Weekly edition	Kernel	Security	Distributions	Contact Us	Search
Archives	Calendar	Subscribe	Write for LWN	LWN.net FAQ	Sponsors

Some development model notes

Posted Oct 28, 2004 10:56 UTC (Thu) by zooko (subscriber, #2589)
Parent article: Some development model notes

I can't complain about 2.6 giving me problems, but that's because I haven't tried 2.6 yet. I'm waiting until it stabilizes in the sense that it has received significant "test and debug" attention from developers while simultaneously receiving zero "add features and re-organize" attention.

Saying "We haven't heard users complaining about it breaking, even though we just added a whole bunch of new and rewritten code." does not persuade me that I should take the risk of upgrading from my reliable old 2.4 kernel.

Linux Counter Project's kernel version count

(Log in to post comments)

Some development model notes

Posted Oct 28, 2004 13:11 UTC (Thu) by iabervon (subscriber, #722) [Link]

I've been running 2.6 on 2 machines for quite some time (one since 2.6.0, one since ~April), and haven't had any problems with its stability. In fact, the only kernel bug I've encountered was on another machine running 2.4.26, and I identified it as a kernel issue in part by finding that it didn't happen in 2.6.4 (evidentally, it was fixed before 2.6.0, unless the Changelog lines were not informative); I applied a patch from 2.4.27, which seemed to be a backport, to fix it. So it looks to me like 2.4 is essentially just stale at this point, and that 2.6 has more effective debugging.

I do agree that the mainline unmarked releases have not been thoroughly tested upon release. It would be nice if Linus would delegate assigning version numbers to releases to someone else. But using an unmarked release after it has been around for a little while is quite reliable.

Some development model notes

Posted Oct 29, 2004 18:39 UTC (Fri) by NAR (subscriber, #1313) [Link]

and haven't had any problems with its stability.

Feel yourself lucky. umount crashes my 2.6.9 kernel about 50% of the time at shutdown. OK, the machine is supposed to be turned off after that anyway, so it's not a big problem but it could be really annoying with an ext2 filesystem fsck'ing at next boot (fortunately I use ext3). I really should google about this problem...

Bye,NAR

Some development model notes

Posted Oct 28, 2004 18:10 UTC (Thu) by tjc (subscriber, #137) [Link]

Thanks for the Linux Counter kernel link. It's interesting that over 20% of the systems counted are running Debian 2.4.18-bf2.4! Time for an upgrade...