370
votes
|
|
404
0
34
|
|
|
|
|
Propose your solution
Attachments
No attachments.
Duplicates
Comments
|
flounder
wrote on the 24 Mar 08 at 16:25
|
|
|
|
This is not to suggest dropping the 2.6.24 development on the floor (it's too late in the cycle for that), but rather to get a 2.6.25 kernel into universe or as an alternate in base (like RT or SELinux support).
|
nand
(Brainstorm admin)
wrote on the 24 Mar 08 at 16:43
|
|
|
Then please edit your idea.
As you pointed out, it's too late in the cycle to drop the 2.6.24 by default.
|
|
|
|
Aren't the odd version numbers considered to be unstable?
|
|
AndrewC
wrote on the 24 Mar 08 at 18:08
|
|
|
|
@Ralf: Not any more, that was a long long time ago. I still don't see this happening though.
|
|
|
I would also like to see Kernel 2.6.25 in the repositories.
In Hardy-proposed or something.
|
|
jiu
wrote on the 24 Mar 08 at 18:21
|
|
|
|
can't that just be an incremental change done sometime later by package updating?
|
|
|
I want HOT stuff in hardy
+1
|
|
maltes
wrote on the 24 Mar 08 at 22:13
|
|
|
Ask Linux how long 2.6.25 is going to take. Hardy is LTS. And LTS was moved to June last time. Maybe this time again. If important software is delayed, then this is a good reason. And I do think the kernel is important enough.
RH pays lots of important kernel hackers. If 2.6.25 is going to be the base of a Fedora release, then Ubuntu might save a lot work on the kernel in the future by taking that route.
|
|
flounder
wrote on the 24 Mar 08 at 22:36
|
|
|
|
FWIW There will probably be at most one more RC for 2.6.25 before the final release. Excluding 2.6.23 & 2.6.24 which suffered from the addition of a new CPU scheduler Linus almost always releases after RC6 or RC7. Note Feisty was almost perfectly aligned with the _very_ stable 2.6.22 kernel. 2.6.24 is more stable today, but in the long run certain ABI/API changes in the 2.6.25 cycle might make it a better baseline for an LTS release. Ingo Molnar, Greg KH, and Jeff Garzik (RH staffers) all seem to have worked really hard to make the 2.6.25 kernel free of corner cases and ABI deficiencies relative to the old scheduler and driver infrastructure they maintain.
|
|
|
There is a kernel freeze in Hardy anyways.
I suggest we ship with 2.6.24, then have 2.6.25 in repository under hardy-proposed.
|
|
|
|
^^^ IMHO, the best variant!
|
|
csulok
wrote on the 25 Mar 08 at 11:20
|
|
|
|
an up to date kernel version in one of the repositories would very much appreciated
|
|
|
i've been playing about with the beta of hardy, thinking i could help with bug reporting. the trouble is, i haven't found any bugs yet. for this reason i think that hardy is going to be really good. anything that would rock the boat at this stage should be considered very carefully.
having said that, your list of advantages for the funky weasel kernel is very interesting, so it would be great if 2.6.25 were available in the repositories
|
|
artilec
wrote on the 26 Mar 08 at 13:41
|
|
|
|
Fedora 9 is in beta and will be using 2.6.25 kernel which as already mentioned has support for more wireless stuff. +1 for hardy delay and .25 implementation!!!! Lets keep up folks!!!!!!
|
|
dino99
wrote on the 29 Mar 08 at 10:28
|
|
|
Some distros already play with 2.6.25 ( opensuse, fedora, ...) and their final release are very close now.
I am testing several Os: hardy 2.6.24, suse 2.6.25, fedora 2.6.25 and i can say that 2.6.25 is a best one
with 2.6.25 we have better Hal performance
So, i hope Hardy coming with 2.6.25 as soon as possible.
Thanks
|
|
vicho
wrote on the 29 Mar 08 at 17:17
|
|
|
|
I think it's better to include it in 8.04.2 :)
|
|
px33
wrote on the 6 Apr 08 at 21:13
|
|
|
|
I think that 2.6.25 package should be shipped just as optional one in universe repo, until 2.6.26 comes. And 2.6.24 as default.
|
|
|
|
I would really like to have some option to install the 2.6.25 kernel when it is released, either in hardy-proposed or something similar. Installing 2.6.25 is currently the only method of getting stable working drivers for the Ralink rt61 chipset (rt61pci driver) and possibly other ralink chipsets.
|
|
epineda
wrote on the 9 Apr 08 at 16:28
|
|
|
|
And in 2.6.25 you would get working drivers for b43 rv01 and rev02 (Broadcom Corporation BCM4312 802.11a/b/g for example)
|
|
|
2.6.25 is critical to users of the intel 965.
When using 3d applications (blender,k3d,all/any first-person-shooter games) there is a complete system lockup which can only be fixed by pulling the plug.
This being in Hardy is critical.
http://bugs.freedesktop.org/show_bug.cgi?id=14937
|
|
|
|
While Hardy is already freezed, at least you should consider putting 2.6.25 in repositories after release.
|
|
timw06
wrote on the 19 Apr 08 at 19:00
|
|
|
|
Allowing Hardy to release with a kernel that causes systems that worked fine in Gutsy to hard-crash would be unforgivable, especially as it is a LTS release.
|
|
|
If we consider that this release will be installed in enterprise production desktops and servers, it is mandatory to have the latest and/or better kernel out there and the same for all the drivers that will be released after Hardy and most important applications.
Like it happen with Dapper if Canonical only does security updates and not new version updates not only for the kernel but also for drivers, firefox, openoffice, gimp, etc... then it's silly to maintain a distro for three and five years.
Think about it, how many people or companies use dapper today?
So it's very important to make available the 2.6.25 kernel, but also the new versions that as time goes by will be released if they improve the old one.
Remember, we are talking about a LONG TIME support release.
|
|
|
|
IMHO it's a very good idea. But it is also fundamental that all restricted drivers (NVIDIA in primis) work correctly.
|
|
yammosk
wrote on the 3 Jun 08 at 07:37
|
|
|
|
Kernel 2.6.25 needs to come in. Like that of many others my WLAN chip can't connect to encrypted networks on Hardy because of the whole "ndiswrapper and symbols"-dilemma. This is supposed to be fixed in 2.6.25. I can live without wireless temporarily, but not for three to five years.
|
|
|
Yes, please.
2.6.24 has very big problems leading to lockups and hard freezes on many computers.
For reference, see the long discussion on the forums at http://ohioloco.ubuntuforums.org/showthread.php?t=768200
or http://ph.ubuntuforums.com/showthread.php?s=f0a7e1e71fc8753bf83472f31604107e&t= 765510
or the many bug reports such as https://bugs.launchpad.net/ubuntu/+source/linux/+bug/204996 and the tens duplicate thereof.
2.6.25 is said not to have such issues.
Also, 2.6.25 is the fedora 9 kernel. Using the same kernel as fedora would mean that bugs can be fixed faster, because there is a larger userbase and no need to identify the precise bits causing problems to backport solutions to a former kernel. So, even if 2.6.25 could not fix right now the above bug, moving to it would be a win anyway.
Wasn't the Ubuntu founder the big supporter of "distribution syncronisation"? This is a situation where using the same kernel version could really help.
So, please make 2.6.25 available for hardy asap.
|
|
|
|
On a more frilly note, 2.6.25 supposedly includes working support for sensors on lots of machines (esp. lots of HP/Compaq notebooks.)
|
|
|
|
Can we close this? Hardy will not carry the 2.6.25 kernel.
|
|
|
ridiculous
Hardy is therefore a DEAD release.
|
|
edmon
wrote on the 6 Jul 08 at 20:10
|
|
|
maybe with next release of Ubuntu developer will make
much wider and wiser decision which kernel to use.
Current situations is a little bit like a own goal in soccer!
2.6.24 isn't solid for LTS release!
|
Post your comment
|
