LWN.net Logo

Stable kernel 2.6.1 released

Linus and Andrew have released the 2.6.1 kernel. There are very few changes since the last release candidate. The 2.6.1 kernel as a whole adds a few new features (MSI support, EFI support, a couple of internal API changes, etc.) but is mostly concerned with fixing problems found in 2.6.0. See the long-format changelog for the details.

Linus is headed off to Linux.Conf.Au for the week, but Andrew will apparently be around and dealing with patches. "The fact that I'm fleeing the country should in no way be construed as anything sinister at all, no siree."

Andrew, meanwhile, has released 2.6.1-mm1 which adds a big ALSA update, more kgdb support, an NFS update, the "kthread" abstraction (covered in this week's LWN Kernel Page), and a number of fixes.


(Log in to post comments)

Stable kernel 2.6.1 released

Posted Jan 9, 2004 16:10 UTC (Fri) by alspnost (subscriber, #2763) [Link]

I see that LWN is now officially putting its neck on the line and declaring this "stable"!

This accolade was not given to 2.6.0 itself, and I noted that this week's kernel page diplomatically refers to "the current 2.4" and "the current 2.6" kernels. ;-)

Anyway, I trust that the editors are now all running 2.6.1 on their systems and finding it wonderful. I only wish I could run 2.6 at home, but due to stupid Nvidia driver hassles, I seem to be stuck on poor old 2.4 for now.

"stable"

Posted Jan 9, 2004 16:16 UTC (Fri) by corbet (editor, #1) [Link]

Old habits die hard...:)

FWIW, however, I am running 2.6 on just about everything. I normally keep development kernels on my desktop unless they are really broken, but I've also just dropped it onto the laptop in preparation for the trip down under. Even suspend/resume works, which was a pleasant surprise.

Stable kernel 2.6.1 released

Posted Jan 9, 2004 17:20 UTC (Fri) by linuxa (guest, #18568) [Link]

I've been running 2.6.x with Nvidia drivers for quite some time using the patches from http://www.minion.de.

It's my ADMtek 8211 wireless card that gives problems.

Stable kernel 2.6.1 released

Posted Jan 9, 2004 19:21 UTC (Fri) by mmarq (guest, #2332) [Link]

If is not because of the legs is because of the trousers!...

I mean, wouldn't it be nice if "at Least" the major Hardware manufactors, would line up behind a major release of Linux, with "in-time" hardware support (technical papers or binary or source code)... more or less what happens in Windoze World.

I belive, everybody can agree that is not kernel developers fault;..., and it isnt really.... But if you think about the all Open-Source movement, it strikes as evidence that should be the major hardware manufactors leading the financing for commercial purpose of the movement, instead of IBM, Red Hat, and or Novel...

I mean Intel, AMD, Nvidea, ATI, SiS, VIA, CREATIVE, have much more money, and much much more to gain from defaulting at a OpenSource(Linux) software OS platform, than they do with the actual Windows paradigma that pressures and accelerates hardware prices down...

The only problem is that they are so affraid of losing money !!?...
PERHAPS THEY HAVENT BEEN APPROCHED THE RIGHT WAY BY OPEN-SOURCE LEADERS.

Stable kernel 2.6.1 released

Posted Jan 9, 2004 19:54 UTC (Fri) by xorbe (guest, #3165) [Link]

with what, wads of cash and incentives?

Stable kernel 2.6.1 released

Posted Jan 9, 2004 20:42 UTC (Fri) by sbergman27 (subscriber, #10767) [Link]

If you want a driver written for newer hardware, do it the way most things in open source get done. (And no, don't write it yourself!) When you buy your hardware, buy an extra and send it (anonymously) to a poor starving computer science undergrad who can't afford anything. Then lurk on the linux kernel mailing list. Within a couple of months at most (best to time your hardware purchases so as not to conflict with finals) you should see a patch submission with an explanation to the effect that 'I had this cool new video card that I wanted to use but it didn't have linux drivers so I reverse engineered the protocol and wrote this but I'm not using my real name because NV*a would probably have me arrested.'

One more tip from personal experience. I doesn't hurt to follow up the anonymous hardware shipment with an occasional anonymous pizza shipment. It's frustrating when they keel over in the middle of the project. You have to start all over again, and it's hard to even get your hardware back.

Stable kernel 2.6.1 released

Posted Jan 10, 2004 22:30 UTC (Sat) by mmarq (guest, #2332) [Link]

So,... its a good or stupid idea to have a split driver model ?

I mean, if you think of it, there is some part of the code that "always" could stay with that particular piece of hardware, unchanged or with minor tweaks unltil the end of useful live of it... while the kernel and above could change dramatically.

And a split driver model, with a some form of API/ABI that would cut tangencial to the hardware, i.e. would be hardened to not allow the "hardware half" code to expose much more than the BUS and those hardware registers...

This kind of protocol would:
a)facilitate debugging
b)facilitate reverse engeniring
C)facilitate portability to other OpenSource OSes (BSDs)
d)faciliate maintenance

I belive ALSA is "half" the way there, with the only big lacking of hardening.

IMO this could prove to be the deal that the hardware industry could not refuse.

Stable kernel 2.6.1 released

Posted Jan 9, 2004 20:29 UTC (Fri) by NAR (subscriber, #1313) [Link]

I mean Intel, AMD, Nvidea, ATI, SiS, VIA, CREATIVE, have much more money, and much much more to gain from defaulting at a OpenSource(Linux)

Are you sure? I think they have much more competition than e.g. IBM, and creating drivers for leading Linux distributions would at least double their "driver-creating" costs. And what would they gain? The Linux desktop market is tiny compared to the Windows desktop market.

Bye,NAR

Stable kernel 2.6.1 released

Posted Jan 9, 2004 20:50 UTC (Fri) by dlang (subscriber, #313) [Link]

you don't have to develop drivers for each distro, you just have to develop a driver that makes it into the core kernel and have someone maintain it there. if it's there it will end up in every distro without you having to do any extra work.

Stable kernel 2.6.1 released

Posted Jan 10, 2004 19:33 UTC (Sat) by XERC (guest, #14626) [Link]

Ahah, and how big is their "driver development cost"? It's strange, that
a few fellows can do it in a weekend or two and a big company can't afford to ask their employees to do the same at workdays? hmm, I thing that it's just
a matter of attitude and nothing else.

Regards,
XERC

Stable kernel 2.6.1 released

Posted Jan 9, 2004 22:54 UTC (Fri) by alspnost (subscriber, #2763) [Link]

Yeah, what sucks the most is Intel with its Centrino stuff. I mean, after how many months, their web site still says that wireless drivers for Linux are "under development". That's a really poor show.

BTW, are there any 802.11g wireless cards with proper, open source, in-kernel support under Linux? I wasn't aware of any, but if there is one, I'd love to know about it.

Prism drivers fairly advanced

Posted Jan 9, 2004 23:43 UTC (Fri) by goonie (guest, #4252) [Link]

Whilst they're not in the kernel yet, the Prism drivers project is pretty well advanced - the patches apply cleanly, work with the standard wireless tools, and worked straight away for me.

Stable kernel 2.6.1 released

Posted Jan 10, 2004 3:56 UTC (Sat) by chad.netzer (subscriber, #4257) [Link]

I couldn't get the newest Nvidia drivers to work with either my 5600 or 5700 based graphics boards, using a motherboard with a VIA KT600 chipset. I worked on it for several days, followed the forums, and tried EVERYTHING. So I went out and got an nForce 2 motherboard (ASUS A7N8X), and now the newest Nvidia drivers (minion.de versions) work fine under both 2.4 and 2.6.

I needed to use the newest nvidia drivers (which support the 5700), so if you are using a VIA (or other) chipset, and don't want to wait for possible fixes from Nvidia, etc., then consider testing with an nForce 2 motherboard to see if it solves your problem.

PS. The ASUS board comes with both the Nvidia ethernet and a 3Com ethernet port (I use the 3COM), and needs the Nvidia IDE support compiled in to use DMA on IDE (otherwise things are insanely slow)

Not a happy situation, on the whole, but hopefully the information might be useful.

Stable kernel 2.6.1 released

Posted Jan 10, 2004 7:45 UTC (Sat) by allesfresser (subscriber, #216) [Link]

I don't think it's that much of sticking out one's neck to use the word "stable" when the even-numbered series have been called that for many years. I have found that the 2.6 kernel codebase has been very stable, and I've been using it occasionally since around 2.5.50-something, and steadily since 2.6.0-test8. The kernel developers don't call something stable just on a whim... but by dubbing it "stable" they don't mean that it has no problems, and that it will never, ever fail, either.

Stable kernel 2.6.1 released

Posted Jan 10, 2004 21:21 UTC (Sat) by ken_i_m (guest, #4938) [Link]

It is beginning to look like Linus and Andrew are intent on changing the historical linux pattern of "stable" regarding early releases in such series. If this continues I will be willing to install an even numbered kernel much sooner on a production server then I would have been in the past.

Here's to truly stable kernel releases...

cheers,
ken

Copyright © 2004, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds