Forgot your password?
typodupeerror
Linux Software

Linux 2.6.26 Out 288

Posted by CmdrTaco
from the kernel-about-town dept.
diegocgteleline.es writes "After three months, Linux 2.6.26 has been released. It adds support for read-only bind mounts, x86 PAT (Page Attribute Tables), PCI Express ASPM (Active State Power Management), ports of KVM to IA64, S390 and PPC, other KVM improvements including basic paravirtualization support, preliminary support of the future 802.11s wireless mesh standard, much improved webcam support thanks to a driver for UVC devices, a built-in memory tester, a kernel debugger, BDI statistics and parameters exposure in /sys/class/bdi, a new /proc/PID/mountinfo file for more accurate information about mounts, per-process securebits, device white-list for containers users, support for the OLPC, some new drivers and many small improvements. Here is the full list of changes."
This discussion has been archived. No new comments can be posted.

Linux 2.6.26 Out

Comments Filter:
  • by Anonymous Coward on Monday July 14, 2008 @10:23AM (#24180897)

    It adds support for read-only bind mounts, x86 PAT (Page Attribute Tables), PCI Express ASPM (Active State Power Management), ports of KVM to IA64, S390 and PPC, other KVM improvements including basic paravirtualization support, preliminar support of the future 802.11s wireless mesh standard, much improved webcam support thanks to a driver for UVC devices, a built-in memory tester, a kernel debugger, BDI statistics and parameters exposure in /sys/class/bdi,

    Does it disturb anyone else how many words the bsdm & linux kernel community have in common? (this is not a troll).

    Frankly, I blame IBM.

    • by something_wicked_thi (918168) on Monday July 14, 2008 @11:10AM (#24181585)

      Nah, SATA gets rid of all that. No more master and slave. Now, we submit to the controller.

      • by value_added (719364) on Monday July 14, 2008 @12:36PM (#24182773)

        Nah, SATA gets rid of all that. No more master and slave. Now, we submit to the controller.

        Actually, submitting to the controller is redundant. I guess that makes the above a joke within a joke for those who thought otherwise. From the relevant Wiki article [wikipedia.org]:

        In fact, the drivers in the host operating system perform the necessary arbitration and serialization, and each drive's controller operates independently. Both are really "slaves" to the driver in the host OS.

        And because SATA presents the ATA interface to the system (the difference being how the chips are connected to the drive), you could say there's an additional joke in there, but one only those using SCSI would find funny.

        • Re: (Score:2, Funny)

          by Bloodoflethe (1058166)

          And because SATA presents the ATA interface to the system (the difference being how the chips are connected to the drive), you could say there's an additional joke in there, but one only those using SCSI would find funny.

          Too true.

    • by Eudial (590661) on Monday July 14, 2008 @12:35PM (#24182767)

      It adds support for read-only bind mounts, x86 PAT (Page Attribute Tables), PCI Express ASPM (Active State Power Management), ports of KVM to IA64, S390 and PPC, other KVM improvements including basic paravirtualization support, preliminar support of the future 802.11s wireless mesh standard, much improved webcam support thanks to a driver for UVC devices, a built-in memory tester, a kernel debugger, BDI statistics and parameters exposure in /sys/class/bdi,

      Does it disturb anyone else how many words the bsdm & linux kernel community have in common? (this is not a troll).

      Frankly, I blame IBM.

      Well, the kernel sources are (or were) pretty explicit in their sexual deviations. I remember several occurrences of the following comment: /* Fuck me gently with a chainsaw... */ in the 2.4 tree.

  • Intelfb still broke (Score:5, Informative)

    by CRCulver (715279) <crculver@christopherculver.com> on Monday July 14, 2008 @10:28AM (#24180967) Homepage
    They have still not enabled mode switching in the intelfb driver on laptops, meaning that I am forced to use ugly, unaccelerated VESA instead of the right driver for this sytem. This bug has been reported on kernel dev mailing lists and forums for at least three years, but no one with the skills seems to want to fix it.
    • by Nutria (679911) on Monday July 14, 2008 @11:10AM (#24181577)

      They have still not enabled mode switching in the intelfb driver on laptops

      Do any desktops really need a fb, or is it only so that there can be pretty pictures during boot, before [xkg]dm starts?

      • by MBGMorden (803437)

        Well, while I am happy as can be to use the VESA fb, writing of a bug in an existing driver simply because "no one needs it" (when as the original poster has demonstrated a desire for it) is just a bit disingenuous. Software developers tend to make really poor judges when it comes to features that users need. If you want to develop for an audience other than yourself, the best strategy is to listen to what the users ask for, and then implement that, assuming it doesn't break anything and that it's possibl

        • by samkass (174571)

          Well, users aren't always right about what they want either, especially when they request specific technical implementation details. The parent poster didn't say what they wanted to actually do or accomplish, just listed a specific technical item that hasn't been fixed. When "users" start asking for implementation details, you're probably better off either not doing anything or getting to the heart of the user's actual needs before implementing anything.

      • My Mac doesn't have any character video modes, so the only way you can get a console without X is virtually via a FB. Unless I'm misunderstanding how this works.

      • by rbanffy (584143) on Monday July 14, 2008 @01:09PM (#24183265) Homepage Journal

        It's actually useful if you, for some reason, need to drop to a text console and do something there (like restarting a firefox that started running amok an hour ago and now has all the system resources taken). I like my console to use the exact resolution of the laptop screen so that there are no weird pixels and, as a nice plus, the screen can fit a lot more text.

        Having a 900x1024 screen and a text mode that's about 480x640 pixels and 24x80 is kind of ugly.

    • by Viol8 (599362)

      Thats because frame buffers were one of those seemed-a-good-idea-at-the-time. But that time is passed.

      • by Chemisor (97276)

        > Thats because frame buffers were one of those seemed-a-good-idea-at-the-time. But that time is passed.

        Well, gee, that's exactly what I think about X. I'd much rather get rid of the huge bloaty X.Org and work on the framebuffer console, as I am doing right now. If there was kernel mode switching support, I'd actually upgrade my video card again. Currently my nVidia 7600 is the best one supported (correct me if I'm wrong). I would certainly like to get something newer to play Windows games faster, but if

    • by Anonymous Coward on Monday July 14, 2008 @12:51PM (#24183021)

      It's coming in 2.6.27 along with the GEM branch that was merged into master. Read Phoronix if you're into this sort of thing.

  • by FeatureBug (158235) on Monday July 14, 2008 @10:29AM (#24180999)
    What I would like to see more emphasis on in future kernels is a discussion of possible clever new tools and methods for configuring the thousands of kernel config options. None of the existing in-kernel-tree or out-of-tree config tools seems ideal.
    • by Zarhan (415465) on Monday July 14, 2008 @11:18AM (#24181715)

      Aye - would be great if there would be tool that I could eg. say "Ok, right now, at this moment, I have all my hot-pluggable USB/PCI devices plugged in, please detect and configure the options as needed". After all, that's what I do with a new comp: use lspci and similar tools to find out what's in the guts of the machine and then set options appropriately in menuconfig.

      • by repvik (96666) on Monday July 14, 2008 @11:26AM (#24181849)

        um... you have a distro that doesn't hotplug all the necessary modules for you?

        • Re: (Score:2, Informative)

          by nawcom (941663)

          You aren't following at all; the concept is that the modules havent been compiled and linked yet. More classic development distributions like Slackware don't provide 2 gigs of precompiled modules for different kernels (it usually comes with enough to pick up your hard drive, chipsets, etc and boot. That's where the kernel source comes in. you take 3 minutes and set it up and another 3 minutes (or hours, if you prefer the good-ol 386) to compile it. It's always been a ton faster than fighting with precompile

          • Re: (Score:3, Funny)

            by Zero__Kelvin (151819)

            "That's where the kernel source comes in. you take 3 minutes and set it up and another 3 minutes (or hours, if you prefer the good-ol 386) to compile it."

            Either:

            1. You have never configured and compiled a Linux kernel before, or ...
            2. You are the absolute worlds worst estimator of temporal resource requirements on the planet.

            There are 1155 options to configure:

            [Zero__Kelvin@stormbringer linux-2.6.git]$ git branch
            master
            * v2.6.26-deeppurple-eldd
            v2.6.26-rc7-deeppurple
            v2.6.26-rc8-deeppurple
            [Zero__Kelvin@st

        • by pbhj (607776)

          um... you have a distro that doesn't hotplug all the necessary modules for you?

          I think he means he doesn't want to compile/build all modules, just the ones that the system will use.

      • Thats what udev does. And it does it without even recompiling the kernel - it works with standard kernels. Awesome!

        Seriously, better config tools are not needed because normal people does NOT need to compile the kernel. Compiling your kernel "just because" it's like recompiling libc. You CAN do it, but not many people does it and people who does do not need better config tools.

    • Clever, but takes some time:

      1. make randconfig
      2. Compile, install and boot the kernel
      3. If your system doesn't boot or lacks a driver, goto 1.
    • by jd (1658)
      At least, to do a pre-configuration tool. You can currently list the processor(s), scan many of the physical and virtual busses, discover nearby network resources, etc. If that information were to be collected and then compared with "best practice" values for configuration options, it should be possible to set up many of the kernel options automagically. Not all, perhaps, but many. If you want something really sophisticated, you could do this as a two-pass thing. First pass uses "best practice" values and e
  • Just curious,

    If you have compatible wlan hardware like Atheros [atheros.com], would it be possible to configure a mesh network on them? Or do you need special 11s compatible hardware?

    I know the OLPC has specialized hardware for this.

    • Re: (Score:2, Insightful)

      by Shadow_139 (707786)
      • by leuk_he (194174)

        I guess that is a yes? Or yes, sometimes? THat still does not give the state of that project.

        • dd-wrt is a special firmware for Linksys WRT54G (and compatible) embedded routers.
          Chancec are that someone has written it, but I'm not sure it will run on any other type of kernel/hardware.
          At best you'll be able to port the functionality to x86 hardware and generic (or caompatible) network cards.
          At worst, you'll have a bunch of gibberich that only makes sense if you know the peculiarities of the hardware.

      • by Klaus_1250 (987230) on Monday July 14, 2008 @11:33AM (#24181939)
        802.11s != OLSRv2 . 802.11s seems to work at the MAC layer, whereas OLSRv2 works at the IP layer. They are not the same. There are quite a few mesh networking protocols out there and in development, but I haven't seen a clear winner yet.
    • Re: (Score:3, Informative)

      by gradedcheese (173758)

      It's part of the mac80211 layer, so it in theory supports any mac80211 driver, that is a 'soft MAC' WiFi chipset. There are minor driver changes required to support mesh (basically adding mesh beacons) and right now the zd1211rw and B43 drivers work. We have more details here:
      http://o11s.org/trac [o11s.org]
      B43 is your best bet at the moment, if you have a few of those, give the HOWTO on the o11s website a try and you can have your own mesh network.

      Eventually other soft-MAC chipsets can work, such as Intel's iwlwifi,

  • by redelm (54142)

    OK, 2.6.26 is out, and kudos for all the good work. But where is a truly writeable NTFS? Many larger USB drives are shipping with this pre-installed, so true write support is needed in the kernel.

    AFAIK, current kernel "write" support does not including creating files or directors (presumably just modifying/appending to existing files).

    I've tried ntfsprogs, but not got it to compile x86_64.

  • Good Featurelist (Score:5, Informative)

    by Doc Ruby (173196) on Monday July 14, 2008 @10:40AM (#24181171) Homepage Journal

    I wish every kernel release announcement included a highlevel featurelist like that. Not just a ChangeLog, as each bug is fixed or small feature is added. But rather a fairly highlevel list of new and improved (and fixed) features like the one in this Slashdot story. Best if in the announcement itself, but at the very least always in the release package.

    That way most of us can decide whether to upgrade, or to wait (perhaps for the x.1 version, which is typically a higher quality bugfixed delivery). Since kernel upgrades require rebooting (and again to downgrade after test), knowing whether to ignore a release based on its highlevel upgraded features itemization is a very effective announcement feature, which makes all of us using the releases more productive.

  • Translation please? (Score:2, Informative)

    by Vellmont (569020)

    Some of these I know what they are, and some I can guess at. But what is:

    read-only bind mounts
    x86 PAT (Page Attribute Tables)
    basic paravirtualization support
    BDI statistics and parameters
    per-process securebits
    device white-list for containers users

    And what might I see as a result of these improvements somewhere along the line?

  • by tucuxi (1146347) on Monday July 14, 2008 @10:59AM (#24181417)

    Reading on it, it seems that Linus never has been a great fan of kernel debuggers. From a famous post [lwn.net],

    I happen to believe that not having a kernel debugger forces people to think about their problem on a different level than with a debugger. I think that without a debugger, you don't get into that mindset where you know how it behaves, and then you fix it from there. Without a debugger, you tend to think about problems another way. You want to understand things on a different _level_. [...]

    I agree that stepping with a debugger instead of thinking real hard about the code (and using abundant log statements) is generally a waste of time, and that expecting to catch rare occurrences of weird race conditions with a debugger is not worth the effort. Sloppy programmers don't take the time to think, and rely too much on fixing what they could have not broken. Unit tests, although more expensive to code, can be reused many times - debugging sessions are one-shot.

    On the other hand, even good programmers can get stuck and benefit from a debugger every once and then. I guess this argument finally won the day.

    • Big thing that a debugger generally gives you isn't the trace through the code.

      Its look at the state of the system when you know there's a problem.

      Now you can probably get there by using logs...assuming that someone has written all of the state information you need into the logs for that particular instance.

      If they haven't, though, frequently that'll save a lot of time - ESPECIALLY when you're debugging other people's code.

    • by Fallen Andy (795676) on Monday July 14, 2008 @11:54AM (#24182229)
      These days I'm too lazy to bang around fiddling with OS's, but back in the early 80's when I ported the UCSD p-system to many machines, we didn't usually have *any* kind of debugger except our own log statements. So, one day I got given an Orion Instruments logic analyser (which could do hardware debugging for MC68000). Beautiful. Best productivity disabler I've ever seen. On the other hand, because of a really bad experience on my first p-system port, my own diagnostic code for a later port made me screw up my deadlines badly.

      With high level code, a decent debugger is really really useful. With low level code, not so much.

      (It's amazing though how many high level programmers don't understand the way debugging changes program behaviour (variable initialization etc - don't even mention heisenbugs)).

      The best ever debugger is the "cardboard man". If you really get stuck you explain the code to anyone (including the cleaner). That way, (even though the cleaner doesn't understand anything) you exercise another part of your mind and *see* the problem (... well here we shift left (wtf? right?) oops).

      Andy

    • Re: (Score:3, Informative)

      by 0xABADC0DA (867955)

      On the other hand, even good programmers can get stuck and benefit from a debugger every once and then. I guess this argument finally won the day.

      Actually after programming in C the past five years I find a debugger completely worthless. Pretty much all problems boil down to:

      1) memory / pointer errors
      2) usage errors (bad casts, unset variables)
      3) code too complicated to follow by reading

      The first is covered by valgrind, or if your system doesn't have valgrind then first writing for x86 then porting. The second is covered well by gcc warnings. The third is covered better by logging than a debugger, or better yet just not writing complicated softwar

    • by Black Art (3335)

      The reason I want a kernel debugger is so I can figure out where the kernel was when the whole system locked up. Trying to guess takes a whole lot of time. (Yes, kernel lockups are rare, but I think I am fighting bad hardware that is not handled gracefully by the kernel.)

      Frozen kernels are the hardest to debug. (Insert Korn shell reference here.)

    • Re: (Score:2, Insightful)

      by Anonymous Coward

      stepping with a debugger

      using abundant log statements

      I love how people with anti-debugger attitudes always seem to rely instead on printfs. as if getting the exact same info from printfs is somehow more noble than from a debugger.

      they're both tools. it's up to the developer to be intelligent and an intelligent developer will use the tools that help them achieve the job best. in some cases that's a debugger, in other cases it's debug printfs and logfiles

  • Kernel Debugger (Score:2, Interesting)

    by wvdmc (1227780)
    While the debugger is, in fact, remote, it appears that perhaps the Linux Kernal really is JUST NOW getting an actual debugger. From TFA:

    For many years Linux has not included a kernel debugger. Linus Torvalds vetoed them for years, for reasons that he explained quite well in a know email: "When things crash and you fsck and you didn't even get a clue about what went wrong, you get frustrated. Tough. There are two kinds of reactions to that: you start being careful, or you start whining about a kernel debugg

  • When are the patches at http://people.redhat.com/heinzm/sw/dm/dm-raid45/ [redhat.com] going to be included? I'm running a dualboot box so have to run the BIOS-fakeraid that works with Windows. I had to run through a few hoops to get it working with 2.6.24 (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/220493 comments) and for now it works...but what if I want to update kernel at some point?

  • I see that windfarm support for the PowerMac 12,1 series has been added.

    Does this mean I can finally run Linux on this late-model iMac G5 without the fans exploding?

    Anyone running it now?

A freelance is one who gets paid by the word -- per piece or perhaps. -- Robert Benchley

Working...