Slashdot Mirror
Kernel 2.6.1 Released
jnf writes "And so he said it is released, and then jumped on a plane to Australia. Linus announced the release of 2.6.1 a few minutes ago, fixes include AGPGART, a fork() bugfix, and misc changes to XFS, and those are just the patches applied since v2.6.1-rc3. Full changelog is avialable, kernel at the usual places, i held off posting this until kernel.org was updated." 2.6.0 is now in Debian unstable...
still doesn't build on cygwin..
... it gets released the minute I've installed 2.6.0.
:)
Perhaps I'll wait until at least 2.6.2 before doing it again
Linus Flees! The time of judgement is upon us. Oh, repent!
Besides Debian, What distros have 2.6.x ?
Does this release fix the do_mremap() exploit? I coulden't find it in the changelog. I got the impression from security sites that 2.6.0 had this bug.
We will know that it is time to use 2.6.x in anger when Patrick ships his distro with it as the default kernel. This is usually a sure sign that stability and maturity is upon us.
Stick Men
I download it, double click on the .exe, click next a few times and restart?
Thanks,
Clif
clifgriffin > blog
Well, you see, my BLT drive just went AWOL and I got this real big presentation due tomorrow for Mr. Kawsaki and if I don't get it in he'll ask me to commit Harry Carry. Could you read me the numbers on your version? It's the thing that you get when you type 'uname -a'
Oh wait, we are all about reality here. My mistake.
I'm going to be in Australia (and on airplanes) for the week, but we're
:-)
all in the capable hands of Andrew, so why worry? The fact that I'm
fleeing the country should in no way be construed as anything sinister at
all, no siree. Nope. I'm innocent, and nobody saw me do it.
Linus is not only a great project manager, system architect and coder, he's funny as hell too.
(If that isn't an underhanded slap in the back of the head of Dalek McBride, I don't know what is. "I'll be in Oz all week, try the veal!!")
I hope SCO sticks around for a while just for the comedy factor.
Soko
"Depression is merely anger without enthusiasm." - Anonymous
Native support for SATA hard drives!
I've been wanting to dual-boot for several months now, but the Linux installer (any distribution) does not recognize my SATA hard drive.
For an OS that's supposed to be innovative and cutting edge, Linux is really dropping the ball on this one!
Socialism: A feeling of discontent and resentment caused by a desire for the possessions or qualities of another.
To Sum it up:
/proc/net/tcp fixed
o lots of USB-Updates, eg. for storage-devices and BUGS
o seeking in
o some more use-after-free()-fixes
o [libata promise] fix another ugly bug (for those who use it)
o lots of misc small fixes
o lots of ARM stuff
o dvb: Update DVB core (and more stuff, for those video-people)
o Fix via686a/KX133 TSC failure (for ppl with an Abit KA7/KA7-100 etc)
o Fix memleak on execve failure (memleaks are always bad)
o cpuqfreq stuff/additions
o "at least" one important X86-64 fix
o mremap() security fix
- UML
- ipsec
- ebtables & bridge-netfilter
- robert love's preemptable patch
- LSM-hooks (which make not everybody happy:grsecurity, RSBAC
- LS-module SE-linux
- filesystem-encryption
- apci 2.5 backports
- Kernel
.config
- DVB-support
I got really tired of applying combinations of those patches to newest kernel source (due to security issues). They 're now all included to 2.6! Only MPPE-support seems still to lack.They must have beaten up Linus to get all those accepted ...
Cygwin is great but a full linux would be even better. In theory at least, User Mode Linux should be able to run under Windows. Possibly with a MinGW compile under Cygwin so after building, it doesn't need the Cygwin layer.
Slackware kernel 2.6.x compile/upgrade guide available here.
That's a heck of a lot of changes for a "stable" kernel.
Me: How many fingers do you have on your right hand? ...All right, I can see you're upset. How much would it take to clear this up? Patches? A syctl named after you? The head of Alan Cox?
Linus: What?
Me: Oh, how I have prepared for this moment. The coding, the studying, the kernel crashes, never seeing the sun...
Linus: What the hell are you talking about?
Me: My name is Saint Aardvark the Carpeted. You killed my kernel. Prepare to die.
Linus: How the hell did you find me? Did Darl send you?
Me: My name is Saint Aardvark the Carpeted. You killed my kernel. Prepare to die.
Linus:
Me: My name is Saint Aar--
Linus: Stop saying that! Guards!
Me: --killed my kernel.
Linus: What do you want?
Me: I want my -rc3 kernel back, you son of a bitch.
Carousel is a lie!
This is a bit off-topic, but:
I was an early user of Linux (1.2.8 and earlier w/ Slackware). ("Back in my day, we had to compile our own kernels!").
Anyway, I've been screwing around again lately. I've got two machines running Mandrake 8.2 and one w/ Mandrake 9.2 (VMWare actually). Also planning on messing w/ Redhat 9 and Suse. Knoppix rules, etc.
What I want to know is: What are the complications/problems with upgrading your kernel? I remember there being all sorts of problems with shared libraries versions since they don't have any internalized versioning system to run things side by side.
Is it still true that I might break half the apps running on my system if I try to update my kernel?
Please help to re-educate a guy who has lost his way.
Thanks.
Tom
---- It puts the lotion on its skin or else it gets the hose again. It does this whenever it's told.
Here.
"Res publica non dominetur"
Uhm, it's the second number that determines stable or development version. 2.6.0 and 2.6.1 are both stable releases.
That was the point of the whole 2.6.0 test series. The reason they did that was because it was likely that someone who wouldn't touch 2.5 with a ten foot pool would run 2.6 text X, and they did. 2.6.0 probably had more testing than any other kernel. I've been using it on two machines ince test1, the only problem I encountered was that DRI was broke in 1 release.
Have there been any articles or reviews comparing the performance of various kernel versions? I'd find it interesting how much progress has been made in areas such as network throughput, disk access, etc. I guess its possible there isn't any more room for progress in some areas, I would find that interesting too.
There are many benchmarks that illistrate how much better 2.6.0 is than 2.4. You can always boot to it and see for yourself and then switch back to 2.4 if you have problems. Just make your old kernel the second option in Grub/Lilo so you can go back and remove 2.6. You might be able to find more info on Kerneltrap.org
I'm not how how important network throughput and disk access are since they don't have much to do with the kernel and more to do with the network interface and filesystem respectively.
In looking at the changelog, I see lots of 2.4 fixes are being added into 2.6.1. I understand the kernel versions are completely parallel development paths. So does this mean there are lots of 2.4 bug fixes that are still not in 2.6? If so, I would think that might be something worth waiting for before upgrading.
I fail to see how 2.6.1-rc3 (rc == release candidate) is confusing. rc numbering is pretty standard, even Bill's boys do it (Windows 2003 rc-1 for example).
There's also the question of why exactly people new to Linux are compiling their own kernel rather than using that provided by their distro of course.
For even more clarity the ftp site now has the rc-whatever releases in a 'pre-releases' subdirectory, so I really don't see an issue here.
Blaming GW Bush for the Iraq war is like blaming Ronald McDonald for the poor quality of food.
This is a way to make the Windows 85% desktop share vanish overnight.
Just compile the most expansive possible kernel. Package it and "sell" it to cnet as the p2p app to have. Include boot loader.
No one reads warnings/lisenses anyway...
And voila! 85% linux on the world's desktops overnight! And what a night it will be!
I pity Dell support and the Indians....
/. Where the truth
Use APT and use Arjan's RPM repository. Cheers.
Well I've been poking around in the kernel for years now. Mostly just trying things other people have told me to do to fix what ever problem I was having. Then I'd say, "yeah, that fixed it", and the author of said code would submit the patch.
This time, I attempted to do the same. But the author didn't tell me much of what to do at all. So I just started looking at the one function he pointed me to. I ended up surprising myself. I found I could easily follow what was going on, and quickly found my problem. I tried a fix, and it worked. I reported back to the author, that I fixed my problem and how, and he asked me to submit a patch to Linus.
I've used to think of the kernel as some beast, full of black magic. Some of the parts dealing with broken hardware, are a little arcane. But the more I look at it, the more I see that most of it is just C. Now that Linus is subscribed to the linux-kernel mailing list, I see more developers interacting with him. He really does have good taste in code.
Recently Linus has shown his dislike for the scsi emulator for ide devices. He went on to say that there werent any common devices that needed the scsi layer. One such device are zip drives. I use a Zip 750 in my server to backup a small but important set of data. The only way this drive will work is with the scsi emulated layer. Has this been fixed in the 2.6 kernel series (it wasnt fixed in 2.6 pre4), or does someone have another way to use this device without scsi?
Tis better to be silent and thought a fool, than to open your mouth and remove all doubt --Abraham Lincoln
When I went from 2.4.xx t 2.6.0-testxx (on a Gentoo 1.4 system) I downloaded the 2.6.x kernel and checked in Documentation\Changes. That file will list several packages and the minimum version needed. It also has the command to check the version and the site to download updated packages. Once you have verified that you have the correct versions of extra software compile the new 2.6.x kernel. Boot it and see what breaks. Of course you want to keep a backup of your current working 2.4.x kernel to boot.
As for breaking half your apps: no. I built my Gentoo system under a 2.4.x kernel and now run a 2.6.x kernel with no problems.
the_crowbar
Have you read the Moderator Guidelines
Been playing with 2.6 since test9 and been upgrading the kernel since.
They keep on top of things with Gentoo.
"Music is everybody's possession. It's only publishers who think that people own it." - John Lennon.
It wont break your system, but a few subsystems change:
* Module loading: There are new tools for this (usually called module init tools). These are MANDATORY.
* Logical Volume Management: lvm2 is available but possibly not required (not sure on this)
* Alsa: Can now be compiled in the kernel. Might need minuscule tweaking
* A few modules have been renamed. (e.g. printer.o -> usblp.ko)
Divide et impera!
I have 2 Broadcom NICs in my new server. They work just fine, haven't had a problem with them.
From what I remeber, there are 2 drivers. The ones from Broadcome, which are a bitch to install and have poor performance, and the other thats in the kernel it self. Its listed as tg3. I am still running 2.4 kernel on that server, so I imagine 2.6 should work fine.
Stock Debian kernels don't appear to have this NIC compiled in, I can't figure out for the life of me why not.
*shrug*
until (succeed) try { again(); }
I have been using test9, and have been having a hell of a time with my PS2 mouse. On my system, with 2.6 kernels, whenever there is any load, the mouse becomes very erratic bouncing about the screen and clicking on things at random. I've tried as many of the "fixes" as I can find, but none have done the trick. I assume it is limited to sometihng unique about my system (KT600 mobo), otherwise people would raising a big stink. Have I missed the fix somewhere>
I just downloaded the tarball, modified my 2.6.0 configuration a little bit (what the heck do I need UFS support for?), compiled and rebooted. Without any tweaking, ALSA, LVM2, ide-scsi (in your face, Torvalds!), the Promise IDE-Controller and all the usual tidbits work. Nice one.
BTW, does anybody know how to make K3B understand the new ATAPI cdburning stuff?
Divide et impera!
Yep. Go to this site and download the diff for your version of the driver.
You'll need to run the NVIDIA installer with the --extract-only argument to untar it, then cd usr/src/nv and patch -p1 the diff file and then cp Makefile.nvidia Makefile. Then just run make install in the top-level directory of the nvidia installer and it'll build and install a 2.6.1-compatible module.
You sir are what Linux is all about :)
Thank you for fixing our code and making it a little more stable for us all. Hopefully your comments will spur others to have a peek under the hood and see what they can discover.
liqbase
I know many people will not read the documentation so I'm posting it here.
You need module-init-tools with the 2.6.x series.
<torvalds@home.osdl.org>
Fix silly mremap test.
Get off the drugs, Linus.
Under 2.6.x, you simply type:
make (menu|g)config
make
make install
make modules_install
"make install" tries to figure out whether you're using LILO or GRUB and tell you what to do next, though it didn't quite work in my case since I never bothered setting up a boot menu (I just use the GRUB boot prompt). Another thing you should watch is that, by default, you can't remove modules from a running kernel. Be sure to check out the options for this.
Anyone else notice that you don't see the actual gcc commands anymore? Compiling Linux now looks eerily similar to compiling DJB's software.
Hopefully my HPT370 chipset will work under 2.6.1; it locks up 2.6.0 at boot unless I disable it in the BIOS (or don't compile in support for it). But it does "feel faster" than 2.4.x on the desktop, and ALSA and my nVidia card worked like a charm (thanks minion.de).
WMBC freeform/independent online radio.
apart from upgrading and watching tentacle-porn I mean.
Don't know what "tentecle-porn" is...but sounds interesting.
I work all the time on my machine. To upgrade a kernel takes all of 5 minutes...including boot time. And I've only upgraded from test9 like 3 times in the last 2 months...so 15 minutes to upgrade per 2 months. Wow, yes, I can see where you would think I would not have any time left to do anything.
Of course, you also could just be an anonymous coward troll...nah.
"Music is everybody's possession. It's only publishers who think that people own it." - John Lennon.
AFAIK, the SATA command set uses SCSI commands, not ATA, SATA is much more SCSI-over-serial than ATA-over-serial. The drives are ATA-like in quality, performance, and features, but the commands they respond to are SCSI-based in nature.
FireWire, USB, and ATAPI also implement SCSI commands, ATAPI implements SCSI commands -OVER- ATA wires.
What I'd like to see is an abstraction of the SCSI-over-[anything] idea, so new drivers are basically just cutting up an input stream for their respective mediums. ATA as a whole could be implemented as part of this, the drivers would just say that your ATA drive is a SCSI drive, on an ATA bus, with a command set capable of features X,Y, and Z. It would make it a lot easier to implement TCQ and other stuff on ATA drives, and pre-ordered queing on dirty write buffers for slower serial devices.
All storage should be based on the most capable and broad command set, with lower-level drivers disabling features (fom said command set) to their needs.
"Sometimes, I think Trent just needs a cup of hot chocolate and a blankie." -Tori Amos on Nine Inch Nails
Have you tried Solitaire?