Linux Kernel v2.6.23 Released

diegocgteleline.es writes "After 3 months, Linus has released Linux 2.6.23. This version includes the new and shiny CFS process scheduler, a simpler read-ahead mechanism, the lguest 'Linux-on-Linux' paravirtualization hypervisor, XEN guest support, KVM smp guest support, and variable process argument length. SLUB is now the default slab allocator, there's SELinux protection for exploiting null dereferences using mmap, XFS and ext4 improvements, PPP over L2TP support. Also the 'lumpy' reclaim algorithm, a userspace driver framework, the O_CLOEXEC file descriptor flag, splice improvements, a new fallocate() syscall, lock statistics, support for multiqueue network devices, various new drivers, and many other minor features and fixes. See the changelog for details."

Re:Answer: Linux will never be GPL3.

[QuantumG]

Unless Solaris is released under the GPLv3 and Linus sees some stuff he wants.

Really, he said that [lkml.org].

Re:Answer: Linux will never be GPL3.

[GigsVT]

You can't just relicense code that was GPL2 only. It would all have to be rewritten, from scratch. Linux will NEVER be GPLv3.

Re:Answer: Linux will never be GPL3.

[sconeu]

No, the problem is finding *all* the copyright holders and getting them to agree to GPLv3.

The copyright holder can license the code however he damn well pleases.

Re:Answer: Linux will never be GPL3.

[phantomlord]

[citation needed yourself]

What Linus said was "I was impressed in the sense that it was a hell of a lot better than the disaster that were the earlier drafts. I still think GPLv2 is simply the better license." [lkml.org]

A couple days later, he expresses more angst with the GPLv3 and the FSF [lkml.org].

The bottom line is

I consider dual-licensing unlikely (and technically quite hard), but at least _possible_ in theory. I have yet to see any actual *reasons* for licensing under the GPLv3, though. All I've heard are shrill voices about "tivoization" (which I expressly think is ok) and panicked worries about Novell-MS (which seems way overblown, and quite frankly, the argument seems to not so much be about the Novell deal, as about an excuse to push the GPLv3).

So... I'd hardly say, as you did, that he doesn't mind the GPLv3. In fact, the FSF shills really ticked off a lot of kernel devs by trying to berate them into switching to the GPLv3 back in June/July.

Re:Answer: Linux will never be GPL3.

[QuantumG]

Dude, if you actually read the kernel mailing list you would know that Linus has said that he can change the license whenever he wants. All he has to do is post a notice to the list, and add the same notice to the license file specifying a date when the license will switch over. Anyone who doesn't agree will have an opportunity to opt-out, at which point their code will be pulled out and rewritten, or opt-in. The ones that don't do either can be assumed to opt-in until such time as they complain.

This has been done before.. with the syscall interface exception.

Stop repeating myths and do some research.

Re:A pre-packaged ISO, please...

[Anonymous Coward]

basically archlinux = slackware's simplicty with debian's package manager with gentoo's compiling power

i'm not joking...
ala slackware:
- it's BSD startup scripts
- all packages are pretty much untouched and rely on upstream releases...there is no backporting

ala debian:
- awesome package manager
- and for me, i find it easier to use and especially better when the package manager breaks (i've never been able to recover from a crapped out dist-upgrade without reinstalling...)

ala gentoo:
- obviously not everything is compiled...but if you do, it's 3 commands:
    - abs (sync with PKGBUILDs which is the equivalent of ebuilds)
    - makepkg (compile)
    - pacman -A package-1.0.0.tar.gz (install)

Re:Ummm. Neat.

[wanderingknight]

Linux has a steep learning curve

...only if you come from another OS. No OS is inherently hard or easy to use--you just need to get used to it. I'll agree, there are people that have an easier time getting used to something new (I'll include myself in that group, it only took me two days to do on Linux everything I did on Windows, and a week to nuke my XP partition), but it doesn't mean Linux is hard to use per se. It's not what you're used to it, that's all. Hell, if I had to go back to XP and have to hunt on Google to find a piece of software I need, instead of using Ubuntu's Add/Remove or Synaptic, or SUSE's YaST, I would be bothered. I would also be bothered if things didn't work like in the GNOME desktop I'm used to. Of course, *I* have an easy time adapting myself (and, besides, I always enjoy trying out new stuff), which doesn't mean *you* have to have an equally easy time. I'm tired of people bashing Linux "non-user-friendlyness". It's just that you're not used to it. It's not a crime, but it's not a reason to bash it as unfriendly, either.

Re:Great! In other news, RIP linux for the desktop

[Randle_Revar]

Thunderbird is not dead, and David and Scott are leaving Mozilla, but retaining their roles as module owners of Thunderbird.

http://robert.accettura.com/archives/2007/10/08/thunderbird-in-crisis-no [accettura.com]

http://standblog.org/blog/post/2007/10/08/The-future-of-Thunderbird [standblog.org]

Re:Ummm. Neat.

[tomhudson]

"It takes a specific type of person to get Linux running and to a point where it can be productive even for nontechnical users (which is the majority of users that use computers)"

WTF???

Linux installation for dummies, PHBs and Windows sysadmins (but I repeat myself)

1. Stick a second hard disk in your machine (don't be a cheap SOB - the OS is free, give it some room to live)
2. Stick a modern distro in the dvd drive.
3. Boot up
4. click for your time zone and geographic location
5. Tell it that its okay to start your internet connection automagically.
6. click on the packages you want (or just accept the defaults if you don't know what you're doing)
7. set your partitions the way you want (or just accept the defaults if you don't know what you're doing)
8. click ok
9. go do other stuff while the dvd installs 5 gigs of software ...
10. enter your root password, a user account and password.
11. click okay
12. watch as your computer boots into your new linux install.
13. pick the gui you wnat to use
14. log in
15. do whatever you want - your web browser(s), office suite(s), email program(s), server(s), etc., are already installed and configured.

If you can't follow that, print it out and pay some PFY* in grade 9 $20.00 to help you.

(if you don't recognize the reference, you're obviously new here and deserve to be beaten with a clue-by-four, both ways, in the snow, etc...)

Re:Answer: Linux will never be GPL3.

[QuantumG]

Linus has outlined a number of possibilities, a lot more favorable than you are suggesting, but yes, there is no motivation at the current time to change the license.

But that wasn't the point of my post.. the point of my post was to stop the meme that the license can't be changed. It can. Or, at least, Linus has said it can, and that should be good enough, cause if he thinks it can be changed and there is a reason to change it, then he will, and we'll be having a different discussion.

Re:A pre-packaged ISO, please...

[Martin Blank]

It's not yet released, but Fedora 8 Test 3 has been running the 2.6.23 kernel code. I suspect that within days (hours?) the RC labels will be pulled from the RPMs.

Re:What about O_CLOEXEC for sockets?

[Jaxoreth]

Sockets are not close-on-exec by default, so you may pass a sensitive socket to a child.

No, because sockets are unbound when created. If you set FD_CLOEXEC prior to calling other socket routines, the worst that happens is the child gets a fresh socket that's not connected to anything.

Pipe endpoints are bound together when created, so that might be a problem.

Re:A pre-packaged ISO, please...

[Nikron]

It's not source based... Coming from an Archlinux user..

Re:A pre-packaged ISO, please...

[kcbanner]

No. Your wrong there sir. Its binary based, however non-repo packages (Arch User Repository) are compiled via build scripts. You can also compile packages from source with pacman (pacman -Sb instead of pacman -S).

real Linux news from 2 weeks ago

[hexfortyfive]

Not to nitpick, but the milw0rm main page says '2007-09-27' beside that exploit. I'd hardly call that today's Linux news.

Re:Userspace drivers?

[QuantumG]

Umm.. they are releasing specs etc, so people can write drivers. All those stupid arguments about patents were based on speculation and are now known to be false because NVIDIA and ATI are now doing the stuff that before they wouldn't and the sky hasn't fallen.

Re:Why do that much work?

[SnowZero]

...and watch Theo actually turn into a demon.

Re:Userspace drivers?

[cnettel]

ATI has also stated that some features won't be implementable, exactly due to patent issues, by just following those specs. That's one of the reasons for why they release specs and redo the driver in an "external cleanroom", rather than releasing an open-source driver themselves. (Yeah, cleanroom applies to reverse engineering and copyright, but it would be a good practice to protect the company in a patent suit as well.)

Re:What about O_CLOEXEC for sockets?

[kasperd]

If you set FD_CLOEXEC prior to calling other socket routines, the worst that happens is the child gets a fresh socket that's not connected to anything.

If the child already got a file descriptor for the socket, it doesn't matter what state it was in at that exact time. The file descriptor does not go away, the child will still have the file descriptor by the time you are doing something sensitive on that socket. Looking back I think it was a design mistake to make file descriptors inherited across exec by default. Close on exec should have been the default, and you should disable that flag yourself on those few file descriptors you want to keep open across exec.

Of course it is easy to look back and point out mistakes. It is much more tricky to fix design mistakes later. I think it is great when some people insist on at least trying to fix design mistakes rather than keeping them around forever just for compatibility. Of course in this case it is not trivial because the design mistake is probably not implementation specific, but rather in the standard that multiple implementations follow.

Re:Answer: Linux will never be GPL3.

[petermgreen]

I can't see the more legally paraniod distros accepting a change in license without the permission of the original authors of the code.

pulling legally dubious licensing crap (e.g. the xfree86 non GPL compatible license which is a problem because nearly every X app links against X libriaries and the shift of large parts of cdrtools to the GPL incompatible CDDL while the rest was still under the GPL) is a damn good way to get your project forked and lose your influence.

Re:The real Linux news today.

[Anonymous Coward]

Old news. This was fixed [kernel.org] in 2.6.22.7

Re:What about the license?

[00_NOP]

The parent post is a troll. The facts are this: one person attempted to change the licensing terms and submitted a patch. Nobody responsible for the Linux kernel accepted the patch. Various people then went mad and started screaming "fire" in the crowded theatres of various mailing lists.

Re:A pre-packaged ISO, please...

[josephdrivein]

Why do you prefer a ISO to a compressed tar archive?
If meant that you want a precompiled kernel, you should wait until your distro offers a package.

On the other hand, if you want to try the new kernel now, you have to build it yourself.
Many users complain that the "make-based" compiling is too difficult, hence distros usually offer some kernel building facility. Check your documentation - or google.

Here's how I'm building the new kernel right now on a Debian system:

cd /usr/src/linux-2.6.22.9/
ketchup -r 2.6
make-kpkg --initrd --append-to-version=`date +%d%m%y` kernel_image

Re:I love my Thinkpad

[FauxPasIII]

> thinkpad-acpi: enable more hotkeys, add input device support to hotkey subdriver

Although it's not merged yet, Thinkpad owners should also check out this project:

http://tpctl.sourceforge.net/ [sourceforge.net]

Using the tp_smapi driver I can, among other things, clamp my battery charger to
stop at 70%, which makes the battery last a _lot_ longer. =)

Re:bloat

[Chirs]

You do realize that many of the options in the kernel are mutually exclusive? You use the slab or slub allocator...only one of them gets included when the kernel is built.

The CFS scheduler actually *simplifies* the code as compared to the old one, as does the new readahead code.

Sure, the size of the kernel source code is continually increasing, but most of the increase is for hardware drivers. Also, the running binary doesn't increase in size nearly as fast as the source does...and as others have mentioned, you can always turn off the stuff you're not using to shrink it back down.

Re:Ummm. Neat.

[wanderingknight]

I'm not saying "x OS is easy to use". I'm saying there's not such thing as "ease of use". It's not an argument about how easy it is for *me* to use Linux, it's an argument about how easy or hard is to get used to something new, especially if you're a non-techie.

To put an example, my grandfather was used to Windows 3.11. He had AutoCAD for his tailoring business, and it was wonderful. It did everything he needed, and his productivity was excellent. He was used to it. My father came in one day and replaced his old 486 with a new PC and put Windows XP on it. My grandfather went crazy, he didn't understand a thing. He was so used to 3.11 that XP's "user-friendliness" meant nothing to him. It took a couple of months till he could finally get used to XP.

Another example, and a much more radical one: My translation teacher was telling us the other day about the days when she worked with a manual typewriter. She was really good at it. But then there came the PC and the graphical word processors--she also went crazy. She wasn't used to typing straight without manually breaking the lines! This also took a certain amount of learning time.

How many people are there that actually used something other than 9x based Windows OS? I was raised with DOS, my first GUI experience was with the old Mac OS, then I passed through Windows XP and ended up today in Linux. I'm used to figuring how things might work in different systems. Most people aren't. Most people were introduced to computers in the 9x or NT era, and don't know anything else. How can you expect them to find something completely different to what they're used to "easy to use"?

I'm not saying people should move to Linux. I'm just countering the "ease of use" argument. That's it, nothing more, nothing less.