Linux 2.6.0 Expected In Mid-December

Ridgelift writes "CRN is reporting the release of Linux 2.6.0 in mid-December. 'Torvalds, Linux's lead developer and now an OSDL Fellow, and Linux kernel maintainer Andrew Morton this week released the test10 version of Linux 2.6 after a three-year development effort. A final test11 version is expected before they sign off on the production version next month.' Get ready for 'major scalability improvements, faster performance, enhanced support for embedded systems and, to a lesser extent,' a kernel that 'supplies desktop systems with better USB and FireWire support.'"

Comments

Re:Xmas

[Libor Vanek]

Don't even think about getting it this year ;-) Have you ever seen such a large project on schedule? .)

This isn't unexpected

[The One KEA]

I've been running -testX kernels for a while now and the claims made about all the improvements are true -- 2.6 is a far better kernel than 2.4, IMO.

The prediction that akpm made about mid-December sounds about right as well -- 2.6.0-test10 could be 2.6.0 right now and I doub there'd be any showstoppers to block it.

Re:This isn't unexpected

[EriDay]

In the last 12 years M$ has really only sold 3 kernels:

• Win3.0, Win31, Win95, Win98, WinME all built on the DOS kernel.
• Various versions of the WinNT kernel
• Various versions of the XP kernel

There has been some tweaking of the kernel for market positioning and bug fixes, but no other major architectual overhauls. Everything else has been modifications to the window manger which is the equivilent to a new KDE or Gnome version. Let's not forget that M$ is impoverished comared to Linux.

Re:This isn't unexpected

[ultranova]

That's why we have distributions, which can up the minor and major numbers whenever they like, or call themselves whatever flashy names the marketing people come up with. For example, there's Red Hat Linux (was, at leat), Red Hat Enterprise Linux, Advanced Server (?) and so worth. A newbie is unlikely to know or care about the kernel version anyway...

On the other hand, calling things by version number is much clearer for anyone who wants to upgrade/modify their system. If you had Linux ME and Linux XP, which one would be better ? On the other hand, if you have Linux 2.4 and Linux 2.6, the choice is pretty obvious.

BTW. I just have to ask. What kind of IT director decides based on flashy names, and not comparison by technical merit ?

Re:This isn't unexpected

[Ianoo]

The NT and XP kernel are the same. XP's is just an incremental improvement over the NT4 Kernel (it's almost the exact same Kernel in Win2000 aka NT5). Microsoft have just moved things in and out of the Kernel since NT3 to improve performance (such as the full DirectX support added to W2K rather than the awful DirectX emulation in NT4).

Also, while it's true that the 9x series are built on top of DOS, the same as 3.11, there were a significant number of changes to support 32-bit and things like pre-emptive rather than cooperative multitasking. Not quite as many changes as Microsoft would have us believe (it certainly wasn't the 100% new 16-bit-free operating system everyone expected based on what they were saying), but certainly a lot of changes under the hood.

Lastly, 3.x never had a Kernel, it was effectively a big graphical library and program launcher, and not much more.

That given, I'd revise your list to:

• DOS Kernel (if you can call it a Kernel!)
• 9x Kernel (which built on DOS/3x)
• NT Kernel (featuring in 2K and XP with tweaks)

I agree entirely that Linux has to be taken in a larger context. Considering that in the same time it's taken to go from 2K to XP we've seen Gnome go from v1 to v2 and KDE go from v2 to v3, which were much more significant changes, I'd say that Linux on the desktop is advancing more rapidly.

Besides, there's no reason to rewrite a Kernel from scratch if you get it right the first time. There don't seem to be that many fundamental problems with the Linux Kernel, so the continued process of tweaking and gradual improvement seems set to continue. Microsoft, on the other hand, had to write a new Kernel because 9x was such a horrendous mess and lacked quite a lot of modern features and elegance.

Re:This isn't unexpected

[sharkey]

Oh, yeah, that was a leap alright! New splash screen and a different colour scheme. Whoop-de-do.

Come on, give a little bit of credit. Starting with XP, you can open .ZIP files without findng, downloading and installing a archiving software application. What other operating system can read compressed file archives right after install?

Re:This isn't unexpected

[spongman]

You're correct, right up until you imply that NT was written to replace 9x. NT shipped before win95, and it existed within MS well before win95 was even conceived. win95 was writen because they wanted to ship win32 to customers and NT didn't run as well as OS/2 or Win3.1 on consumer-level machines.

There was an effort to write a completely new win32-based replacement for win31, cougar was the codename for the 32-bit DOS kernel, and panther was the win32 core, but panther was canned and cougar was merged into chicago (win95).

What the fuck?

[Anonymous Coward]

Morton acknowledged that the XFS and JFS file systems, which were originally developed under a Unix license and then ported over to Linux, could be a sticky issue that lawyers can exploit. "SGI did develop it. It could be [SCO] has a legitimate case there, not technically, but on the letter of the law," Morton said.

Re:What the fuck?

[Zapdos]

They were actually developed on white paper. The first implementation may have been on a UNIX. Both HP and SGI are aware of license agreements and code ownership. All of their programs are fully developed on white paper, patents applied for, granted or pending, then creation of the FIRST implementation. In doing business this way, HP and SGI own the technologies, SCO has no leg to stand on.

Linux 2.6: I can only recommend it!

[quigonn]

We're running 2.6.0-test9 on several production machines at work, and we had absolutely no problem, so far, but a huge improvement on performance instead. The only thing one has to care about is that 2.6.0 requires module-init-tools instead of modutils. It's especially important to read the upgrade guide, so that one can easily switch back to 2.4.x even when using modules (not that I would miss 2.4.x, but you never know... not all people will have such flawless upgrade processes as I did).

Re:Linux 2.6: I can only recommend it!

[Billly Gates]

Mod this sucker up!

It will save you countless headaches if you read the files about module-init. I have been using the linux kernels since 2.0x but never had an upgrade break a system. It totally ruined my redhat 9 box doing that.

I upgraded to module-init and after that 2.4 wouldn't boot. grr.

Is there a way to have both installed so I could dual boot 2.4 and 2.6?

I switched back to my more upgrade friendly FreeBSD until 2.6 was more stabilized and more distro's supported it. My guess is Gentoo would be the first.

Re:Linux 2.6: I can only recommend it!

[dossen]

I don't know about Gentoo, but as far as installing module-init-tools/modutils, kernel and related stuff Source Mage GNU/Linux has been there for quite some time.

Re:Linux 2.6: I can only recommend it!

[shani]

It depends on the production, doesn't it? I mean, if you have a machine that collates addresses and prints mailing labels, then you can run with the test kernel and if it gives you better performance and works, who cares if it's "only for testing"? What's the problem?

Re:Linux 2.6: I can only recommend it!

[Dave2 Wickham]

make moveold before make install - it moves the current lsmod, modprobe etc to lsmod.old, modprobe.old etc.

Re:Linux 2.6: I can only recommend it!

[BoysDontCry]

Nope. Linux 2.4 needed modutils. 2.6 needs module-init-tools.

good stuff

[inode_buddha]

I've tried several -test kernels, and the desktop performance boost blows my mind. I could probably stay with -test10 and akpm's patches, but that's not what the kernel team is really looking for (I think). Last I heard, They really want people to hammer on stuff like PnP, scalability, USB, and ACPI.

Any ideas on how much akpm's patches end up becoming "mainstream"? After reading the changelogs (and using the patches), I think it'd be a good idea.

Re:good stuff

[crimsun]

All of the critical fixes from -mm are pushed into Linus's current tree. Just take a look at the "Merged" section immediately following "Latest Linus tree" here; repeat with the previous -testX-mmY patch announcements.

Now take a look at this under the "Andrew Morton" heading and notice how many of those patch headings ring a bell. Yessir, he has been kickin' arse and taking names.

Newsflash - Christmas Postponed..

[Channard]

NEWSFLASH: Chrismas has been postponed as Santa has had an injunction placed on him by Linux copyright claimants SCO.

'We intercepted a number of letters from Linux users to Mr Claus, requesting that he bring them the new 2.6 Linux Kernel for Christmas, and given that at least 50% of them have been good, we believe he was going to supply the requested code' said Daryl McScrooge, head of SCO's 'Grabbit and Runne' division.

'Linux 2.6 was of course entirely written by ourselves and the tooth fairy and to protect our rights we have taken out an injunction preventing Father Christmas from delivering any presents this Christmas. We believe this is a fair and legal action. And anyway, I never did get that bike I asked for.'

IDE support on Dell Latitude D600

[cerberusss]

I'm running this new kernel to get full support on whatever IDE chipset my Dell Latitude D600 laptop uses. Combined with the better performance this kernel really rocks.

Not to nitpick, but

[Space cowboy]

I thought this was all Linus' doing - in the LWN text, he says that Andrew is off for a couple of weeks so he may release a test11 before Andrew decides to take it on for release management...

Simon

They seem to have som problems though...

[errl]

A problem, potentially delaying release? Seems that they don't really know what causes it as of yet...

Re:Keyboard still doesn't work

[MrHanky]

Have you tried "pci=noacpi" at the boot prompt? I had complete crashes (no kernel panic) when the keyboard was initialized before I tried that. Maybe your problem is related? (This is for a PS/2 type keyboard, BTW.)

I don't think I have any problems with 2.6.0-test9-mm1 at all.

Re:Xmas

[contrasutra]

It was suppose to be released October 31, so you are right. This sounds about right. 2-3 months behind schedule.

I love 2.6. Ive been using it since -test4 and Ive only had a couple of issues with some of the Morton Patches. Other than that, it IS everything people say it is (on my desktop machine). Fast, stable, and performs amazing under heavy load.

Its the first time I could compile GIMP and surf the web without feeling any slowdown when scrolling.

YES: Gentoo and Source Mage do it

[FreeUser]

I upgraded to module-init and after that 2.4 wouldn't boot. grr.

Is there a way to have both installed so I could dual boot 2.4 and 2.6?

Gentoo GNU/Linux supports this, and I believe Source Mage does as well.

I run 2.6.0-test10 and 2.6.0-test9-mm5 on numerous Gentoo boxes with no problem, and occasionally switch back to 2.4.22 without difficulties.

I'm not sure how they do it exactly. A quick perusal of module-init-tools and modutils revealed that, for example, bot install /sbin/modinfo, and numerous symlinks to things like lsmod.old exist. There is probably a boot script that detects the kernel version on boot, creates the appropriate links, and then loads up the modules, but nothing in /etc/init.d jumped out at me as the culprit.

In any event, it is certainly possible have both installed and functional, and to seemlessly move between 2.4.x and 2.6.x kernels.

My bet

[Vilim]

Sure, mid-December might be when its ready. But I'm sticking with my bet of a Christmas day release. Linus likes to release kernels on holidays (he did one on christmas a while back and noone can forget the greased turkey). Perhaps he will name this one the greased reindeer or something :p

Re:Debian support

[scharkalvin]

Debian supports 2.4 right now. You get your choice with 'woody'. True the default install CD DOES install 2.2, but if you boot with the BF24 image, you will install the 2.4 kernel. AND 2.4 kernel images ARE in the 'stable' package tree.

Red Hat builds

[Bernie]

It's worth reminding RH/Fedora users that Arjan van de Ven maintains kernel RPMs (including new module RPMs etc), and those with yum and apt can very easily test 2.6 using these files.

Read the readme.txt for full details.

2.6 Kernel issues - Is it really ready?

[cronot]

I've been wanting to run the 2.6 on my Debian-unstable box, but I've been scared off by some of the negative reports I've heard so far. Some of them I've seen on comments from this article.

Most notably (for me) is devfs not being actively supported anymore (being shifted in favor of udev). That's sad, at least for me, because I've been using devfs since the early versions of the 2.4 kernel, it always worked well for me, and from what I've heard about udev so far, devfs seems like a more elegant and mature solution.

Then there are problems with USB devices, and others that, being narrowed down, comes down to problems on the APIC interface. From what I've heard so far, it doesn't look stable, so why ship it on linux 2.6?

There's also this problem with Kernel Preemption. I'm using it on my 2.4, and I don't want to go without it on 2.6. Of course I might just be lucky to no stumble on this problem, but the fact that it can trigger an oops on someone just scares the hell out of me.

Finally, there's a problem I've experienced myself, but didn't care to report at the time. It's quite old by now (I think it was around version 2.5.65~2.5.70). It has to do with software-raid. I've got a RAID-0 array with 3 SCSI Disks (6gb + 2x4gb = ~14Gb). The disks are old, I know, just like the controller (Adaptec AIC-7xxx). But they work just fine on the 2.4 kernel. So, at the time I decided to give 2.5 a try, just to find out that my array wasn't being detected/mounted. Googled around, found some similar reports and some possible workarounds, but none worked, so I switched back to the 2.4 kernel and haven't touched the development kernels since. It might just be resolved by now, I don't know... Anyway, I will soon replace these disks by a couple of IDEs, with no RAIDing, to save some CPU cycles, so this will not be such a big problem.

So, anyone care to give me one (or more) reason s to try 2.6 again?

Re:2.6 Kernel issues - Is it really ready?

[Dionysus]

Most notably (for me) is devfs not being actively supported anymore (being shifted in favor of udev).

Well, the reason devfs is not actively supported is because the maintainer disappeared, and nobody has stepped up to take over the code.

udev seems to solve the problem of only have /dev files that correspond to devices you have connected. But they are moving "back" to the old /dev/sg?1 etc. I much prefer the devfs nameing scheme (what do you do if you have more than 24 hds? The devfs had a solution for that).

alsa sound kernel support

[HelloKitty]

I'm looking forward to the end the nasty OSSFree, and the beginning of the silky smooth default included Alsa sound kernel.

no more annoying upgrading my system to Alsa when I want to make it into a professional audio workstation

Dear moderators,

[greppling]

why is it that always people get a +5 Informative when they post a bug report to slashdot instead of sending it to LKML? When Logitech wireless hypersonic ultra XZ763 keyboard isn't working, the responsible maintainer should be informed, but why post it here?

Those who are interested in such reports should read LKML. Let's keep slashdot for posts of general interest.

Re:Welcome to the immaturity of Linux

[dylan_-]

would be stomped out in any organization that had a financial investment in producing professional, streamlined output.

Actually, no. In an "organization that had a financial investment in producing professional, streamlined output" HyperDev would be dropped because the Head of Team SupraDev had recently become engaged to the boss's niece. Only half of Team HyperDev would be informed of this, and no-one in Team CyberDeX (which depends on the correct Dev) would be informed at all. Marketing would state that the product would be ready in three weeks, despite the fact that none of them even know what CyberDeX does.

CyberDeX would be released 3 months later and if you tried to use it with more than 2 users it would fall over. A patch would be released 6 weeks later which allowed you to have 3 users. The product would crash daily.

One year later CyberDeX would finally become something approaching stable with the release of Service Pack 4, which essentially rips out SupraDev and puts HyperDev back in.

The head of Team SupraDev gets promotion and a raise.