|
|
Subscribe / Log in / New account

Stable kernel 2.6.16.9 released

Today's first stable 2.6.16 release is 2.6.16.9. It contains a fix for a potential inter-process information leak on the i386 and x86-64 architectures. (Update: the FreeBSD advisory has some additional information on this vulnerability, which is specific to certain AMD processors).

to post comments

Stable kernel 2.6.16.9 released

Posted Apr 19, 2006 16:58 UTC (Wed) by aleXXX (subscriber, #2742) [Link] (9 responses)

First post ! :-)

Well, every day a new kernel release, honestly, isn't that a bit much ?

OTOH, is there a schedule for 2.4.33 ? RC2 is already 2 months old.

Alex

Stable kernel 2.6.16.9 released

Posted Apr 19, 2006 17:33 UTC (Wed) by JoeF (guest, #4486) [Link] (1 responses)

Doesn't quite strike me as "stable" for a stable kernel...

Stable kernel 2.6.16.9 released

Posted Apr 19, 2006 17:38 UTC (Wed) by smitty_one_each (subscriber, #28989) [Link]

It's a horse thing. Got a whole barn of 'em 'round back. bdump-bump.

Stable kernel 2.6.16.9 released

Posted Apr 19, 2006 17:48 UTC (Wed) by dlang (guest, #313) [Link] (2 responses)

how long should they sit on known, published security holes before releasing an update to fix them?

personally I think that once a day is probably enough, anyone who cares more then that should subscribe to the -stable list and see the patches themselves, but beyond that what _is_ the right schedule?

right now they are doing a release each time they confirm the fix (and that the fix doesn't cause other problems), and the result is multiple releases per day. I agree that this seems a bit much.

but should they sit on the vunerability (or bug that causes crashes) for up to a day?, up to a week?, up to a month?

any value they pick will have people complaining, but what do people think is proper?

Stable kernel 2.6.16.9 released

Posted Apr 19, 2006 18:01 UTC (Wed) by dang (guest, #310) [Link] (1 responses)

For the sort of bug fixes that go into the stable tree, I have no problem with frequent releases. Strikes me as odd that people complain about timely bug fixes. Isn't that supposed to be one of the hallmarks of open source development?

Can't recall exactly, but I *think* role of the stable tree is almost more for developers, certainly the pressure to form it came from that direction. The linus tree is sort of a reference tree that linus doesn't recommend betting the enterprise on ( that is what distros are for ), the distros put out polished kernels but those kernels can lag behind or race ahead of the linus kernel, and developers asked for something that they could work with. Voila, sucker tree.

Either way, I really appreciate the work the maintainers and reviewers put in to the tree and all of the timely releases.

Stable kernel 2.6.16.9 released

Posted Apr 20, 2006 8:42 UTC (Thu) by Wol (subscriber, #4433) [Link]

The problem is that, at present, we don't have a stable and a development tree like we did in the past.

It's all very well saying "use your distro tree", but there are several distros (and it's a marketing point for Slack) that use vanilla "Linus" kernels.

Expecting distros to maintain a stable kernel is hurting the small guys who don't have the resources. That's why this plan to maintain one of the kernels as stable for a long while is so important.

Cheers,
Wol

Stable kernel 2.6.16.9 released

Posted Apr 19, 2006 18:39 UTC (Wed) by jimi (guest, #6655) [Link] (3 responses)

You don't want bugs and security problems to be fixed? Stable does not mean problems will not be fixed. Stable does not mean releases will not be made in quick succession. Stable means that APIs are not supposed to change. Stable means that your computer should not crash. Stable means that a program or interface that worked prior to an update will continue to work after the update (assuming that program did not use the system in ways that are specifically undesired). But as long as we are unable to see into the future, stable cannot ever mean "competely problem free".

Stable kernel 2.6.16.9 released

Posted Apr 19, 2006 21:13 UTC (Wed) by JoeF (guest, #4486) [Link] (2 responses)

Geez, I guess I stepped into a hornets' nest...
Of course, security issues need to be fixed. Non-critical bug fixes usually go into the next release.
However, having such a flurry of issues may indicate that something is wrong with the process. Maybe more testing before a new release?
As somebody pointed out, the 2.4 series doesn't have "a release a day". Now, the issues may not apply to the 2.4 kernel, I don't know that.

Stable kernel 2.6.16.9 released

Posted Apr 19, 2006 21:54 UTC (Wed) by xorbe (guest, #3165) [Link]

Releases is what causes the wide-spread testing... what'cha gonna do?
These minor bug fix updates are a great thing, I say. If you want them to
collate a weeks' worth of fixes, then just update once a week yourself!

Stable kernel 2.6.16.9 released

Posted Apr 20, 2006 2:27 UTC (Thu) by Ross (guest, #4065) [Link]

I think it indicates something was wrong with the old process. There's no evidence this level of security issue wasn't present before, and eventually patched. It was just that it took longer to get the fix, and when it came out you got a lot of other changes you probably didn't care about.


Copyright © 2006, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds