A brief discussion on the lkml looked at which 2.6 Linux kernel patchset offers the best stability. Only four patchsets were mentioned, the bleeding edge -bk snapshots, Andrew Morton [interview]'s -mm patchset [story], Con Kolivas [interview]'s -ck patchset [forum], and Alan Cox [interview]'s -ac patchset [story]. When it was suggested that the -ac patchset was the most stable, and that -mm might be too bleeding edge, Alan offered some insight:

"2.6.x-mm [forum] is more like some of the work the old 2.4-ac [forum] did in merging new stuff (its also worth noting that 2.4-ac ended up more stable than 2.4 at times so -mm might be stable)

"The -ac tree is trying to be fairly conservative. When I merge stuff that is a little less conservative because it has to be done then I've tried to put a note in the relnotes for that release warning people its more testing grade.

From: Maciej Soltysiak [email blocked]
To: [email blocked]
Subject: 2.6 flavours
Date: 	Thu, 16 Dec 2004 21:13:56 +0100

Hi,

AFAICS the -ac tree should be the most stable of all kernels, right?

-mm is totally bleeding edge
-bk the same
-ck is experimental

Others are experimental too.

Looking at the changelogs, the most reasonable kernel to use for
generic use are the -ac kernels, which I am going to use since 2.6.10
as long as Alan is kindly going to continue his fabulous work.

I swear not to use 2.6.10 until Alan publishes 2.6.10-ac1 :-)

Regards,
Maciej



From: Jurriaan [email blocked]
Subject: Re: 2.6 flavours
Date: 	Thu, 16 Dec 2004 21:37:40 +0100

^gt; From: Maciej Soltysiak:
> Hi,
> 
> AFAICS the -ac tree should be the most stable of all kernels, right?
> 
> -mm is totally bleeding edge

I don't agree there. It is bleeding edge, but Andrew makes a conscious
decision when to release it, and I'm sure part of that decision is
thinking about when things will be stable enough to work for many
people.

> -bk the same

bk is the central repository, and depending on when you pull it, you may
have just catched Linus asleep after merging patch 1, meaning the tree
is unstable until he wakes up and merges patch 2.
There is no conscious 'release moment' - this is totally bleeding edge.

> -ck is experimental

There are 'release moments' here too.

> 
> Others are experimental too.
> 
> Looking at the changelogs, the most reasonable kernel to use for
> generic use are the -ac kernels, which I am going to use since 2.6.10
> as long as Alan is kindly going to continue his fabulous work.

I've understood the 2.6.x-ac kernels started with some ide work, then
included some serial fixes, and may or may not have other bug fixes.
>From what I read, they are not as all-including as the 2.4.x-ac kernels
were. Those I recognize most in the 2.6.x-mm kernels.

> 
> I swear not to use 2.6.10 until Alan publishes 2.6.10-ac1 :-)
>
Whatever you think best, of course. That may be the release where Alan
says 'Here's the new, experimental next-generation SATA code. It'll
probably break every partition you have. Send me bug-reports' :-)

My way of keeping my home system up is to test a new kernel first on my
laptop (which has good backups and little configuration), then read this
list for some days and then install it on my main workstation (which
also has good backups). Only after some weeks quiet I think about such a
kernel on my firewall/router. Keeping backups helps when testing
kernels.

YMMV,
Jurriaan
-- 
I never think, sir. Didn't get a degree.
	Chief Inspector Morse
Debian (Unstable) GNU/Linux 2.6.10-rc3 2x6078 bogomips load 0.23



From: Alan Cox [email blocked]
Subject: Re: 2.6 flavours
Date: 	Thu, 16 Dec 2004 19:54:54 +0000

On Iau, 2004-12-16 at 20:37, Jurriaan wrote:
> I've understood the 2.6.x-ac kernels started with some ide work, then
> included some serial fixes, and may or may not have other bug fixes.
> >From what I read, they are not as all-including as the 2.4.x-ac kernels
> were. Those I recognize most in the 2.6.x-mm kernels.

2.6.x-mm is more like some of the work the old 2.4-ac did in merging new
stuff (its also worth noting that 2.4-ac ended up more stable than 2.4
at times so -mm might be stable)

The -ac tree is trying to be fairly conservative. When I merge stuff
that is a little less conservative because it has to be done then I've
tried to put a note in the relnotes for that release warning people its
more testing grade.

> Whatever you think best, of course. That may be the release where Alan
> says 'Here's the new, experimental next-generation SATA code. It'll
> probably break every partition you have. Send me bug-reports' :-)

That would be Jeff 8)



From: Con Kolivas [email blocked]
Subject: Re: 2.6 flavours
Date: 	Fri, 17 Dec 2004 13:05:25 +1100

Jurriaan wrote:
> From: Maciej Soltysiak [email blocked]
>>-ck is experimental
> 
> 
> There are 'release moments' here too.

I keep releasing versions till it is stable. eg 2.6.9-ck3 is more stable 
than 2.6.9-ck1. 2.6.8.1-ck9 is more stable than 2.6.8.1 (for my users at 
least).

Cheers,
Con

• Archive of above thread
• KernelTrap interview with Alan Cox
• KernelTrap interview with Con Kolivas