2.2.1 is out!!!! I better flame the hell out of the entire kernel team!

    Sheesh if you whiners would shut up for 1 second you'd realize that the kernel team is doing an EXCELLENT job!

    I think that a release this quickly is awesome and proves that OSS and Linux can kick the rumpus out of any commercially made OS on the planet!

    Linus and the Kernel team... You guys are GODS!
  • That's why theyre there! Don't keep downloading the whole kernel. You're just wasting bandwidth.
  • No, it takes Microsoft months to admit there are bugs, more months to release a patch, and then they charge for it and call it Win 98
  • I tried to compile 2.2.0, but got an error in checksum.c it turns out (from my limited knowledge of c) that one function is declared twice. Anyone know how to fix this....is it in the -ac2 patch or 2.2.1?

    thanks
    -Mark marks@NOSPAM.magpage.com (you know what to do)
  • Thanks a lot! I have to goout for a while, i'll try those when i get home later, (and d/l 2.2.1 tomorrow at skewl). This is why i love /. and linux in general...It's so freakin' dynamic.

    -Mark
  • And I just finished d/l 2.2.0!
  • We should use "Brown Paper Bag" as our themesong for this kernel. (It's a lot cooler than "start me up")
  • In order for opern source projects to work, people need to grab the source as soon as possible and test it. The developers can't do all the coding and testing by themselves, and testing gives the people who haven't ever written a single line of code a chance to participate.

    The more people who grab it and test it, the fewer bugs that there will be in the next release.

    So, wait for a later release if you feel the need, but I say grab it and go with it.
  • They give us good laughs...
  • I don't know. I d/led 2.2.0. I had initial problems compiling and some other problems with a bad util-linux rpm (couldn't log in b/c of incompatible shared libraries) but once I returned to my old version of util linux, compiled the kernel with auto kernel module loading or whatever, I've had nothing but improved speed, etc., and no crashes yet. I upgraded from 2.0.34 and I am definitely pleased with the results. The only reason I'm going to get the .1 patch is so I don't have that pesky ldd core bug. :)

    -Y
  • A little jealous are we?

    And lose the damn d00d speak -- it's freaking annoying.

  • And at least have the guts to not be an "Anonymous Coward"

    Oooh, is that why you use Hotmail? Is your last name really M?

    (NT == no testicles == eunoch != UNIX)

  • > Maybe we should be the ones wearing the brown
    > paper bag... after all **WE** are
    > the ones who failed to identify the bug
    > prior to 2.2.0, eh?

    Heh. I guess so, though by that logic EVERYONE IN THE WORLD should be wearing a paper bag right about now.

    It might be interesting to see all those people bumping into each other, but I guess it wouldn't add that much to the Linux development effort.

    Aw, who am I kidding. Bring on the bags!
  • Even if NT 4.0 did what it was billed to do, and even if you could claim that MS gets it's bug fixes out (I remember an extremly buggy 95 original and 3-5 MONTHS for a service release, NT sevice pack 1 took several months and even at SR4 it only tenitivly makes the C2 rating it claims out of the box). Your forgeting, IT'S FREE, and the people who support it do it out of the goodness of their hearts, for us and The Industry as a whole.
  • So when do we get a jpg of Linus wearing it ;-)?


    ps. I don't think Linus deserves anything but praise and thanks for 2.2. Linux is moving into 1999 with a bang!!!
  • >When was the last time you saw the NT kernel
    >patched a day after release?

    When was the last time you heard of Microsoft being ABLE to fix a bug within 24 hours?

    -ck
  • Q. How many Open Source Software developers does it take to change a lightbulb ?

    A. Maybe SIX BILLION. One to change the lightbulb and then everyone else on the planet can check
    if it was done it right !

    M$ and others should be scared of this.
  • You know, if you would just take a gamble and type your messages straight onto the board, instead of typing them in Word and running the spell check, it might be easier to read. Hehehehehe...

Slashdot Top Deals

"In my opinion, Richard Stallman wouldn't recognise terrorism if it came up and bit him on his Internet." -- Ross M. Greenberg

Close