| |||||||||||||
|
| |||||||||||||
Stable kernels 2.6.24.6 and 2.6.25.1
The 2.6.24.6 and 2.6.25.1 stable kernel updates are out. They
both contain a number of important fixes, including some with known
security implications. The word is that, in the absence of a "major"
security problem, there will be no more 2.6.24 updates after this one; it's
time for people running mainline kernels to move on to 2.6.25.
(Log in to post comments)
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 2, 2008 1:04 UTC (Fri) by dldeborahl (guest, #51862) [Link] Due to the amount of changes between 2.6.24 and 2.6.25, the kernels really deserve a different versioning scheme, i.e. 2.6.25 should have been called 2.7.0, with the last digit reserved for security updates. At the current rate we will be stuck in version 2.6.xxx for the next 20 years.
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 2, 2008 2:05 UTC (Fri) by charris (subscriber, #13263) [Link] At the current rate we will be stuck in version 2.6.xxx for the next 20 years.I think that is the general idea.
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 2, 2008 6:26 UTC (Fri) by dambacher (subscriber, #1710) [Link] Hey - its only a name ! If Linus does not want to follow the other software marketing people and offer a new shiny mayor version of "the software" every now and then it's okay. We don't sell linux after all. And after all leaving out the first digit makes more sense to me .-)
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 2, 2008 14:51 UTC (Fri) by jabby (guest, #2648) [Link] I think after a while, when it becomes awkward and pointless to keep saying out loud "two point six point" at the beginning, people will just drop it from the spoken name of the release. It can still be in the official version string, but since "twenty-five point one" is easier to say than "two point six point twenty-five point one" by four syllables, once it becomes unambiguous (due to the gradual recession of the earlier major kernels--2.2 and 2.4) language will adapt to take advantage of this efficiency. The point is that there is no need to change the version numbering... spoken language will take care of it on its own.
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 2, 2008 22:28 UTC (Fri) by vlx (guest, #51883) [Link] I actually think its a pity that Linus goes this way. I think we are missing something like "stable" release like 2.4 was (and is). The interfaces are changing too fast in 2.6.x series so it's more problem to keep compatibility with kernel for external modules. And there is not just kernel but also userspace which would benefit from stable releases. Now you see there are things added and reverted during release of 2.6 series. And also for security, you are forced to update to new releases which bring uncompatibility. For 2.4 you got really stable base and you know there are just major bugs and security flaws fixed during releases (yes there were few backports). Now there is more overhead in maintaining 2.6 based kernel for vendors IMHO. Yes, there are pros like adding and testing new features is faster but cost is that we all are sometimes testing on our production boxes.
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 3, 2008 9:10 UTC (Sat) by Thue (subscriber, #14277) [Link] 2.6.16.x is still being maintained, and follows about the same rules as the old 2.4 series did. You can just use that...
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 3, 2008 13:46 UTC (Sat) by ewan (subscriber, #5533) [Link] Or there are things like the RHEL kernels for real long term unchanging stability; RHEL4's kernel is still maintained and is 2.6.9 based. This is one of the classic Free software advantages; you don't have to just take what you're given from on high, you can pick and choose something that suits you.
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 3, 2008 14:31 UTC (Sat) by vlx (guest, #51883) [Link] But I think generally it would be better to have something like stable series of kernels where are no changes to ABI and no new features which could potentially add more bugs (yes I know boring). Everyone lusting stability would just stick with those kernel series and not choose some mysterious kernel version like 2.6.16 for stability. I think it must be quite hard work to maintain this one chosen kernel and have every important fix incorporated. Really we have just no other option but to choose some vendor and his kernel if we prefer stability and want to have security fixes included.
People want "stable" kernels with new features - they are impossible Posted May 3, 2008 22:36 UTC (Sat) by khim (guest, #9252) [Link] The question as always arises: why the kernel developers should bother? No new features and no new drivers mean no new development - and then you can just pick any old version of kernel. New drivers do mean new features (not always but quite often) and then we don't have "stable" kernel anymore... And new programs eventually just refuse to play with old stable kernels (inotify becomes more and more essential, for example) so you are starting to backport new features too... No, stable kernels were never particularly good solutions - they are changing too (count number of patches in latest update of 2.6.9 kernel from RHEL4, for example) but don't benefit from wide kernel development community...
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 2, 2008 6:30 UTC (Fri) by xorbe (subscriber, #3165) [Link] I loosely follow lwn, kerneltrap, lkml... seems like 2.6.25 is a little worse than normal, more reported troubles than what I normally sense from the posts. Think I'll wait for 2.6.26.* stable series on this one, assuming there's not a tremendous churn + issues again.
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 2, 2008 10:33 UTC (Fri) by juhl (subscriber, #33245) [Link] The current plan is to stay with 2.6 for the forseable future and just keep releasing new 2.6.x kernels. There's no plan to create 2.7.x anytime soon (if ever).
2.7 is coming Posted May 2, 2008 15:37 UTC (Fri) by ccyoung (guest, #16340) [Link] 2.7 is the anticipated microkernel system
2.7 is coming Posted May 2, 2008 16:41 UTC (Fri) by MisterIO (subscriber, #36192) [Link] Anticipated by whom? Didn't Linus recently confirm his aversion to microkernels?
Tongue in cheek? Posted May 2, 2008 18:28 UTC (Fri) by AnswerGuy (guest, #1256) [Link] I suspect that the comment about microkernels was intended to amuse. JimD
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 2, 2008 19:33 UTC (Fri) by iabervon (subscriber, #722) [Link] Actually, 2.6.9 should have been called 3.0; changing the development process would have been the sensible point to change the major version and drop a digit. Calling these kernels 3.30.6 and 3.32.1 would match better with the historical numbering. But, since it wasn't done at that point, the next best thing is to rename the kernel from "Linux" to "Linux 2.6", and say that this is version 25.1 of Linux 2.6; after all, this works great for the makers of Java2SE1 (a.k.a "Java") version 6...
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 2, 2008 20:35 UTC (Fri) by proski (subscriber, #104) [Link] It doesn't work great. It's actually very confusing.Linux changes from 1.x to 2.x when it started supporting SMP and multiple platforms. I don't see any changes of that scale on the horizon.
Stable kernels 2.6.24.6 and 2.6.25.1 Posted May 3, 2008 8:40 UTC (Sat) by ahoogerhuis (subscriber, #4041) [Link] On the upside of staying with 2.6.x and keeping from seperating maintenance and development for a long time, when was the last time many of you saw the kind of major transition work having to be done to jump from New Shiny Distro 4.x to 5.x whenever there was a new stable version of the kernel? Most of that is now gone and the kernel is less of a major thing to transition with everything happening gradually. -A
|
|
||||||||||||