Linux kernel 2.6.8 is now available (42 MB). Changelog is available here. Update: A patch (2.6.8.1) that fixes an NFS bug (a typo) has been released. Users who use NFS should upgrade.
Linux Kernel 2.6.8.1 is Out
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Just great news!
woot!
EOM
I still don’t understand why people link to a direct URL, instead of the mirror list , unless you specifically love the so-called “slashdot” effect….
http://www.kernel.org/mirrors/
So go and get it, now (when the mirrors are synced, that is
!
kernel.org: “ISC has upgraded our outbound connection to 1000 Mbit/s. Thanks!”
So mirrors are not necessary.
1Gbit is not that much for the world…
I agree 1Gbit can be overrun and latency is improved when
downloading from a close sever. But so far 1Gbit seems fine. Best not a store the eggs in one basket
At least link to the .bz2 version, which is a few MB smaller!
I am not sure what ‘EOM’ means. But a quick look through
the change log might help, plus kernel trap normally highlights the best improvements.
With in the hour in will be running it and will report back here.
Is anyone els running this version please tell us now it goes.
Off topic but EOM means end of message.
You know, I think Linux is a great kernel, but a 42 MB download is really a bit too much for my liking. Much of that is code for hardware that I don’t have or features that I don’t want. I am a great advocate of modularity, and I would like to see it applied not only to the compiled kernel, but also to the sources. I am aware that this will add some administrative overhead, but it could save a lot of traffic and CPU time.
Here are some ideas:
– Split the distribution in a base that has the common stuff, and optional add-ons for lesser-used network devices, filesystems, etc. etc.
– Employ a BSD ports like system that downloads the sources on request (i.e. when compilation of some part is requested)
– Distribute only the configuration interface, and download only the parts actually needed based on the configuration selected.
I am too occupied now to come up with a proper proposal, but I hope this will set some people thinking.
I think you’ll find that the only thing keeping people from splitting up the sources is the convenience of one tarball. The work on the 2.6 build system means it’s really easy to build modules from a seperate source tree, so that’s not a limitation anymore.
What about the fact that some people want to compile things directly into the kernel instead of as modules?
Warning IMHO….
I disagree it is good to have one source and when you configcan pick what you want. But with other apps it is a good idear.
“What about the fact that some people want to compile things directly into the kernel instead of as modules?”
Yes, what about it? In either case, they will have to fetch the source for that part, but won’t have any use for the things they don’t compile at all.
If 42MB is too much, get the patch then
Anybody got informative details rather than the changelog
Yea ok, maybe it would be a good idea. (Actually, it might even be a great idea
I’m not sure, however, because you would need that ports system to always be available when building a kernel. Hm, to solve that, maybe a tarball could be provided that included all the parts, and the config script knew how to get parts from a local pool?
Hm, to solve that, maybe a tarball could be provided that included all the parts, and the config script knew how to get parts from a local pool?
—
something like this was covered in kerneltrap.org. dont remember the name but scripts exist for sure
Get the patch or the bz2 version.
.bz2 vs .gz: 34MB vs 42MB
ftp://ftp.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.8.1
Wow! No kidding! Must have been a big screwup. Someones face is red.
I had trouble getting DVB to work, please note that the device node major number has changed for DVB from 250 to 212. Then all works fine!
Is that an update to the change log or the kernel and change log ?
Summary of changes from v2.6.8 to v2.6.8.1
============================================
<[email protected]>
[PATCH] Fix NFS client screw-up in fcntl f_op removal
Fix stupid thinkos in the fcntl f_op removal code.
<[email protected]>
Linux 2.6.8.1
———————————————–
when was the last “screw-up” ? 😮
It looks like they forgot to include a one-liner fix for a NFS oops. So, if you use NFS, you should use 2.6.8.1, if not, you don’t really care…
WRT to the “the kernel is too big” problem: You should use vendors kernels (you don’t care that much if you run gnome 2.6.1 or 2.6.2 either) or remove all the cruft not needed and put it onlien for the people with dialup
I don’t understand: It goes through -RC1, couple of weeks testing, fixes, -RC2, couple of weeks testing, fixes, -RC3, couple of weeks testing, fixes, -RC4, couple of weeks testing, fixes, and then there’s suddenly a critical fix that wasn’t found before? How about testing for some time, and then “-RC4 was released with no changes” procedure, like it’s done for older kernels? Does stability of the source tree not matter for 2.6 series or what?
They only human. The linux kernel is coded by robots (yet)
oops, should be “They’re”
The linux kernel is coded by robots (yet)
After seeing the press photos of Linus I’m not so sure :>
So what they screwed up, how long before it was fixed ?
Thats all that needs to be said.
…and it should be “isn’t” of course.
My grammar is crap in the morning
Um.. Is the new kernel 2.6.8.1 supposed to remember my .config settings from the 2.6.7 kernel? I did a make menuconfig and sure enough it has all the settings from my old kernel. I’ve never seen this before, I always had to go back and manually select stuff.
Done properly, there may be a business opportunity.
If you could boot a secure, trusted CD that made an SSH connection, passed your hardware configuration, and validated all of the drivers and stuff on your computer, would it be interesting?
Admittedly, I prefer the sheer do-it-yourself masochist approach, but something along these lines might prove successful…
Hi there.
When you do the menu config it will grab your old kernel’s settings from /boot/config-2.6.7 you may want to have a quick look around and then exit and pick save. Then you should be ready to do the make.
You can use “make oldconfig” to update your config file:
cd /usr/src
cp linux-2.6.7/.config linux-2.6.8.1/.config
cd linux-2.6.8.1
make oldconfig
Greets,
Alex
@Aaron: Thanks, I don’t remember putting my config-2.6.7 in /boot, I must have done it without thinking. Still didnt’ know the kernel looks for that file, I always thought you had to put the .config into the new kernel directory.
@Alex: Yes, I’m familiar with the make oldconfig command, just found it odd that my config settings were there when I didn’t type make oldconfig. I used the make menuconfig and had the same results.
Hi there
When you install the kernel by RPM or make install it puts a config file there in most cases. If you do put a .config file into the kernel directory it will default to that. Tell us how 2.6.8.x goes.
very cool
Actually, you can configure it to use the config of the running kernel. run make gconfig (or xconfig) and choose the “Kernel .config support” and also check the “Enable access to .config through /proc/config.gz”.
This way, you use the configuration of your running kernel. Then do a “make oldconfig” to make sure there are no new options in the kernel which need to be taken care of. A .config is placed there with all the options in the running kernel. You can tweak further if you have to. easy enough.
Isn’t this just a love fest of computer nerdery?
Yes, it’s a love fest like the one with the SP2 for Windows…
FYI if you’re reading this website you’re already a nerd! 😉
As a newbie,
how does it compare to Windows2003?
– Does it thread better?
– Can it handle more load?
– More secure?
– Is it cheaper? — Yes. — I can answer that.
Hopefully, they’ll have had the presence of mind to add that mishap into their regression testing suite (automated, I’d hope!) so that mistake is never made again
(ie, don’t send out a fix too soon that stands a remote chance of needing to be made again)
As a newbie, how does it compare to Windows2003?
– Does it thread better?
– Can it handle more load?
– More secure?
– Is it cheaper? — Yes. — I can answer that.
According to an animated advertisement by Microsoft I just saw, it’s half as fast and 10x more expensive.
Anyone else will tell you it’s a few percent faster, more load, etc.(depending on usage specifics) and much cheaper. Threads better, all that.
I think it’s easier to make secure. But my only linux security experience was a few years ago with Red Hat boxes, which btw were very insecure in the default setup.
Runs faster, but takes an expert to realy make it secure. While windows 2003 is quite secure by default, but not as secure if youre a good linux admin.
Well, if you are a competent Windows admin 😉
http://jalla.no-ip.org/kde/kde33.png
🙂
Howdy
May i ask what theme do you have in that screen shot ?
When does 2.6.8.1.3.4b come out?
I think Bjorn is using the “Plastik” theme with the Noia KDE icons and a gradient background for the taskbar.
I seem to remember Xandros has that theme by default (Plastik).
You can find it all at http://www.kde-look.org
Is it just me, or doesn’t make menuconfig work in the new kernel sources?
I seems to be working well with 2.6.8 and 2.6.8.1
What error message is coming up ?
Have you got ncurses-devel?
Yes, I have ncurses headers.
The output I get is this:
scripts/kconfig/mconf arch/i386/Kconfig
#
# using defaults found in arch/i386/defconfig
#
arch/i386/defconfig:175: trying to assign nonexistent symbol PCI_USE_VECTOR
arch/i386/defconfig:251: trying to assign nonexistent symbol BLK_DEV_CARMEL
arch/i386/defconfig:569: trying to assign nonexistent symbol NET_FASTROUTE
arch/i386/defconfig:1247: trying to assign nonexistent symbol X86_STD_RESOURCES
interrupted(11)
make[1]: *** [menuconfig] Error 1
make: *** [menuconfig] Error 2
The four nonexistent symbol messages also appear with xconfig and gconfig, which both work, so I guess the only relevant information here is interrupted(11).
Hi there,
What distro are you using ?
[quote]I think Bjorn is using the “Plastik” theme with the Noia KDE icons and a gradient background for the taskbar.[/quote]
Yes.
but when well kernel 2.7 be out ?
or is this a beta kernel?
http://developers.slashdot.org/article.pl?sid=04/07/22/0138244&tid=…
“but when well kernel 2.7 be out ?”
See http://slashdot.org/developers/04/07/22/0138244.shtml
Hi all,
I have just installed kernel 2.6.8.1 on my rig.
Everything is fine, except for burning cd’s as user. K3b says their is no burning device, as well as the selecting of burning device being grayed out.
BUT if I log in as root, I can burn cd’s.
SO I am sure it is a kind of permissions problem.
Any help would be great
Regards
Robert
Install submount
Already installed
Any other ideas?
Sorry I submit instead of preview.
Submount makes a kernel module called subfs. It is needed
to nicely mount and unmount cd’s and floppy and digital or
external devices. When you install the new kernel it is
not getting loaded. You need to make and make install
subfs to make it work again.
Check the permissions in /dev/scd0 or whatever the device is. Check the owner, group, and check the other permissions.
I suggest to chgrp the device to the group you are member of (e.g. “burn” or “cdrom2”) and then chmod it to 770 OR to chown it to your username and then chmod it to 700. Remember, root can still burn anyway, but you don’t want to give permission to other hence xx0.
It seems that many other people are having the same problem. It looks like 2.6.8.1 broke cdrecord.
More details on k3b’s web site under feedback.
Adding my username to the group “disk” fixed the cd problem for me. I didn’t see this issue come up until the upgrade to KDE 3.3 today, I had kernel 2.6.8 yesterday and had no problem. I don’t think the kernel broke this.