Submitted by Linus Torvalds has announced the release of version 2.6.30 of the Linux kernel. “I’m sure we’ve missed something, and I know we have some regressions pending. At the same time, we do need the coverage of a eral release, and on the whole it looks pretty good. We’ve fixed a few regressions in the last few days, and there’s always 2.6.30.x.” The list of changes is interesting.
A big new feature is the log-structured filesystem NILFS2, which we already discussed earlier on OSNews. NILFS2 also allows you to mount its snapshots in read-only fashion. There are more interesting filesystems in this release, such as preliminary support for NFS 4.1, EXOFS, and FS-Cache.
We also see work to make the kernel boot even faster by using fastboot, which scans storage devices in parallel to the rest of the kernel boot process. We can also add another architecture to the supported list: Microblaze.
The new kernel will find its way to your machine soon enough, but if you can’t wait, then you probably already know where to get the source and compile it. Have fun!
proofreading the announcement, for one
It is a kernel regression. The kernel isn’t fast enough to keep up with Linus’ typing speed.
What does kernel regression mean?
It means that a feature that worked in the last release will exhibit buggy or completely broken behaviour now.
its a nice way to say “bug”
(yes yes, its a specific kind of bug: feature that worked before stopped working, effectively “regressing”)
“bug” is too ambigious, regression is when functionality gets worse.
I belive in this case the topic is performance regression, which means that while new features was added or other modules/features got optimised one or more existing features got slower , eg uses more cpu cycles to achieve the same work.
Common regressions happens while eg. choosing to optimise towards either latency or throughput, you rarely get to satisfy all users in such a case, which could explain why just about every important module is hot-plugable in the Linux kernel (eg. schedulers optimized for servers vs desktops).
I really hope that Intel’s GEM is stable and giving out good performance.
It is, at least on 3 of my machines
With version 2.6.30?
Not even… I’m using .29.4, xf86-video-intel 2.7.1, xorg-server 1.6.1.901 and KMS+UXA+DRI2 on my netbook (eeepc 1000), a Dell Workstation and a Dell latitude D830 without issues.
This is since a month or two. Before that I often had crashes and freeze. I don’t see those now, I’m happy
It is not stable on my machine (toshiba U200 Intel GM945)
It is stable without KMS though.
I would like to see more things pushed to the user-space (ala uKernel). Anyway I believe they did their best again with this release.
Well, you’re in luck. You can personally change it. As linus likes to say, His tree of the kernel is his tree. If you disagree, create your own tree. All the major distros have their own trees that vary from Linus’ tree.
I know that would be a time consuming proposition for little real gain, but sometimes I think we have it too good in our vast world of open source software.
Feel free to express your desires, we need to hear all voices so that the majority of them can be served by the default behavior. Your particular request, most likely wouldn’t be best for most users so its unlikely to be adopted in a major tree.
Edit: my grammar filter is broken again. Sorry.
Edited 2009-06-10 19:51 UTC
Has 2.6.30 been released or are they preparing to release it … reason I ask is that at the time of writing, on kernel.org it shows the latest stable version of the Linux kernel as 2.6.29.4.
Update: looks like the website is lagging behind.
Edited 2009-06-10 19:24 UTC
Hi,
Why they don’t mention the new multitouch support ?
http://www.lii-enac.fr/en/projects/shareit/linux.html
Great release, the NILFS2 file system looks really great and interesting, I can’t wait for Btrfs to be stable though
.