Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
Linux Software

Linux Kernel 2.4.5 Released 115

John Jasen writes: "Join the kernel of the month club! Order yours now!" See the Changelog, I would link to the mirrors but I doubt they're updated yet, so just head to kernel.org.
This discussion has been archived. No new comments can be posted.

Linux Kernel 2.4.5 Released

Comments Filter:
  • by Anonymous Coward
    Does anyone know if USB is working yet ? I keep getting the infamous "USB not accepting new address" error (happens right when you plug in a USB device). I found plenty of posts in newgroups stating the same problem but have not seen a reply yet. I want USB bad............
  • by Anonymous Coward
    MacOS X is NOT UNIX

    MNU?

  • by Anonymous Coward
    Veritas is doing a Linux port, but it
    is a closed-source module. It is not done
    yet either.

    In fine Linux tradition, we write our own
    damn VxFS driver. Why wait for a binary
    when you can have source?

  • by Anonymous Coward
    You mean that the transition to the FreeBSD VM system is finally complete?
  • by Anonymous Coward
    Admittedly the last one I've tried personally was 2.4.4-ac8, which started killing pieces of KDE when I was simulataneously recompiling the kernel and Qt!

    Well, he does work for Redhat you know ;)

  • Upgrade to Slackware-current (pre-7.2) and 2.4.4 on the kernel.

    --
    WolfSkunks for a better Linux Kernel
    $Stalag99{"URL"}="http://stalag99.keenspace.com";
  • Perhaps now is the time to ask if anyone has had the same problem as myself. I'm not sure if it's related to the kernel, to X, or to software that runs under X, but there has been a serious problem with my system for around four months.

    So I fire up X, and use it for a while. Within a short amount of time, the screen blanker starts to come on only in X within one second. This usually happens after heavy disk usage, perhaps after a run of mkisofs to create a CD image that I am about to burn. So I say what the hell, and fire up a terminal window and type 'xset off' to kill the blanker.

    I decide to fire up a timer based video game. Quake 3 will suffice for an example. I enter a map, and the sky is moving very quickly. I fire a rocket and the explosion blinks at the end of the hallway immediately.

    I have written code in the past with SDL [libsdl.org]'s cross platform timer, and however they implement it in Linux is also broken, because code that I have written is no longer timed correctly.

    The timing for my machine is broken. The only way I can fix this is to reboot. My system clock is running at a normal speed, and this doesn't happen in Windows 2000. This is a bug in the GNU/Linux system.

    I'm going to try the new kernel version, but after reading the changelog, I don't have much hope for 2.4.5.

    Has anyone else experienced this? It's VERY troubling.

    X -version returns 4.0.2, though I have tested 4.0 and 4.0.1.

    uname -r returns 2.4.4, though I have tried every kernel since 2.4.0.

    /lib/libc.so.6 returns version 2.1.2 compiled by egcs.

    I have a Quadro 2, however I have also tested a TNT2 Ultra.

    hdparm is set to DMA on on my IDE drive.

    I have an AMD/750 with a VIA chipset.

    I'm using the NVidia binary-only drivers, though the problem persists with the stock NVidia drivers that come with XFree (tried 4.0 and 4.0.2)

    My distro is Slackware 7.0.

  • All the major distros are now including openssl/openssh standard (Red Hat, Mandrake, etc.). Why not include the full crypto support for loop devices and the like? Make it an option in the setup to create secure, passphrase mounted filesystesm using blowfish, AES, IDEA, cast128, etc.?
  • As of Solaris 7, you can remount your volumes
    with the "logging" parameter.

    Solaris also comes with volume management built in.

    -fialar
    Solaris Administrator
  • Linux 2.4.5 and Freeswan 1.9 is broken. If you want to use the Freeswan patch, either:

    1. go back to linux 2.4.3 (2.4.4 has serious problems)
    2. wait for an official freeswan update, http://www.freeswan.org/ [freeswan.org]
    3. Try the bleeding-edge snapshots from freeswan.

    (BTW, Freeswan adds IPSec to Linux)

  • Service Packs, I believe, contain fixes for a ton of different programs and such. Since this is exclusively the kernel that is getting updated, it's not really a "pack".

    The point about not changing things that work is still valid, although in this case, it's only changing a single thing; it won't make all of your software behave differently.
  • Non-executable stacks can still be exploited, but the exploits that are easy to write and work on most machine won't work. While it does give you a false sense of security, it means that someone searching for a machine to exploit will probably move on. Of course, if everyone used it, attackers would get around it, but it's helpful for now.
  • by Tet ( 2721 )
    Not that I've had much trouble with the earlier 2.4 kernels on the whole, but I wouldn't run my server farm on them.

    I would. For me, the turning point was 2.4.4. We had a few stability issues with earlier 2.4 kernels, but 2.4.4 has been rock solid. That's not to say I'd rush out and upgrade a perfectly functioning 2.2 server farm, but if I was building one from scratch, it'd be 2.4 all the way.

  • Erm, I'm using an Epson 1240U USB scanner (and a USB mouse, for that matter) on my FreeBSD box. I'm also printing to a Winprinter (Epson Color Stylus 440). All this on a 4.3-STABLE installation, meaning that I'm running a standard, non-developmental system.

    I assure you that FreeBSD's hardware support is not nearly as dire as you might think.

  • by stripes ( 3681 ) on Saturday May 26, 2001 @04:38AM (#197511) Homepage Journal
    The Real Time Scheduler does not really make Linux an RTOS because in and of itself it does not provide kernel pre-emption - the ability for the kernel to interrupt kernel-space code to deal with incoming events that _must_ be processed. This is a requirement of a 'proper' hard-RTOS because such an OS must be able to guarantee a response time, and if it cannot interrupt kernel code the OS scheduler may be stuck waiting for kernel code to return before it can go on to deal with the input.

    As a minor nit real time kernels do not require kernel pre-emption. The require a bounded maximum interrupt latency time. In theory the bound can even be high (100+ms, or hours even). In practice the bound has to be low just like you said. In practice kernel pre-emption is the simplest way to do it (one could also use a true micro kernel that only passes messages, and does that really fast, making all of the "real work" done in premptable user level code).

    There is also the difference between hard and soft real time. Soft real time like a video game can't handle going above the stated latency very much or the animation will stutter and the user will become displeased and play a different game, but it can handle once in a while blowing the stated latency. Hard real time can't handle missing the promised latency, a computerized fuel injector might be a good example of this. If it misses, even just once it could inject fuel at the wrong time, and might blow out a delicate gasket and cause $1000 of damage to your engine (this may also be a bad example, I'm not sure how tight the timings really are for CFJ).

  • You might want to consider FreeBSD then. FreeBSD fans have long touted the superior NFS implementation. Not sure if it's still the case -- there were lots of improvements in 2.4, NFS being one of them.
    ___
  • by otis wildflower ( 4889 ) on Friday May 25, 2001 @07:25PM (#197513) Homepage
    (2) Disksuite is a free Volume Manager that does various levels of RAID.

    However, it requires a lot more legwork.

    You pretty much have to slice and dice your HDDs identically (and have identical HDDs in the normal case). Thus, you are still limited to 7 partitions (minus some for metadbs of course) within a "volume".

    Also, you only get concatenation when you want to increase filesystem sizes, and fairly dumb concatenation at that. And IIRC if you want to concat you have to take the filesystem offline.

    Disksuite is nice for small systems and root/boot/swap mirrors. Much nicer IMHO than setting up similar service (converting a single disk R/B/S system into a mirrored one) in Linux using md. I just did both on separate boxes in the last week, and I am still cringing from the md mirror "procedure" (though it did remind me that I actually don't suck ;)

    A true LVM beats DS up and down the square. Many flavors of Unix come standard with LVM for "free" (though you usually have to license the OS, and Sun now beers it away up to 8 CPUs) and IMHO it's about time for Sun to give it away as well, whether they license Veritas or port/write another solution.

    ps: when you've got your kernel installed into the boot sector and you've gotten your / to start mirroring (by in my case booting from rescue cdrom, copying over the /dev/md instances into the ramdisk /dev, insmodding md, raid1 and reiser, mounting the partitions to mirror, chrooting to that mntpoint, editing lilo.conf (btw that's the LATEST lilo with md bootsector support) and /sbin/lilo), BE SURE to specify your md=X,/dev/hdeX,/dev/hdgX for your root drive in your kernel append for your mdX mirror label.

    It's days like that you don't feel overpaid. ;)

    Your Working Boy,
    - Otis (GAIM: OtisWild)
  • by Taco Cowboy ( 5327 ) on Friday May 25, 2001 @05:20PM (#197514) Journal


    FYI -

    The mirrors are updated !

    I have tried the mirrors at .no (norway), and
    version 2.4.5 is now available at
    ftp://ftp.no.kernel.org/pub/linux/kernel/v2.4/

  • I'm stuck on this G4 PowerBook. If you had installed YD or SuSE PowerPC version you would not have been is such a predicament ;-)
  • Even if they're not up the instant you type, they might well be a minute later, and certainly will be by the time most people get to read your article later in the day. It would help kernel.org to not get slashdotted.

    This ``probably not updated'' rubbish just doesn't cut it, either. How long would it take you to check? In seconds? Whatever happened to responsible reporting, the kind so often bemoaned on his very site for its lack?
  • I would link to the mirrors but I doubt they're updated yet, so just head to kernel.org.

    That's why you open a new browser window, and take a look! This is what we call "Investigative Journalism".

    Sheesh. Not to sound like an asshole, but is it really too difficult to check a few mirrors, and say "The mirrors we checked have/have not been updated. Your mileage may vary."? Better yet, don't say anything at all and just link to the mirrors!

    The point of the mirrors is to offload traffic from kernel.org. Slashdotting it defeats that purpose.

    -- Give him Head? Be a Beacon?

  • The changelog speaks about VxFS... Has it been ported or something? Not that I really care about VxFS too much anymore, but it would be funny to see VERITAS start making commitments to porting stuff to Linux. How would such an implementation take shape? Since it'll probably be a bit hard to convert from ext2fs/xfs/reiserfs to VxFS, I'm thinking that they'll have to do a ''Red Hat a'la SGI'' installer... Closed Source (tm) of course with lots of expensive licenses to buy, but that's another story.

    What do you think?
  • doh! I forgot MOSIX [mosix.com]! MOSIX is a clustering thing! It looks cool.
  • Here are some patches for 2.4.X I find essential
    • Alan Cox's Patches [kernel.org] - Nice!
    • Real Time Scheduler [sourceforge.net] - Aside from making Linux a RTOS, it improves app performance!
    • GetRewted [getrewted.net] - Similar to the Openwall [openwall.com] pacthes for 2.2.X - NonExec stack, improved filesystem security, stealth networking, Trusted Path Execution
    My personal box runs all but GetRewted. My server will run them all very soon. Enjoy!
  • I think there's a couple differences between this pattern of kernel releases and MS service packs.

    These don't seem to break programs that previously worked - at least the ones I've used.

    2.4.4 on my machine broke the Realtek 8139 card for DHCP - it would freeze for a few minutes and then start without network. Reverting to 2.4.3 fixed the problem.

    I believe that this may be fixed in 2.4.5 from reading the changelog
    -pre2: - Jeff Garzik: 8139too net drvr fix

  • What's with the troll moderation?

    Anyway, at least some of the ftp.us.kernel.org servers have the update at this point. If the one you try doesn't have it, try another.

    -----

  • However, the source files are still owned by non-root users :)

    If you're extracting as non-root, don't be surprised ;) If you're extracting as root, try the --no-same-owner option.

    -----

  • Where did he say that? Last I heard he used 3.2 in (at least one of) his machines because of sb-optimal 2.4 vm...

    Also, the ac-series is different from Linus' tree; the diff between 2.4.4ac17 and 2.4.5 is many megs. (Incidentally 2.4.2 OOM rambo killed my diff process when cache filled up (230 megs cache), so I can't give you exact figures) Alan includes much more experimental stuff, but also has some other differences in his tree. Merging in between does happen piece by piece.
  • 900K vs 14M sure makes sense if you're on a modem as many of us still are, as well as not screwing kernel.org on their bandwidth costs.

    Get kernel patches here: http://www.bzimage.org/ [bzimage.org]
  • Really? I've been following this on lkml and up to yesterday it seemed people are still reporting VM problems under heavy swap.

    Admittedly the last one I've tried personally was 2.4.4-ac8, which started killing pieces of KDE when I was simulataneously recompiling the kernel and Qt!
  • Well, several kernel updates (like 2.2.19) are heavily suggested as upgrades -- 'cause they fix security holes. Anyone running a 2.2.x less that .19 is crazy if it is in anyway connected to the 'net.

    The big difference is that MS usually holds SPs and you get about 1 every 6 months to a year (sometimes longer) with an occasional hotfix for serious ("highly publicized") problems. Point releases for the Linux kernel usually are out much quicker and don't encompass as much.

    --
    Charles E. Hill
  • There are still many places in the world where having/using crypto is against the law. In some cases, it is a capital offense.

    Many of these places (developing nations) are places Linux would do well in. Including crypto as standard would rule a lot of that out.

    Also, there are still a few hangups in the U.S. about the distribution of crypto. They might not have all the paperwork ironed out, yet. (They might not want to go through the hassle.)


    --
    Charles E. Hill
  • Frequently there are updated/new device drivers. In the cases of .3, .4 and .5 the big one would be lots of additions/changes/fixes to various things USB.

    Aside from that, if you use any of the hardware that has been updated/changed (I have an AIC7xxx controller) or use ReiserFS (there are several minor patches/updates/fixes to this).

    But you are right -- if it ain't broke, don't fix it. I can make strong arguments for using 2.2.19 instead. On several machines at my last job they are still running (happily) 2.2.x kernels. [I suggested to the new SysAdmin to update to 2.2.19 due to security.] I wouldn't recommend an upgrade beyond that 'cause they don't need anything or gain significant advantage with 2.4.x.

    --
    Charles E. Hill
  • The help to VxFS talks about FreeVxFS a veritas free implementation - currently read only, but ya never know...
    With XFS/JFS just becoming usable, I dont know if they'll have much success...
  • by throx ( 42621 ) on Friday May 25, 2001 @06:29PM (#197532) Homepage
    Once the dev kernel gets forked off, the kernel releases become much more like service packs. I know if you are running anything less than 2.2.16, most people will suggest you upgrade.

    In the early numbers however, it is probably worthwhile upgrading now and again to get rid of those bugs that surface up in the major version change.
  • If you're not using *at least* SP4 on NT4, I think people'd roll their eyes at you and you'd get a round of suggestions to apply the latest SP
    And if you already have SP6 you'd get a round of suggestions to apply the latest SP *again*!
  • What about the user-space NFS daemon? (Or, is it the NFS client that's broken?)
    ------
  • I know this won't fix the problem, but it does render it moot: try installing ntpd and ntpdate. (NTP - Network Time Protocol).
    ------
  • Could it be a weird interaction between Linux and a Y2K bug?
    ------
  • Like ntpdate?
    ------
  • Well, if the mirrors haven't had time to update, lets make sure we give the main source a good slashdottin! I'm sure *that'll* help them catchup...blah.
  • by treke ( 62626 ) on Friday May 25, 2001 @06:33PM (#197539)
    get the patch for the 2.4.4 source? It's only about 900k
  • Both the tar.gz and the tar.bz2 files from kernel.org are broken - all the files extract being owned by non-root users and don't have the appropriate permissions (scripts and stuff aren't set +x, etc).

    Is anyone else having these problems? Or is my copy of tar screwed? Or am I screwed? :)
  • Well, the wrong perms turned out to be because of a weird umask that was set on the box, that was stopping the newly compiled binaries from getting +x.

    However, the source files are still owned by non-root users :)
  • Actually, it's not wholly just me :)

    Other people have seen that the source extracts as owned by weird UIDs.
  • Apart from the fact that a kernel release or patch is inherently different from MSFT's "service packs", I despise the fact and wonder why so may unix/linux users nowadays want to take over MSFT invented terminology.

    The UNIX community has had its own slang for 25 years. I can think of no reason to submit to MSFT "culture" in adopting words like "service pack". Newcomers (many from the MSFT world) to LINUX should adapt to an older and richer culture; maybe take a look to the jargon file [tuxedo.org].

  • X Windows? O horror, the old mistake, also induced by MSFT contamination.

    It is X window (without s!!!, and it is short for the X Window System). See, for example, this news article [google.com] (there are many many more).

    To the other person w.r.t. what the difference between a service pack and a kernel patch: The patch simply takes your kernel up to the new level. In fact it is just a mechanism to get to the next (full) release.

    The service pack doesn't take you to a new full release (you'll still see build 1395 or whatever while booting), when you buy NT 4 now you'll still get the original and have to apply service packs, hotfixes and the like. It is something completely different. Another big difference is that a service pack is a kind of permanent patch to the whole system, not only to the kernel.

  • and I just installed 2.4.4.. another 2 hour download.. I hate dialup. :(
  • Frequently there are updated/new device drivers. In the cases of .3, .4 and .5 the big one would be lots of additions/changes/fixes to various things USB.

    One USB fix that I really appreciated: devmode, devuid and devgid mount options for usbdevfs had no effect in 2.4.4 (I hadn't used previous 2.4 releases, so I don't know about them). They work again in 2.4.5, and I'm very happy that I no longer need to su to get the photos out of my camera.

    Thank you kernel developers. You rock!

  • Where can we actually find out what has changed? This is always a problem with each new kernel version. The changelog includes information about everything that was changed, but says nothing about what the change was. The diff is too detailed to read and could take days to understand.
  • I guess they post it because otherwise they'd get clobbered with 10001 "hey, didn't you see 2.4.5 is out yet?" submissions.

    Correct?

  • by FattMattP ( 86246 ) on Friday May 25, 2001 @06:09PM (#197549) Homepage
    Alan Cox: further merging
    I'm glad to see that we're still working hard to merge Alan Cox directly into the kernel. And news of this right after the AI post, too!
  • ...I'm all for "it takes a villiage to raise an OS" or whatever, but am I the only one here who thinks "GetRewted" is a bit of a sketchy name for a security patch? The name isn't exactly inspiring a feeling of security (or credibility) in me. Kind of like buying a home security system from "GetBrokenInto", or soup from "GetBotulism". My general theory of security is "AvoidGettingRewtedAtAllReasonableCost". Takes different strokes to move the world. That's what the source is for, anyway, I guess...give it something of a read-through before piping it through patch.

    Laugh, folks...this is supposed to be funny.
  • ...the software is written and the site is run by some high school kid, as far as I can tell from reading the news items.
  • What's with the troll moderation?

    Don't ask me. I found I had some mod points and gave in an Informative, as the norway mirror does indeed have 2.4.5 up.

    I guess the moderator just thought that since the post was contradicting one of the editors of this fine site it must be trolling. :)

  • It's just you. :-)

    [pts/0:saint-ides] [/home/mikedotd] > uname -a Linux saint-ides.geekpitlabs.com 2.4.5 #1 Sat May 26 00:19:01 EDT 2001 i686 unknown



  • I can think of no reason to submit to MSFT "culture" in adopting words like "service pack".



    Yeah, I was annoyed when popular lexicon stole the Windows out of my usage of the term, as in:
    "X Windows"
    or, when nontechnical folks, finding I was a programmer, would ask me if I programmed in "VisualSeePlusPlus" instead of:
    "C++"

    The best defense being a good offense, I propose a counter attack!

    Volley One: The next Linux kernel release will not be named "2.4.6".

    Rather, it shall be named

    Linux 2002 Professional Enterprise Enhanced.NET Business Productivity Extreme Suite XP Next Generation for Large Data Centers Whose Budgets Are Controlled by PHBs with Unenviable Small Phalli®

    That should hit home with the target market...no?

  • Veritas is working on a version for linux. They're in beta right now. I actually love veritas. It's practically required if you want to run a file server on a Sun box. Sun, in their infinite wisdom, hasn't gotten around to including a journaling file system or logical volume management into their OS. Check out http://beta.veritas.com
  • You're right, of course. UFS does do journaling, but it's slow as butt. Solstice disksuite doesn't compare to Veritas' volume management though. Either in terms of ease of use or features. But what do I know, I just run file servers for a living. By the way, I currently prefer Sun/Veritas for the task over Linux. That may not be the case in a few more months, but the tools aren't there yet for Linux.
  • I'm going to be setting up a smallish (~300 gigs) linux fileserver for testing. I'll do some benchmarking (with Reiserfs and XFS) and see how it looks. Since my company has lately restricted my choices for hardware and software (no more sun :-( ), I need to find a decent alternative. Since I hate AIX (the rs6000 hardware is nice, though), I'm pretty much stuck with using linux. When the hell is somebody going to put some work into NFS to get it up to par?
  • Must be something broken on your box. As I write this, I am *running* 2.4.5 with no issues or problems. Perhaps you might email the errors you're getting to me? I might be able to help.

  • Our mirror here at the Dalarna University in Sweden is also updated... Come get it!

    ftp://ftp.du.se/pub/linux/kernel/v2.4 [ftp.du.se] or
    ftp://ftp2.se.kernel.org/pub/linux/kernel/v2.4 [kernel.org]

  • <i>2.4.4 also has the patch for the iptables hole included</i><br>
    <br>
    Is this the one for ftp servers? I seem to recall something like that a while back, but I don't run an ftp server so it didn't sound like it affected me.<br>
  • Apart from the fact that a kernel release or patch is inherently different from MSFT's "service packs"

    Could you clarify how it's different? You're still taking a common code base, applying a patch that fixes problems, introduces new problems, and adds new features. Sure you're working with source code instead of binaries, but the process and outcome are still the same. If you don't believe me that it introduces problems, perhaps you should read some of the previous posts in this thread.

    I just don't want the Linux community to be disillusioned into thinking that their new kernel releases are any different than service packs, because they aren't.
  • Linux 2.4 Service Pack 5. I'm running Service Pack 2 just fine and I haven't really seen a reason to apply the latest Service Pack as soon as it comes out, unless the changelog mentioned a significant security fix. Otherwise, if it's not broke, don't fix it.
  • As of 2.4.4 it's sort-of working for me. I can use some devices fine, others hang the kernel occasionally. I've never had the error message you describe. The bug probably resides in the host controller code (I use UHCI).
  • Well, i've experienced quite a number of kernel hangups when using various USB hardware. Apparently the USB support isn't quite stable yet; not much of an issue for servers, but it illustrates that there are still serious bugs in the 2.4 kernel, and some may also still be hiding in server-critical areas.
  • Hmm...

    "[kernel releases] don't seem to break programs that previously worked"

    ...and...

    "the attitude I see from many is 'if it's working, don't touch it'."

    ...seem contradictory. If people were sure that kernel releases wouldn't break their stuff, they would be much less cautious about upgrading. The fact that people are reluctant to apply minor updates in security or stability implies they're worried something will break.

    -Erik

  • A search on Freshmeat for "debian reiser" reveals 2 projects (actually, I already knew they were there :)

    These are not, however, with the 2.4 kernel, but the important part is that you can do the installation with Reiserfs as your filesystem. It's much easier to compile a new kernel (I'd be doing that anyway) than to deal with converting from ext2 to Reiser.

    Sotto la panca, la capra crepa
  • or M-XINU? (hint: read backwards)
  • linux-2.4.4-ac10 ? Now an anonymous coward can add code to the kernel ?

    AC -10 AWARD [rareware.com] is also something you get if you make heavy use of Body Armor in Rare's Goldeneye 007 for N64.

  • You can't miss this! Go check it out right now!

    *Sigh*. How long are we going to have to read kernel = kernel + 0.0.1 just released stories? What is the relevance of this, truly? This thing shouldn't even be at freshmeat [freshmeat.net], for christ's sake.

    If not, then I want daily CVS announcements. Please, either completely bore me, or do not bore me at all.
  • OOPS! Discovered while rebooting that unmounting a reiserfs partition causes a panic! There is a fix available, search on the mailing list for the patch, seems to have worked here.
  • VFS layer cleanups, USB enhancements, oh my! And the march to rock-solid stability continues.

    Not that I've had much trouble with the earlier 2.4 kernels on the whole, but I wouldn't run my server farm on them. Soon, though, from the look of things.

    Keep up the great work.
  • yah, duh..it's un*x
  • by marm ( 144733 ) on Friday May 25, 2001 @07:08PM (#197573)

    The Real Time Scheduler does not really make Linux an RTOS because in and of itself it does not provide kernel pre-emption - the ability for the kernel to interrupt kernel-space code to deal with incoming events that _must_ be processed. This is a requirement of a 'proper' hard-RTOS because such an OS must be able to guarantee a response time, and if it cannot interrupt kernel code the OS scheduler may be stuck waiting for kernel code to return before it can go on to deal with the input. The rtsched patches do appear to integrate with MontaVista's kernel pre-emption patches however, and together they would indeed form a proper hard real-time OS.

    Kernel pre-emption does not come without a price though - it can make a significant dent in overall performance, and it is tricky to implement in a clean way, and this is why kernel pre-emption will probably stay out of the mainstream kernel for the forseeable future. It also isn't necessary for 99.9% of people, who, as long as the latency, the time to respond, is on average less than a few ms, are happy. This is called 'soft' real-time and is more than adequate for any video or audio work.

    Linux is actually pretty bad at soft real-time as standard, with typical latencies around the 100ms mark, which is rather worse than any version of Windows 9x or NT, and a lot lot worse than BeOS, which has latencies in the sub-5ms realm. Andrew Morton's Low-Latency patches [uow.edu.au] deal with this quite nicely, taking typical latencies down to the 1.5ms mark by improving various kernel algorithms and adding a few points where the kernel can reschedule itself during long periods in kernel-space code. This represents the best latency in just about any OS that does not do hard real-time with kernel pre-emption (QNX, vxWorks etc.) and does not hit performance in the way that pre-emption patches do.
    What would be very interesting is to combine the low-latency patch with the improved scheduler in the rtsched patches...

    As for GetRewted patches... well, I'm not entirely convinced about the value of a non-executable stack. The problem is whether they actually do any useful good - they give a warm fuzzy feeling of security while only actually preventing a limited subset of attacks. In addition, it's in the wrong place. It's a kernel-space fix for what is really a user-space problem - and certainly I think it's better to fix problems at source than patch them up elsewhere - otherwise you end up with code spaghetti.

    My own personal favourite anti-stack-smashing add-on is libsafe [avayalabs.com], originally a Bell Labs project, which overrides dangerous libc functions with its own, safe functions, either by using the LD_PRELOAD feature of ELF shared objects to protect existing binaries, or by being linked in to a binary at compile time, preferentially to the existing libc functions. In addition, version 2 of libsafe now includes protection against format-string attacks that appear to be the new scourge of unix. Of course, the best place for this protection is in libc itself, and glibc 2.2 does include some protection like this, but it is a compile-time option only, and further, is primarily designed to help developers fix overflows during program testing rather than helping sysadmins in the wild - it causes more of a performance hit than libsafe does.

    Anyway - as for 2.4.5, nice to see the VM is sorting itself out - I was that close to turning my desktop machine's ext2 partitions into UFS. I think I might convert them to ReiserFS now. :)

  • "Sun, in their infinite wisdom, hasn't gotten around to including a journaling file system or logical volume management into their OS." Bzzt, NEXT! I do hope /. readers do take information here with a regard it deserves.. in this case, bodly > /dev/null. (1) UFS does support journaling-- it logs metadata and one can enable it with the mount options in /etc/vfstab "logging" (2) Disksuite is a free Volume Manager that does various levels of RAID. It works however if you're using a large clustering environment, VxVM is recommended by Sun. I do both env. and like Linux since 95 however posting this mis-information does little for the redibility of the Linux Community. Whatever, patster
  • We also have serveral Filers (~14TB) consisting of Network Appliance and Auspex where I work in the Valley. I don't believe you will be tempted to consider Linux as a Filer since NFS is STILL very broken as of 2.4.4 series.
  • That's just a binary that's linked to whatever kernel/XFree that you're currently using when you 'compile'.

    --
  • I guess I can answer my own question by compiling and installing the new kernel, but I also wanted to know if anyone else has experienced the same thing:

    Ever since the kernel has supported USB (I used the backport to 2.2.something too) I've had a problem with my system clock. When I reboot from Linux, the hardware clock goes back to 1998.

    No, it's not the battery. I can reboot from Other OSes (Win9x/2k, BeOS) without resetting the clock. I've checked all of my shutdown scripts, and find nothing accessing the clock at all. All this leads me to think it's a kernel problem, and no one has been able to figure otherwise.

  • Already done. I've got it set to update the time on startup and at some point during the night. However, when I boot to other OSes, I have to reboot, enter BIOS, reset the time, save/exit/reboot. It's just a bit of tedium I'd rather do without.

    In response to the other comments:

    "What makes you think its a USB issue?" -- I'm not sure it is... but about the time I started using USB features to link w/ my Visor is when I noticed the problem.

    "How do you know its not happening on startup?" -- If I disable my auto-ntpdate, set the BIOS time, and boot into Linux, it keeps the date just fine. I can read the software and hardware clocks, and they're both correct. However, after a reboot from Linux, I can hit [del] and check the BIOS clock (or allow it to boot into any OS) and it's back at 1998.

    "When you shutdown you might want to make it set the hardware clock from the system clock." -- This seems redundant. I can check the hardware clock before shutdown, and it's fine. After reboot, it's not.

    Info I probably should've given in my first post (duh):
    Slackware 7.1
    Kernel 2.4.5 (now) (No, didn't fix the problem.)
    AMD K6-III
    128M RAM. Though, I don't really think that's relevant.

    Thanks for the suggestions!

  • I've checked all of my shutdown scripts, and find nothing accessing the clock at all.

    When you shutdown, you might want to make it set the hardware clock from the system clock:

    /sbin/hwclock --systohc

    On RedHat, at least, that's in the shutdown script (/etc/init.d/shutdown).

  • 245 is also qabalistically 'Spirit of God'. I bet there's some cutsey analysis one could do on linking that to Lord of the Hosts (ADONAI, I believe).

    I'll put 777 away again.

  • Ever since Solaris 7 (or 2.6 if you had DiskSuite) there has been a journaling option for UFS.

    Add the 'logging' option to your mount command (-o logging) and say goodbye to fsck.
  • is that the rhine driver you are using? I am having similar symptoms and going crazy trying to find the cause.

  • How does the real time scheduler compaire with VTRR
    VTRR [columbia.edu]
    And does anyone know what the status is of the linux scheduler?

    Anthony
  • -pre1: - Alan Cox: big merge

    You missed the really 'big merge' earlier. I've seen the 'further merging' for a number of kernel releases, but it looks like it's getting stepped up even more now.

  • According to David L. Kitts [fortunecity.com], well-known specialist of God's Divine Number, the phrase "Lord of Hosts" appears 245 times in the King James version of the Bible.

    Furthermore, 245 is divisible by 7, and everybody knows 7 is God's Divine Number.

    Don't you understand ? 245 times == kernel version 2.4.5 ??? It's OBVIOUS : God has decided that this version of Linux will be Lord of Hosts, therefore making NT and Solaris server looking like toys for pagans !

    Many thanks to Michael and Slashdot for reporting such a CRUCIAL event in the history of Humanity !

    "A door is what a dog is perpetually on the wrong side of" - Ogden Nash

  • by Cardhore ( 216574 ) on Friday May 25, 2001 @05:25PM (#197589) Homepage Journal
    According to Alan Cox, the VM system seems (finally) sane now (since 2.4.4-ac10). Check out Alan's full changelog for extreme details of changes at http://www.uwsg.indiana.edu/hypermail/linux/kernel /0105.2/1618.html [indiana.edu].
  • by stud9920 ( 236753 ) on Friday May 25, 2001 @11:04PM (#197593)
    linux-2.4.4-ac10 ? Now an anonymous coward can add code to the kernel ? I guess Cowboy Torvalds and Linus Malda did that to allow people to protect their anonymity in coding, e.g. for Microserfs who would want to participate.

    But then again, don't expect them not to add boot code to display an ascii goatse.cx picture. Or to call all identifiers nathalie_portman. An also expect the karma whores to add empty for-loops to gain karma.
  • I too would hesitate to go out and upgrade a server farm in any way without first testing the software in a controlled environment. Too many things an go wrong, and business critical systems are not the place to be experimenting.

    If it ain't broke, do some testing before you fix it.

    That being said, I am continually looking at migrating over to later kernels as the performance boosts I have seen have been pretty incredible particulalry for my Athlon boxes. One of them is a PDC/File/Print server (SAMBA), Apache+PhP web server, MySQL and PostgreSQL database servers, etc. mostly for development work, and the memory usage is way down from the 2.2.x kernels.

  • by V50 ( 248015 ) on Saturday May 26, 2001 @02:32AM (#197596) Journal

    Dude! That's my post!! Now if I were a big company I'd probably sue you... But seeing as I'm not I'm glad more people get to see my post, which I came up with at 2.4.4-pre6 or something and had to wait until 2.4.4...

    But anyway glad you got modded upto 3 so I don't have to look like moron "recycling" my old post...


    --Volrath50

  • / to start mirroring (by in my case booting from rescue cdrom, copying over the /dev/md instances into the ramdisk /dev, insmodding md, raid1 and reiser, mounting the partitions to mirror, chrooting to that mntpoint, editing lilo.conf (btw that's the LATEST lilo with md bootsector support)

    Yeah .. it's not very elegant yet. :)

    The best way I've figured out how to do it is to install your OS on a your first disk (like usual), bring up a mirror set in degraded mode (/dev/md0 with only second disk online), copy your OS over (in single user mode), lilo, and boot off the mirror. Then, bring your first disk online and let the mirror sync. Lilo one last time, and you're set.

    --
    All men are great
    before declaring war

  • Just when I read this my gf is playing The Sims on my work box so I can't mess around with the new kernel. I'm stuck on this G4 PowerBook.

    Fear for me.

  • Try mounting a file on a loopback device in 2.4.2 and you'll learn real fast why you should upgrade! Aside from that, less than 2.4.4 has some file system problems, which could lead to filesystem corruption. 2.4.4 also has the patch for the iptables hole included. (There a patch for 2.4.3 that fixes this too.).

    Haven't had time to figure out what the major changes are yet in 2.4.5.

  • Sun, in their infinite wisdom, hasn't gotten around to including a journaling file system or logical volume management into their OS.

    I see no reason for using a journaling file system; you pay significantly in terms of performance and you do not get any meaningful improvements in reliability. And LVM actually greatly decreases system manageability and reliability. After several years of using that junk on IBM boxes, I'm glad I don't have to put up with it on Linux. Linux does need to have it, of course, but only to be buzzword compliant.

  • Maybe you are using the wrong file system: even without any journalling, a few hundred GB of disk space shouldn't take "hours". It doesn't on Linux. Technically, journalling is overkill if all you want is fast reboots: you don't need the complexity of XFS or JFS.

    Also, the tradeoff doesn't work out. Let's say journalling has 5% overhead. That's an hour a day on disk-bound machines. For reasonable server uptimes, you spend more time journalling than you would if you just ran fsck once in a blue moon.

    If you say that you can't tolerate the downtime in a single chunk, journalling isn't going to protect you. To deal with hardware failures, you still need hot backup systems.

  • Some comedian always posts this hilarious comment, here's what it looked like when 2.4.4 came out.

    From the Changelog.... (Score:5, Funny)
    by V50 (V50DX@yahoo.coop) on Saturday April 28, @03:39AM EST (#182)
    (User #248015 Info) http://www.canada.com

    - Alan Cox: more resyncs (ARM down, but more to go) - Alan Cox: more merging (S/390 down, ARM to go).

    Oh my GOD!!! Alan Cox is being merged into the Kernel!!! They have his ARM merged in now, what next, WHAT NEXT?????

    --Volrath50
    Canadian and proud of it.

    This place is full of comedians. I didn't realize there were so many Paul Risers and Bob Saggets out there. Slashdot is a regular laugh factory, cranking out the funny stuff day and night.

The difference between reality and unreality is that reality has so little to recommend it. -- Allan Sherman

Working...