Linux Kernel 2.6.7 Released

conrausch writes "German Heise News reports among others that the new Linux Kernel 2.6.7 was just released, and that it fixes the previously mentioned bug in the floating point exception handling. Whether or not you offer shell access to other people, get it now from kernel.org or one of the mirrors."

Comments

Hopefully

[Anonymous Coward]

nForce2 chipsets aren't an arse with it now.

(This is a damn early post for me).

NVidia?

[xhorder]

Does the NVidia driver work with it?

Re:NVidia?

[Anonymous Coward]

why wouldnt it ?

(but to answer your question, its working fine on my box)

Re:NVidia?

[CyBaer]

It works.

Re:NVidia?

[Anonymous Coward]

I think he means 4496, which dont work with it

Re:NVidia?

[Anonymous Coward]

Something about the latest versions of the kernel has changed in such a way that there was a problem with earlier versions of the nvidia kernel driver

But it's not a problem as this was fixed with the later versions of the Nvidia driver
in my case I just had to use a later version within gentoo
nvidia-kernel-1.0.5336-r3.ebuild

Re:NVidia?

[misleb]

I had problems with the NVIDIA driver and kernel 2.6.6.

-matthew

Re:NVidia?

[Anonymous Coward]

Then you must be stupid RTFM!

Re:NVidia?

[Halthar]

Were you using a precompiled kernel? On the one machine I have 2.6.6 on at the moment, my NVidia drivers work fine, but I made sure not to use the new stack size when configuring it. If it was precompiled then you may have hit the stack size problem. I didn't see anything in the Changelog mentioning that for 2.6.7 they made the stack size change permanent (but I only skimmed it), so it is probably still an option. Admittedly I haven't had a chance to download 2.6.7 and check yet, so I may be wrong.

Woohoo, another kernel compile.

[Gabrill]

On your mark! Get set! Make!

Re:Woohoo, another kernel compile.

[Anonymous Coward]

wget http://www.kernel.org/pub/linux/kernel/v2.6/patch- 2.6.7.bz2
cp patch-2.6.7.bz2 /usr/src
cd /usr/src
mv linux-2.6.6 linux-2.6.7
bunzip2 patch-2.6.7.bz2
cd linux-2.6.7
cp .config ../config.saved
make mrproper
cat ../patch-2.6.7.bz2 | patch -p1 #(or maybe -p0)
make mrproper
cp ../config.saved ./.config
make oldconfig
make && make clean modules modules_install install
vi /boot/grub/grub.config # check it make sure it's right
reboot

Not to difficult. Could make it into a script, to bad I'm going to loose my uptime to patch the kernel, but oh well. Shit happens.

Re:Woohoo, another kernel compile.

[smoking2000]

Here they go into the first round: make oldconfig.

Some sort slowdowns here and there but they're coming into the second round none the less: make modules modules_install bzImage.

Their CPU's are running full speed and they're almost at the finish, the excitement!

There we have it folks, 2 rounds and the AMD XP's are doing great, followed by the Intel P4's.

Now comes the most feared part for all the contestants: mv System.map /boot && mv arch/i386/boot/bzImage /boot/vmlinuz-2.6.7 && update-grub && reboot

The crowd is holding its breath while the contestants approach the finish line. Owww, 1 P4 crashed with a kernel panic just before the finishline, what a shame!
The Athlon XP's crossed the finishline and were awarded with a beautifull new kernel! The Intel P4's came in second place and got the beautifull new kernel anyway!

Tune in next week for more kernel sports, when we will visit the senior compile contest between a 486sx and a 486dx

Re:Woohoo, another kernel compile.

[Anonymous Coward]

Oops. line cat ../patch-2.6.7.bz2 | patch -p1

should be

cat ../patch-2.6.7 | patch -p1

or something like that.

Re:Woohoo, another kernel compile.

[Short Circuit]

Wait! The P4 is crying foul! Someone ran SETI@home as a background process!

Re:Woohoo, another kernel compile.

[Taurim]

Kernel 2.6.7 is the solution for this situation. It integrates the very useful following patch from Con Kolivas :

[PATCH] sched: SMT niceness handling

From: Con Kolivas

This patch provides full per-package priority support for SMT processors (aka pentium4 hyperthreading) when combined with CONFIG_SCHED_SMT.

It maintains cpu percentage distribution within each physical cpu package by limiting the time a lower priority task can run on a sibling cpu concurrently with a higher priority task.

It introduces a new flag into the scheduler domain
unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */

This is empirically set to 15% for pentium4 at the moment and can be modified to support different values dynamically as newer processors come out with improved SMT performance. It should not matter how many siblings there are.

How it works is it compares tasks running on sibling cpus and when a lower static priority task is running it will delay it till high_priority_timeslice * (100 - per_cpu_gain) / 100
eg. a nice 19 task (Note : Seti@Home for example :-) ) timeslice is 10ms and nice 0 timeslice is 102ms On vanilla the nice 0 task runs on one logical cpu while the nice 19 task runs unabated on the other logical cpu. With smtnice the nice 0 runs on one logical cpu for 102ms and the nice 19 sleeps till the nice 0 task has 12ms remaining and then will schedule.

Re:Woohoo, another kernel compile.

[Short Circuit]

For HT, doesn't the process on the second logical CPU get blocked until the resources it needs on the first logical CPU become available?

If so, wouldn't it make more sense to put higher-priority tasks on the first logical CPU, and lower-priority tasks on the second?

Re:Woohoo, another kernel compile.

[Taurim]

With the previous kernels the P4 HT is treated like a true SMP which it "nearly" is...

From the OS point of view there is 2 "real" processors available. They share the same ressources (cache, execution units...) but there is not a "primary" and a "secondary" processor. They are equivalent like in a true SMP system.

The OS currently has no way to know if a process to be sent to one of the logicals processors will eat the ressources of the other processor or not.

I use the Con Kolivas SMT patch since 2.6.3 and it really makes my P4 3.0C fly compared to a stock kernel (Yes Seti always runs on it ;-) )

Re:Woohoo, another kernel compile.

[Short Circuit]

Understood...I was talking about a bit-lower level, though, regarding how the physical CPU divides resources between its logical processes.

Re:Woohoo, another kernel compile.

[Anonymous Coward]

Now comes the most feared part for all the contestants: mv System.map /boot && mv arch/i386/boot/bzImage /boot/vmlinuz-2.6.7 && update-grub && reboot

make install?

Re:Woohoo, another kernel compile.

[Unknown Lamer]

make-kpkg kernel_image && sudo dpkg -i ../kernel-image-2.6.7-mm1-smp-_123.Custom_k7.deb

And my two processors just beat all of your machines ;)

Re:Woohoo, another kernel compile.

[jusdisgi]

There isn't typically a 'make install' step in the 2.6.x build process. Usually just 'make && make modules_install' and then move your bzImage and System.map to /boot. Oh, and fix up your bootloader.

Re:Woohoo, another kernel compile.

[Taurim]

With the current P4 en Xeon, there is no way for the OS to find these informations which would be very useful for the scheduler.

Maybe in future HT processors from Intel ?...

Re:Woohoo, another kernel compile.

[Short Circuit]

I suppose you could write a kernel boot-up benchmark module to do it. But I don't know enough about the subject matter to do it effectively.

what about 2.4?

[PatrickThomson]

When are we going to see 2.4.27 with this bugfix? not all of us can afford to, or are able to switch to 2.6

Re:what about 2.4?

[dema]

I applied the patch noted in the bug report, and am now free of the problem.

Re:what about 2.4?

[Anonymous Coward]

Out of curiosity, what would prevent someone from being able to switch to kernal 2.6?

Also, why do people always speak of 2.4.x and 2.6? What happened to 2.5?

Re:what about 2.4?

[andrej73]

patch for i386

Re:what about 2.4?

[Dogers]

x.odd versions are the development versions inbetween. while 2.6 is current, towards the end of its life, Linus or whoeever will create a development 2.7 kernel. Once thats finalised, it'll be released as 2.8.x

Think of the odd versions as release candidates, if youre from a Windows world! :)

Re:what about 2.4?

[reub2000]

If it's working, and it's imporatant they don't want to switch for no reason. 2.5 was a devolopment kernel. All odd numbered versions are devolopment.

Re:what about 2.4?

[PatrickThomson]

er... what? last time I checked, 4 was even.

Re:what about 2.4?

[Hobart]

I'm not sure, but I think this may be the fix--

http://linux.bkbits.net:8080/linux-2.6/diffs/inclu de/asm-x86_64/i387.h@1.12?nav=index.html|ChangeSet @-4w|cset@1.1754.6.3

Re:what about 2.4?

[PatrickThomson]

heh, I applied the patch, recompiled, rebooted into the new kenel, was definitely running a new kernel, and it still halted ... luckily I only give out shells on my ppc machine :D

Re:what about 2.4?

[kfg]

What happened to 2.5?

It became 2.6 and has been supplanted by 2.7.

That's the way things work around here, odd numbered point releases always being development models for the next stable release which is always even numbered.

There are a lot of good reasons for maintaining older stable releases. Maintaining obsolete development models would be a bit silly.

KFG

Re:what about 2.4?

[egreB]

Out of curiosity, what would prevent someone from being able to switch to kernal 2.6?

The driver architecture in Linux kernel 2.6 changed somewhat from 2.4. Drivers will have to be patched or rewritten to work with 2.6. This is being worked on, but lots of unofficial patches to the kernel haven't caught up yet. My laptop, for instance, was unable to get X up at adequate resolutions with 2.6 (albeit this was around christmas - I might give it another shot with this release).

Then there's low-level userspace programs (stuff not running as a part of the kernel itself) that needs some change. Examples are the PCMCIA-suite.

Re:what about 2.4?

[Dogers]

he was asking about 2.5 kernel! :)

Re:what about 2.4?

[ahaning]

In case there's any confusion, when people say "odd point releases", they mean that middle number. You've probably gathered this by now, though. 2.0.0, 2.2.0, 2.4.0, 2.6.0 were "stable" while 1.9.9, 2.1.0, 2.3.0, and 2.5.0 were development.

So, 2.4.25 wasn't a development version, even though it ends in an odd number. The last number just indicates a minor version revision and it's still considered stable.

Re:what about 2.4?

[Anonymous Coward]

The full speed free version of the drivers for connexant winmodems doesn't work with it.
2.5 was a development version

Re:what about 2.4?

[dinivin]

For me, none of the 2.6 kernels (2.6.1, 2.6.3, 2.6.5, and 2.6.6) I've tried will work with my firewire drive... All requests for help on the linux-kernel and ieee1394 mailing lists have either fallen on deaf ears or resulted in suggestions that didn't work. The same drive works fine with the 2.4 kernels.

Dinivin

Re:what about 2.4?

[Short Circuit]

Email the maintainer for the relavant kernel module...if something was broken from the transition from the 2.4 kernels to the 2.6 kernels, he ought to know.

Re:what about 2.4?

[dinivin]

Ben Collins, the maintainer of the sbp2 driver, is the ieee1394 project administrator and, I'd hope, is subscribed to the list. I also posted a bug to the projects bug tracker on sourceforge:

http://sourceforge.net/tracker/index.php?func=deta il&aid=933965&group_id=2252&atid=20225 2
which has gone unanswered and even unassigned.

Not particularly uncouraging, huh?

Dinivin

Re:what about 2.4?

[generic]

Same happened to me, applied patch for 2.4 kernel rebooted and system still halted.

Re:what about 2.4?

[Short Circuit]

No, it's not. But I still suggest you email him directly. It's much more likely to get his attention than a message on lkml...I don't think there are many people who could read all the posts on that list and still manage to get work done.

Re:what about 2.4?

[Nutria]

For me, none of the 2.6 kernels (2.6.1, 2.6.3, 2.6.5, and 2.6.6) I've tried will work with my firewire drive... All requests for help on the linux-kernel and ieee1394 mailing lists have either fallen on deaf ears or resulted in suggestions that didn't work. The same drive works fine with the 2.4 kernels.

My external drive works on 2.4.24 (the last 2.4 kernel I tried) and 2.6.3, but not 2.6.[456]. I also emailed bcollins, but also got no response. :(

Maybe 2.6.7 will work, but I'm not holding my breath.

Re:what about 2.4?

[dinivin]

Well, I'd hope he's subscribed to the linux ieee1394 mailing list, which gets much less traffic than lkml. In any case, I'll follow your suggestion and e-mail Ben Collins directly if 2.6.7 doesn't work (and so far, it doesn't work when I have serialize_io set to 0 in the sbp2 driver.... I'm about to try with serialize_io set to 1). However, I have heard from others who have e-mailed Ben directly with similar problems but who never heard back from him, so I have my doubts.

And, frankly, this is a shame. It's the one item keeping me from upgrading to 2.6 on three separate machines (all suffer from the same problem), and has me thinking about switching to FreeBSD where I know the drive works with -CURRENT.

Dinivin

Re:what about 2.4?

[Seehund]

Out of curiosity, what would prevent someone from being able to switch to kernal 2.6?

Well, if you type "kernal.org" in your browser, you won't get very far. So the inability to spell might be a stumbling block when switching to kernel 2.6. ;)

Re:what about 2.4?

[Anonymous Coward]

Well that's *almost* as easy as Window's Update! Who says Linux isn't user friendly? How do I patch my kernel. Visit an obscure website linked to in an obscure comment in an obscure story on an obscure website, duh!

Re:what about 2.4?

[Alexis de Torquemada]

When are we going to see 2.4.27 with this bugfix? not all of us can afford to, or are able to switch to 2.6

You can't afford it? It's free as in "free beer"!!1!11!1

Re:what about 2.4?

[ross axe]

What happened to 2.5?

It became 2.6 and has been supplanted by 2.7.

2.7? What 2.7?

Try "It became 2.6 and will be supplanted by 2.7"

Argh! Dilemma!

[troon]

Do I demonstrate my machismo via my large and increasing 2.6.6 uptime, or do I impress the chicks by running the latest kernel release?

Help!

Re:Argh! Dilemma!

[Anonymous Coward]

If you are a true Linux user, you will keep the uptime, regardless of any new kernel that comes out. Do not break culture!

Re:Argh! Dilemma!

[sucker_muts]

By chicks, do you mean the children of tux and his (yet to be discovered) mystery girlfriend? I allways thought tux was up to something...

Re:Argh! Dilemma!

[TwistedSquare]

Chicks like an experienced, mature man - switch back to 2.4 or even better 2.2 or 2.0.

Re:Argh! Dilemma!

[jcinnamond]

Real men chroot and do both.

Re:Argh! Dilemma!

[kfg]

Just get a "Type R" sticker. Chicks dig it and it's the same thing as a kernel upgrade really.

KFG

Re:Argh! Dilemma!

[mrjackson2000]

how would you go about doing that? seems possible, but i wouldn't know where to start

Re:Argh! Dilemma!

[Anonymous Coward]

wait a sec, how can you demonstrate your "large" uptime in 2.6.6 when 2.6.6 has not been out that long? - unless by large, you mean a month or so.

When most of us say large, we mean LARGE, like years! My 486 laptop (next to my bed for email checking and irc etc... ) was up for 325 days until we had a powercut :( being a 486, the battery lasts for a "large" time of 1 second!

Re:Argh! Dilemma!

[Short Circuit]

Well, I'd try user-mode-linux. I don't know what chroot has to do with it.

Re:Argh! Dilemma!

[deadmongrel]

chicks? where (looks around) where? where? where?

Re:Argh! Dilemma!

[Anonymous Coward]

run 2.6.7 in UML (user mode linux) under your huge uptime in 2.6.6

Re:Argh! Dilemma!

[jcinnamond]

You can't really chroot the kernel (at least as far as I know). The kernel is the only thing that stays consistent across a chroot. (If you are interested in using chroot to hold a different installation you could start by looking at the debian manual.)

I guess you could install something like bochs and run a kernel inside a kernel, but that is a lot of effort just to improve your uptime.

Re:Argh! Dilemma!

[DrNibbler]

So... following that arguement to its logical conclusion I should get more sex then Ashton Kusher?

Re:Argh! Dilemma!

[Sloppy]

Modify your kernel to save the uptime to a file when it shuts down, and to read from the file on boot.

It's the best of both worlds, like having a cute puppy that you can take to the park (chicks can't resist it), but goes into suspended animation when you get home, so you don't have to take care of it.

Re:Argh! Dilemma!

[Anonymous Coward]

Judging by your username, your technique might be what's driving them away.

Re:Argh! Dilemma!

[Punboy]

But no woman would respect a man doing it with 2 kernels

Re:Argh! Dilemma!

[vijaya_chandra]

doesn't chroot just give you a new root i.e the / directory!?!

or may be i am not yet worthy enough to be a root and am wrong

Re:Argh! Dilemma!

[It'sYerMam]

They're a bit pixellated, but not bad.
Perhaps they need a new kernel.

Re:Argh! Dilemma!

[squeegee_boy]

I have 382.7 days uptime on 2.6.6. Beat that.

Also, my resume says I have 10 years experience with .NET ...

Re:Argh! Dilemma!

[radiophonic]

#!/bin/bash
while read CHICK; do
echo 'Wanna go out on a date?' | /dev/$CHICK
if [ $? := 0 ]; then
rm -f $CHICK
else
echo $CHICK >> ~/my_yes_list
fi
done < ~/my_chick_list

Re:Argh! Dilemma!

[morleron]

Go for the latest version. Most chicks say they aren't impressed by size anyway.

Just my $.02,
Ron

Re:Argh! Dilemma!

[jusdisgi]

Yeah. I was just thinking the same thing. My main desktop at work was up for 344 days when we lost power here last month. The worst part is, that's what happened at the start of that time too.

Of course, before that power outage I'd had only ~50 days uptime...but that was me being a jackass. Keyboard in mouse-hole (I share a keyboard/mouse/monitor, but we're too cheap for KVM's) caused machine to not wake-up...ssh in and everything looks fine....after totally failing to figure it out, I told it to reboot via ssh. When that didn't change anything I eventually pulled my head out.

But anyway, point is, you have to *really* endure an old kernel if you want uptime worth bragging about. I don't think any 2.6 kernel users are there yet.

Re:Argh! Dilemma!

[beeblebrox87]

chroot just selects a new root filesystem/userland. The kernel remainst the same. Real men use user mode linux, which accomplishes what you describe, letting you use one kernel as a process running under another.

Re:Argh! Dilemma!

[Alexis de Torquemada]

Hi, my name is Frank Parts. I served as a Private under General Error, Colonel Panic, Major Mistake and Corporal Chaos.

Re:Argh! Dilemma!

[Punboy]

What about General Protection Fault? --The Dev-Nulls Advocate

Re:Argh! Dilemma!

[KidSock]

And nobody likes a snitch.

Re:Argh! Dilemma!

[Punboy]

Nor does one enjoy being treated like a bitch

Just curious

[Anonymous Coward]

Given that 2.6.x has been out for a while now, is anyone running the 2.6 series in a full blown production environment yet (say, database or web server)? If so, how does it compare to the 2.4 series?

Re:Just curious

[djhankb]

I have implemented a Samba PDC Fileserver for one of my clients. It's very fast, and has been 100% stable. So far i've been quite impressed.

-Henry

Re:Just curious

[Spacejock]

Yes.

Long answer: 1 web/file server which also runs a 20-user LTSP setup, and 2 web/file/print/email servers with 5 users each. None of them are really heavily used so it's hard to say what kind of difference it made going to 2.6, but I've not had any problems.

Re:Just curious

[bwindle2]

I run 2.6.6. on an SMP machine, ext3 on SCSI RAID5 that runs MySQL/Apache/MRTG/BigBrother. It has been completely stable.

bwindle@balrog:~$ uptime
09:06:48 up 36 days, 22:03, 2 users, load average: 1.00, 0.55, 0.43
bwindle@balrog:~$ uname -a
Linux balrog 2.6.6 #3 SMP Mon May 10 10:55:43 EDT 2004 i686 GNU/Linux

Re:Just curious

[Superfly_rh]

I've got it on two servers, had it on three but had to back down to 2.4 for one.

It's running a Qmail/Courier IMAP server w/ webmail interface. And it's running a rather busy nfs/samba server.

I had it running on a second NFS/Samba server that was using LVM2 (only difference that I can tell). With the 2.6 kernel I got kernel panics 2-3 times a week. So I went down to the 2.4 kernel and it hasn't crashed since.

Re:Just curious

[leandrod]

> I had it running on a second NFS/Samba server that was using LVM2 (only difference that I can tell). With the 2.6 kernel I got kernel panics 2-3 times a week.

Similar experience here. Had 2.6.3, if I remember well, with LVM, software RAID5 and ext3. Didn't got kernel panics, but abort logs that forced a reboot 'cause the filesystems were remounted readonly. Eventually I lost the /, so backed down to 2.4.

Tried to follow the issues in the relevant mailing lists, there was little interest by the powers that be.

I guess Tannenbaun was right, monolithic kerni are getting just too complex. If only the Hurd got critical mass...

Re:Just curious

[dangerz]

I dont know if this counts, but I've been using 2.6.6 on my laptop. 3.2 ghz p4 with 512 ram..

I have had no problems at all with the kernel; very stable. I'm just waiting for better laptop suspend support. It's a pain having to completely 'shutdown -h now' each time I wanna go somewhere.

Re:Just curious

[Anonymous Coward]

I switched from 2.4 to 2.6 when test 10 was released. Machine was a Celeron 300 overclocked to 450. We used it in a production environment (non-profit organization, couldn't afford anything better at the time). Speed improvements were very noticeable. Have not had a single issue. We actually just got some decent server class hardware donated to us. We right now run a database server, email server, list server, and webserver all off of 2.6 kernels without any stability problems.

Re:Just curious

[datan]

have you tried the software suspend patches from swsusp.sourceforge.net?

Re:Just curious

[Shaman]

I have approximately 20 machines using 2.6 since last fall and with the exception of one (an AMD-64 box), they have all been exemplary. That machine became stable with 2.6.6 though its BIOS seems flakey (hardware problems.. ugh)

In particular, my HT machines seem to perform very well with 2.6.3 and up.

Re:Just curious

[fire-eyes]

Yes, I have been doing it since 2.6.1 or so. We have databases, mail servers, ftp servers, apache servers, samba, blah blah.

I've had ZERO issues, in fact, it's been fantastic. I strongly recommend it.

Only thing to watch out is that you must have module-init-tools installed, though there may be other gotchyas per distro.

Re:Just curious

[dangerz]

ya i've tried those to no avail.. it's for a dell 5150, so it's a rather newer notebook

Re:Just curious

[jusdisgi]

Yeah, I've been running a varietty of production boxen (web, email, mysql, etc.) on 2.6 since about .3 and I have no complaints. I don't see any reason to install a server with 2.4 anymore...the stability is the same as it always was--flawless.

Re:Just curious

[hackus]

Been running Oracle 9iRelease 2 on Kernel 2.6.6 quite nicely.

No problems yet!

The real problem I have is if I want to use the Java utils with 9i, which given the compiler and lib requirements with Oracle's jre, can get quite dicey.

I rarely use the Java admin utils but they are nice to play with.

-Hack

Re:Just curious

[KidSock]

... is anyone running the 2.6 series in a full blown production environment yet ...

I know a UML VPS provider that is upgrading some of their host kernels to 2.6 tonight (although that was schedualed for some time now and has nothing to do with this release).

Re:Just curious

[radiophonic]

I have it running on about 8 machines here on the LAN.

All of the machines' performance increased slightly, especially in the networking department.

Re:Just curious

[Anonymous Coward]

Reading the early pro-Linux posts, even Linus made it clear that his kernel was only meant to fill in the gap while we waited for Hurd. Looks like that will never happen now, so we just have to make do with what we have. However, all of the major competitors are monolithic as well, so at least we look good in comparison, eh?

Re:Just curious

[Lennie]

MacOSX/Darwin and supposedly NT/200x/XP are not.

Re:Just curious

[lintux]

Darwin is just a monolithic kernel on top of a microkernel. Windows NT used to be a microkernel OS, but that's a long time ago already. AFAIK NT4 went "back" to monolithic already.

Re:Just curious

[kaltoft]

kernel.org runs on 2.6, right now they are serving kernels at ~250MBits/s

Re:Just curious

[sydb]

Was 2.6 not released in December?
Yes, it was:

ChangeLog-2.6.0 17-Dec-2003 19:04 12K

Re:Just curious

[sydb]

I'd guess it's pretty good for a server.

As a desktop, I've had no problems either.

But the Linux Audio people still recommend 2.4.xx with low-latency and preempt over 2.6. However I've had weird USB problems with that particular combination.

Now everyone...

[ThisNukes4u]

Is going to rush to download this because it fixes probably the most destructive kernel hole in a few years. And its on slashdot.

Re:Now everyone...

[alex_tibbles]

it's not really destructive, it's only a DOS (local too).

Re:Now everyone...

[Short Circuit]

Sure..."local" as in, "anyone who can install cgi scripts can do it."

Or any way a remote user can run arbitrary code (not even as root.).

Re:Now everyone...

[Anonymous Coward]

Hmm, I still think that the following bug is still in there:

echo -e "t&t&" > t
chmod 777 t
./t

and there you go... Well, doesn't crash the machine directly, but nearly ;)

ineiti

Re:Now everyone...

[ThisNukes4u]

System downtime costs business money, especially when its from some stupid prick who has a web access shell that is running the code.

Re:Now everyone...

[tunah]

probably the most destructive kernel hole in a few years.

Just recently there were a bunch of mremap exploits allowing local root access. This is less severe (but still very bad).

Re:Now everyone...

[mko]

[fork-bomb]

That's what ulimits are for (max user processes). Those have been in UNIX for at least 20 years, Linux must have gotten them in '92 or so.

Re:Now everyone...

[Short Circuit]

They must not be enabled by default...I murdered my machine while trying to test a safety trigger mechanism.

It was so much fun, I changed the script to write the system load and timestamp to a text file. When I got the chance to examine it, I saw it had pretty much slowed to a halt at about 350.

It was a fun experiment. At least now I know to make the trigger automatic instead of keyboard-controlled. :)

Re:Now everyone...

[ThisNukes4u]

OpenBSD is all about securing the system from REMOTE holes. The openbsd home page boasts: Only one remote hole in the default install, in more than 8 years! REMOTE is the key word. They could care less about anything else(well maybe not care less, but thats not their focus)

Got it

[pcmanjon]

I just compiled and installed it. It's not that bad.. or good... orr... how the hell should I know?

System doesn't seem to run much different, I haven't read the changelog

but for those of you who want to read the changelog it can be found HERE:

http://www.kernel.org/pub/linux/kernel/v2.6/Chan ge Log-2.6.7

Re:Got it

[pcmanjon]

Sorry, bad link above

Here's a good link

http://www.kernel.org/pub/linux/kernel/v2.6/Change Log-2.6.7

AGAIN, I APOLOGIZE!

Re:Got it

[MalikChen]

Run the exploit script. See anything different?

Time for download then

[Impie]

And see if my Radeon 9800 Pro will work with my Nvidia Nforce2 chipset.

Anyone gotten Nvidia Nforce2 and Radeon 9800 Pro working with 3d Accel?

Re:Time for download then

[Dagny Taggert]

that's my big complaint about 2.6: it still didn't address 3D issues with ATI Radeon cards. Is this an issue with ATI not releasing binaries to people like Suse? I'm just a casual GNU/Linux type, so I'm not sure.

Re:Time for download then

[Short Circuit]

What issues are there? Do you have a reference?

My ATI Radeon 9000 works fine under the DRI that comes with the kernel.

(And before someone says, "That's an old card; you should upgrade," I'll point out that it was a significant improvement over the Riva TNT2 I had until this past January.)

Re:Time for download then

[Anonymous Coward]

And see if my Radeon 9800 Pro will work with my Nvidia Nforce2 chipset.

Get a GeforceFX card. They work great on the Nforce2 chipset.

Re:Time for download then

[Pelops]

Well, I am not exactly answering your question but, i managed to get my Radeon 9500 Pro to work with my Nvidia NForce 2 chipset with 3D Accel.
The 9500 Pro is very similar to the 9800 Pro.
I use Gentoo, a 2.6.6 kernel, Xorg-X11, and ATI-Drivers 3.9. I know a few people will disagree on using proprietary drivers but again, i have 3D accel working.
I am not using any of the Nvidia proprietary drivers to run my system and it is working perfectly well.

Re:Time for download then

[Impie]

Well .. it seems I have problems with the DRI thingys when I try out ATI:s driver.

I am running Mandrake 10.0 official and Kernel 2.6.3-7 on this box.

I bought the card because it offered best price/perf at that time, now I begin to wonder if I should toss the card out and buy an Nvidia card instead. As it is now the only OS I can run with 3D accel is WinXP and I do not want that (at least not all of the time)

Re:Time for download then

[Anonymous Coward]

The 9000 & 9800 are different chip revisions. There is a problem with AGPGART on nForce 2 boards with R350 Radeons.

Re:Time for download then

[Short Circuit]

Which card? And why is it necessary to use ATI's driver?

Like I said, the driver that comes with the kernel works fine...3D accel and all.

Re:Time for download then

[Impie]

Well .. I tried the kernel driver and it did not work either.
I'll do a fresh install when I get home today and give it another try.

Does anyone have a good pointer to any howto on the net? And by good I mean step by step exactly how they managed to get it working. I have run Linux since 94 so I have the experience, but I haven't managed to solve this specific problem. This is very annoying and a big blow to my ego :-).

It is a Sapphire Radeon 9800 Pro with 128 Megs of memory. I use an ASUS A7N8X-E Deluxe motherboard with 1 gig of memory.

Re:Time for download then

[pmjordan]

Well, I've got a Radeon 9800 XT on an nForce2, which works fine. The trick is to compile agpgart and nvidia_agp as modules, and disable using fglrx's internal AGP drivers in XFree86Config-4 or xorg.conf.

It's been working fine for me that way since 2.6.3 or something.

Re:Time for download then

[Impie]

Thanks for the tip. I'll try this one out as well.

Re:Time for download then

[Short Circuit]

Well, a quick googling didn't turn up a howto, but here are a couple of related links for your perusal:

ATI's Radeon 9800 Pro technical issues page

A LinuxQuestions.org thread on ATI with the NForce2 on Mandrake 10.0.

Re:Time for download then

[Impie]

Are you using the ATI prop. drivers?

Thanx for all of the tips folks.

Re:Time for download then

[pmjordan]

Yes, I'm currently running fglrx 3.9.0 on my machines with Radeon 9800XT, Radeon 8500 and Radeon 9000 Mobility. Previously I've had various other versions (I think back to 3.2.8 or so) working as well.

In gentoo you can do an "emerge ati-drivers".
If you're running a recent SuSE kernel you can find special rpms at
ftp://ftp.mirror.ac.uk/sites/ftp.suse.com/pub/ suse /i386/supplementary/X/XFree86/ATI/
but if you're using a vanilla kernel (such as I am) you'll have to use the ATI drivers.
I think the ATI drivers come pre-compiled for various Mandrake and Red Hat kernels as well.

I've always compiled my own though.

There's a lot of info on the web about the fglrx drivers, google is in fact your friend. (at least on this matter)

Re:Time for download then

[Jason Hood]

nforce3 + 9800Pro works here with no problems. unreal flys. (32bit native mode)

Re:Time for download then

[Anonymous Coward]

Well, so the fglrx driver works with 2.6... My question, then, is does it work for AMD64, or only for IA32?

Re:Time for download then

[Anonymous Coward]

I can't even get the NForce2 working with 3D acceleration.

Well, not until NVidia work out how not to call deprecated kernel functions from their binary driver, anyway.

Re:Time for download then

[Pelops]

You could go check on the gentoo database application to see if there is any AMD64 flag

Re:Time for download then

[Impie]

The tips at linuxquestions.org worked like a charm. 4700 fps in glxgears and about 800 in fgl_glxgears.
On top of that DualScreen works as well.

Now I just need to install my Neverwinter Nights linux client and off we go.

Thanks for all your help.

Re:Time for download then

[Short Circuit]

Thanks for all your help.

Not at all. :)

Err

[Anonymous Coward]

It's not on ftp.uk.kernel.org or ftp.fr.kernel.org?

Re:Err

[cs02rm0]

Just use kernel.org (it's definately there), it'll take a bit to work down through to the other mirrors I guess.

Re:Err

[nathanhart]

It was just released so a few mirrors probally haven't ran resync yet

Starting to take a bit longer for a release...

[tejohnson]

Wow, this one actually took over a month! There for a while there was a kernel being spit out about once a month.

Re:Starting to take a bit longer for a release...

[goldspider]

Indeed, and where are all of the zealots who claim Linux superiority over Micro$oft in the bug-fix turnaround department?

Re:Starting to take a bit longer for a release...

[Punboy]

We still kick ass. Microsoft has yet to fix issues in IE that have been there since IE 3. And they're at 6

Thank you, Torv^WSCO!!!

[Cronopios]

I bet that pesty security hole was written by Darl himself ;D

Re:Well I'll be...

[sucker_muts]

Doesn't matter, all kernels are affected. (2.4+ anyway)

Obligatory luser question

[ObsessiveMathsFreak]

I've got Fedora Core 2.

yum update kernel*

should install 2.6.7 right?

I'd much prefer to use an auto installer, I'm not ready for the full shebang yet.

Enlighten me here.
How is a full compile of the kernel done and how long would it take on a 3GHz,756RAM computer?

Re:Obligatory luser question

[Sunspire]

Sort of, it'll definitely fix the local DoS bug, but it won't give you 2.6.7 (yet). Red Hat patched the bug some days ago already.

You can keep up to date regarding regular updates and security errata at Fedoranews.org FC2 updates page.

Re:Obligatory luser question

[Phekko]

Ok, I'll bite

It won't take a lot of time, not with your system specs. We're talking about 30min tops. Depends a lot on what kind of modules etc you need to make.

If you've never done it, though, you may want to a) ask someone who knows this stuff for help or/and b) read a few docs on the subject. If you don't know what you're doing, you most probably will end up with something missing, stuff like soundcard support or something. A good place to start (in my biased opinion ofcourse) would be here for instance Some googling never hurts, either.

Re:Obligatory luser question

[ObsessiveMathsFreak]

I've just remembered to ask!!

Is this going to fix the Firewire issue?

Re:Obligatory luser question

[alexbartok]

Making a Kernel isn't too hard. You find the correct image, (yep you probably want 2.6.7 ;), download the .tar.bz2 version. To keep this short, we'll stick it in /usr/src: 1) su - 2) cd /usr/src 3) wget http://full.url.to.file.including.filename 4) tar xjvf linux-...tar.bz2 5) ln -s ./linux-2.6.7 ./linux 6) cd linux-2.6.7 7) make menuconfig 8) configure options - read the help for information about options, if necessary google around, usually it's pretty easy to find information about specific options 9) make bzImage 10) make modules 11) cd arch/i386/boot/ 12) cp bzImage /boot/ 13) edit /etc/grub.conf with your favorite editor, basically adding the new image, using the existing one as template. reboot and enjoy :)

Re:Obligatory luser question

[Anonymous Coward]

Extract the tarball, read the README. You'll probably want to do a make oldconfig and you'll probably want to look at Documentation/modules.txt. That's really all there is to it. Time taken on your machine will be of the order of minutes.

Re:Obligatory luser question

[Advocadus Diaboli]

Fedora Core 2 has already a kernel update. See the announcement of the 2.6.6-1.435 kernel. So all you need is getting the RPM and install it.

The question is, when will this patch show up on other distributions. People are sometimes not able to compile a vanilla kernel or a vanilla kernel can cause headache, e.g. SuSE 9.1 formats your filesystem with reiserfs and ACLs, but a vanilla kernel might not support this backported ACL feature.

Seen the kernel release from this point of view means, that the sistributions should hurry up to provide fixed kernel packages for their users.

Re:Obligatory luser question

[canadiangoose]

I've got Fedora Core 2.

yum update kernel*

should install 2.6.7 right?

Why not try it and find out?

I'd much prefer to use an auto installer, I'm not ready for the full shebang yet.

Enlighten me here. How is a full compile of the kernel done and how long would it take on a 3GHz,756RAM computer?

Also, why not just try it? You have the machine in front of you, give it a whirl! That's what makes computers so much fun; you get to try out new stuff all the time.

Re:Obligatory luser question

[flex941]

Vanilla kernel supports ACLs for reiserfs and data=ordered and other nifty things starting from 2.6.6 (if I'm not mistaken).

Re:Obligatory luser question

[goodvilhunting]

'yum update kernel' will install kernel-2.6.6-1.435 that fixes this problem

Re:Obligatory luser question

[linuxcoder]

Actually with the new KBuild system it's not necessary to make the individual components. Just do a: make menuconfig (or make oldconfig if you already have a working configuration) make make modules_install cp arch/i386/boot/bzImage /boot Now configure lilo/grub if necessary. If using lilo, don't for get to run it before rebooting.

Re:Obligatory luser question

[gunpowder]

Just a few comments to your build guide:

1) su -

Not neccessary at this early step, ie. not needed for compiling the kernel. Do a 'su' when you are about to install the compiled kernel&modules.

2) cd /usr/src
5) ln -s ./linux-2.6.7 ./linux

Not needed. 5) is actually discouraged by Linus. Just unpack the linux kernel sources somewhere and cd into that directory.

9) make bzImage
10) make modules

both commands are only needed for 2.4.x kernels and if you compile a kernel for the ix86 platform. If you are using 2.6.x, you should simply do a 'make'.

11) cd arch/i386/boot
12) cp bzImage /boot/
13) edit /etc/grub.conf

are also only needed if you compile for the Linux/ix86 platform. 13) only applies if you are using GRUB as a bootloader on a RedHat system. BTW, GRUB's config file is usually in /boot/grub/menu.lst.
You are also missing a 'make modules_install' and it might be a good idea to save your System.map and .config aswell.
Instead of 11-12) you might want to do a simple 'make install' (for 2.6.x kernels). Also try out 'make help'.

Re:Obligatory luser question

[Anonymous Coward]

30 minutes, on a 3GHz box? Hell no. By my experience, a build of an unusually featureful 2.6 kernel, on my Athlon64 3400+, takes about 2.5 minutes. So, on a 3GHz (doesn't specify, so I'll assume P4), with 768MB, ought to take about 2.5-4 minutes. (Mine is unusually fast because I have RAID, which speeds up the disk-bound parts by 65% or so)

Re:Obligatory luser question

[parksie]

What should be under /usr/src/linux is a version of the kernel headers that your glibc was compiled against. Thus, you may have /usr/src/linux as the 2.6.6 headers that you built glibc with, when you upgrade the kernel to 2.6.7, you just use it under /usr/src/linux-2.6.7 and leave /usr/src/linux alone.

Re:Obligatory luser question

[Dravik]

When does this start to get simple?

Re:Obligatory luser question

[imroy]

e.g. SuSE 9.1 formats your filesystem with reiserfs and ACLs, but a vanilla kernel might not support this backported ACL feature.

Not anymore. It looks like EA/ACL's for ReiserFS are in 2.6.7.

Problems with JFS?

[henley]

Anyone else unable to compile with JFS enabled as module?

fs/jfs/jfs_dtree.c: In function `add_index':
fs/jfs/jfs_dtree.c:388: parse error before `struct'
fs/jfs/jfs_dtree.c:389: `temp_table' undeclared (first use in this function)
fs/jfs/jfs_dtree.c:389: (Each undeclared identifier is reported only once
fs/jfs/jfs_dtree.c:389: for each function it appears in.)
make[3]: *** [fs/jfs/jfs_dtree.o] Error 1
make[2]: *** [fs/jfs] Error

Google shows no hits, and it's not important enough for me to track any further at the minute (since disabling JFS is an adequate work-around for me).....

Re:Problems with JFS?

[flex941]

This e-mail might help you out.

Re:Problems with JFS?

[red floyd]

See? It's because you're an evil pirate, trying to sue SCO's code without a Linux IP License! If you pay Darl his $699, then it will magically compile properly!

Re:Problems with JFS?

[Anonymous Coward]

full bottle, jesus, get it right
I'd rather have a full bottle infront of me than a full frontal ....

English: Linux Today has human redable changelog

[VC]

Linux 2.6.7

In German, but no need for translation.

[Alsee]

I was going to post a link to a bablefish translation, but it really doesn't add anything new. There's a 2.6.7 release. It has minor fixes. And Linus Torvalds corrects the expenditure the nose with the treatment of Floating POINT exception.

-

2 questions:

[Anonymous Coward]

Two Questions:
Does grsecurity work with it? I heard they won't do any new releases now.... maybe an old patch will work?

Is the fix they applied just to prevent the exploit that was released or does it actually prevent all FPU errors in signal handlers from causing endless loops?

~Jacob

Latest Linux Release

[pipingguy]

Yeah but does it run Windows?

Re:Latest Linux Release

[Anonymous Coward]

Yes

Re:Latest Linux Release

[vijaya_chandra]

yes it does

WINE Is Not an Emulator

or you can try out vmware

You can thank me...

[bender647]

I knew if I patched and rebuilt 2.6.6 yesterday they would release 2.6.7 today :(

Re:You can thank me...

[Neo-Rio-101]

I'm in the same boat. I just made a 2.6.6 kernel today too and now I feel ripped off.

Re:You can thank me...

[Icyfire0573]

I did you one better, last night at 2AM EST I patched my kernel with 2.6.7-rc3 to get LM-Sensors workin... not even 2 hours later ...

Not upgrading...

[halivar]

Those of use still running P3-500's on old mobo's don't have very many compelling reasons to upgrade from 2.6.3.

If in doubt, don't upgrade unless you need new support for essential hardware or need to cover a security vulnerability. I leanred that after b0rking several systems trying to keep my kernel perpetually updated.

Re:Not upgrading...

[alexbartok]

Well security is not a reason to update your system ?

Re:Not upgrading...

[MartinG]

Are you sure?

I upgraded my p133 and it feels much faster for interactive use.

Re:Not upgrading...

[mrjackson2000]

i've got a p75 running 2.6.6, will be running 2.6.7 in a few days, i dont see what the big deal is upgrading w/i the 2.6 kernel. 2.4 to 2.6 i can see not doing on an old comp but thats it

Re: Not upgrading...

[Anonymous Coward]

Those of use still running P3-500's on old mobo's don't have very many compelling reasons to upgrade from 2.6.3.

Kernel 2.6.3 had a very broken OSS ALSA emulation layer. This is why I switched down to 2.6.2. Version 2.6.5 and above have a major ALSA fix. So if you use your soundcard at all, then it is definitely worth it to upgrade.

Re: Not upgrading...

[Anonymous Coward]

Really? That's good to know. Maybe I'll be able to get my darn soundcard working now.

Is it just me,

[aixou]

or are the kernel version numbers escalating rather quickly. Already at 2.6.7? Isn't the 2.4 kernel still at 2.4.2x? Can someone explain to me the reason behind the quick rise? Are they just anxious to get to v.3?

Re:Is it just me,

[ThisNukes4u]

thats 2.4.twenty-x, not 2.4.2.x, so 2.4 is still way more versions ahead.

Re:Is it just me,

[werve]

The version numbers are not real decimals, i.e. the sequence will be:

2.6.8
2.6.9
2.6.10
2.6.11 ...

So we're along way from .27 in the 2.4 series.

Re:Is it just me,

[aixou]

Ahhh, gotcha. I misread the numbers. Thanks. :)

Re:Is it just me,

[btsdev]

Not sure what you mean by "v.3" but as it's been said in the last couple of kernel updates, things aren't actually "escalating rather quickly." Yeah, 2.4 has a version number in the 20s, and there's no problem with that. 2.6 can and will go into the 20s as well; the last version number in x.x.x isn't restricted to a single digit. When Linus feels like he should move onto 2.8, he will.

Re:Is it just me,

[ville]

The next version after 2.6.9 isn't 2.7.0 but 2.6.10.

// ville

Re:Is it just me,

[His name cannot be s]

The first 8-10 revisions in a 2.X release tend to go quite quickly.

Some even say, that the kernel isn't stable till at least .10 :)

It sure seemed that way when it went from 2.2 to 2.4

The 2.4.0 to 2.4.10 seemed like overnight, and then it slowed down to a small humm :p

Re:Is it just me,

[tunah]

2.4 is no longer being "developed" as such, it's being "maintained". So, in theory, you only get new releases for bugfixes of one kind or another.

2.6, while "stable", is still under development. It seems a little inconsistent, but it seems to work - the kernel guys get it reasonably stable for 2.6.0, a horde of regular users gets it and so there's more feedback/bug reports, and it all develops quite fast for a while, eventually everything calms down and the Downtime Costs Me $1000 A Minute people pick it up, and the kernel guys get to work on a (much more fun, I'm sure) unstable (odd-numbered) branch. At least that's how it looks to me...

Re:Is it just me,

[Jugalator]

v 3?

Hehe, they can speed to 2.6.40 if they wish, but will for that reason not be any closer to 3.0.0. :-)

Re:Is it just me,

[Sunspire]

The 2.6.xx revisions have no bearing at all on when the 3.0.0 or 2.7.0 trees will get created. The quick turn around times are due to many factors; the new versioning and source control procedures put in place for 2.6 naturally encourage a more rapid pace while elimating the "did my patch make it into Linus's tree?" problems of yesteryear, which in turn has people submitting more, perhaps smaller, patches in a very rapid fashion. The 2.6 kernel is also right now being developed by more developers than ever, until the 2.7 branch gets spun all the efforts are basically focused on this single tree, timely releases keep code divergence down and hopefully prevents 20kloc ALSA merges from happening.

What, are you afraid they're suddenly going to run out of numbers for the 2.6.xx branch? ;) Hint: after 2.6.99 comes 2.6.100. With vendor kernels you can't say where in the 2.6 branch you are anyway, when you're running 2.6.6-1.423 it's can be anywhere between 2.6.6 and 2.6.10 feature and security wise.

Re:Is it just me,

[iabervon]

They go through the small numbers quickly for a number of reasons and then slow down. Essentially, there are a lot of small changes that they wanted to make from 2.6.0, and they want to get all of the small changes in before they start on big disruptive stuff (which will start 2.7); once localized changes are less important than rearranging awkward interfaces, 2.7 will start, 2.6 will be turned over to Andrew Morton, and 2.6 will change more slowly, since people will tend to work with the nice new interfaces and have to backport to get changes into 2.6

Re:Is it just me,

[Cro Magnon]

.10 my foot! IIRC .11 & .15 were "brown bag" releases. 2.4 wasn't really safe until .18.

Re:Is it just me,

[Anonymous Coward]

2.4 is no longer being "developed" as such, it's being "maintained". So, in theory, you only get new releases for bugfixes of one kind or another.

Even in theory, I think they're still willing to add drivers to old kernel versions, especially since new drivers don't have to be compiled into the kernel so long as they work with existing frameworks.

Re:Is it just me,

[SirNAOF]

Well, that's the way it happens when you replace the VM system in the middle of a "stable" release. I know people who still run 2.2 because of that.

Re:Is it just me,

[Deusy]

Is it just me, or are the kernel version numbers escalating rather quickly. Already at 2.6.7? Isn't the 2.4 kernel still at 2.4.2x? Can someone explain to me the reason behind the quick rise? Are they just anxious to get to v.3?

Oh poor dear. It's just you.

26 > 7, significantly. I know it would make it easier for simpletons to digest if the number was written 2.6.07 but then again kernel code isn't for simpletons.

Meow.

Re:Is it just me,

[Anonymous Coward]

2.6.07

Then there are the folks that would look as "07" and think it's in octal format. (grin)

APIC Fix?

[Apreche]

Does anyone know if the lockup bug with the combination of APIC/APIC-IO and nforce 2 boards is fixed? Currently I've got 2.6.5 and I have to disable APIC to prevent the system from hanging.

Re:APIC Fix?

[flex941]

It should be fixed. Give it a try!

Re:APIC Fix?

[Halthar]

Based on my reading, someone from Nvidia gave em some hints. Nvidia had fixed the bug in the BIOS and gave the information to the manufacturers, but only one manufacturer actually fixed their BIOS (Shuttle I think?). Since no one else was fixing it, someone from Nvidia explained what the problem was and how to figure out if a BIOS had been fixed or not, as well as information on a work around to stop things from crashing. I could be wrong on this, since I still haven't had my coffee yet and this is from memory.

So far as I know those fixes went in to 2.6.6. Unfortunately, as a result of not being able to move to 2.6.6 yet, I haven't been able to test this fix (My two primary machines have NForce2 chipsets), but supposedly it's fixed.

Re:APIC Fix?

[jadel]

I'm running 2.6.6 on an Nforce 2 motherboard with APIC compiled in. It seems quite stable with a current uptime of over a week (the last outage was caused by a thunderstorm.) Previous versions tended to crash every few hours even with APIC support left out.

Re:Beware the Toll

[PoshSpod]

"LOL OMG M$ sucks" - more probably.

Re:Beware the Toll

[presarioD]

+1 for Informative.
+1 for Interesting.

Somebody gave mod-power to teenagers?
Or is this all part of the collective psychology, the oh-so-powerful:

"we all belong to the same guilt/cult/group/$PREFERENCE_HERE so we should all think/act the same?"

Hmmmm!

Isn't that psychosis responsible for the greatest disasters in human history?

I need more coffee please...

Oh, spare me...

[halivar]

I love how everybody who asks "Why is this important?" gets modded down as a troll. This just fuels the argument that Slashdot is full of elitist pricks.

1. Post inflammatory comment calling everyone "pricks".
2. Claim martyrdom because your valid point was modded down.
3. ???
4. KARMA!

First distro to use it? Ninnle, of course!

[Anonymous Coward]

That's because Ninnle Linux is and always has been, bleeding edge Linux!

Re:Beware the Toll

[REBloomfield]

A new Release Candidate is out for XP2, and I bet if that gets posted here, it'll get bashed all afternoon.... I hear ya buddy...

Re:Beware the Toll

[Anonymous Coward]

Because it's a stupid thing to say. Slashdot has been announcing Linux kernel releases sometimes since the beginning of Slashdot. By now the only reason to complain about it is to be intentionally annoying. It's not exactly high-brow debate.

This is needed if they have ftp/cgi access

[ehack]

If the lamers have ftp upload ability and can execute cgi's via apache you'better have that fix in there too. I guess every single free webhost in the world with cgi's will go down in the next few days.

Re:This is needed if they have ftp/cgi access

[bonch]

In light of that, think how different the discussion threads would be if this was a Windows hole and not a Linux one. Your post is the first one I've seen that even mentioned the real-world consequences of this vulnerability.

Re:This is needed if they have ftp/cgi access

[ehack]

The moderation problem here encourages groupthink: Nothing one can do about that. Besides, exactly what is the real workd ?

Re:This is needed if they have ftp/cgi access

[Anonymous Coward]

... if this was a Windows hole and not a Linux one.

If it were a Windows hole you wouldn't even *know* it existed for a year and you'd have to wait around for MS to acknowledge that it was a problem, much less fix it.

So yes, the discussion would be different, and for good reason. Get a grip MS fanboy.

what really chaps my ass is

[Anonymous Coward]

the fact that a kernel revision needs to get a front page placement. Why dont you guys create a section called 'Linux Kernel' as you do for BSDs and dump all this irrelevant junk there?

Linux Linux Linux......Big fucking deal.

Re:English: Linux Today has human redable changelo

[TreeHead]

;if you could spell pseudo, i might think you knew what you were talking about.

;mod -42 off topic.

;TreeHead

Alan Cox?

[mqRakkis]

Alan needs to get to a linux hacking rehab. He still has couple of months left of his year off and I already see him in the ChangeLog!

2.6 kernels

[Neo-Rio-101]

The 2.6 kernels have a plethora of goodies as well as being faster than the 2.4 branch. Better hardware support, Crypto APIs (for IPsec), and in 2.6.4+ we have the beginnings of dm-crypt which is a better method of encrypting entire filesystems. I'm still yet to find a decent 2.6 distro that is good enough for both production and desktop environments though. Still too many bugs in most of the 2.6 based distros prevents me doing much. My newer hardware goes undetected, soundcard not working, SATA RAID being screwed up, webcam acting weird..... *sigh*

Re:2.6 kernels

[locutus2k]

I agree with what you are saying. I have been using 2.6 since it came out in the gentoo portage, and have been very impressed so far. Especially with the UI optimizations that have been made. Hopefully, 2.7.x will be in portage soon, and will continue this great trend. Thanks to all of the kernel developers who have worked so hard.

Re:Beware the Toll

[egreB]

Your sig:
What on earth is a boxen? Are we serious?
We probably aren't serious, but in fact "en" is and ending in norwegian grammar. I'll try to explain it; In english, you'd say "a box". In norwegian, that would be "en boks". ("Boks" is norwegian for "box", obviously.) Furthermore, in english, a specific box is "the box". In norwegian, the article "the" is "en", but placed at the end of the word. So "the box" in norwegian would be "boksen".

Aha! Thus, "a box" rendered in norwegian grammar would be "boxen". Simple as that! (-:

Yes.

[HaloZero]

From the changelog:

Ben Collins:
* ieee1394: CSR1212 Extended ROM bug fixes
* ieee1394: Fix possible NULL ptr dereference with calls to find_ctx()
* ieee1394: Handle swsusp better in kernel threads
* ohci1394: Handle invalid max-packet-size
* ieee1394: Revision sync
* ohci1394: Fix incorrect HPSB_WARNING to HPSB_ERR

Re:Beware the Toll

[myster0n]

Well, I bet that if this update to XP gets released, there will also be trolls asking "Why is this important", and they WILL also be modded down.
And if you read at -1 you'll find enough "linux is teh suck" comments with this article.

So, that seems pretty balanced to me.

Re:Beware the Toll

[presarioD]

Aha! Thus, "a box" rendered in norwegian grammar would be "boxen". Simple as that!

Gotcha! It makes sense now!

PLEASE MOD ME DOWN!

[gotan]

It worked for him, why shouldn't it work for me?

Re:Beware the Toll

[REBloomfield]

yeah, but that's the problem. The linux sucks comments get modded down. The Microsoft sucks ones don't.

gentoo guide

[big daddy kane]

the gentoo guide is great, it helped me as a ubern00b install gentoo from scratch. but if you're going to compile your own kernel, make sure you know your system specs and read every help item for every option. that way you can know what you need and want and what you dont.

Re:Beware the Toll

[pebs]

I love how everybody who asks "Why is this important?" gets modded down as a troll.

We get jackasses saying that in EVERY Slashdot thread. Fuck them all and mod them down. We also get fuckers like you complaining about how things are modded in EVERY Slashdot thread. Fuck you, I hope you get modded down because you ARE a troll. Then we get assholes like me, who comment on how much of an asshole you are, adding to the noise. Yep, yep, mod me down, too, but not until you've modded all the other jackasses down.

quick fix !

[phreakv6]

Someone told me today morning that linux has a security hole

http://www.eweek.com/article2/0,1759,1612368,00. as p

and just see how fast things get fixed on this side of the planet !!

mindboggling

From the changelog:

[Anonymous Coward]

it look like Linus did all the work:

Linus Torvalds:

* Mark compaq Fibre Channel driver broken
* Fix x86 "clear_cpu()" macro
* sparse cleanup of #include file
* Revert wakeup-affinity fixes
* Linux 2.6.7

This is the i386 patch for 2.4

[Lface]

Well, yes that is the patch if you are running x86_64.

The 2.4 patch for i386 is here:
http://linux.bkbits.net:8080/linux-2.4/diffs/inclu de/asm-i386/i387.h@1.5??nav=index.html|ChangeSet@- 7d|cset@1.1447

Dude, You have a memory leak!

[Anonymous Coward]

...how long would it take on a 3GHz,756RAM computer?

You seem to be missing 12MB of ram! you should have 768MB. You need to check for memory leaks now or you won't have any left. Look around the floor of your computer, ram tends to puddle there....

Re:Dude, You have a memory leak!

[Anonymous Coward]

must have leaked into his shared video card.

Re:Beware the Toll

[Anonymous Coward]

There you go, modded you down as Troll. And guess what? Before the others! Sore loser...

-mm patches?

[Ice_Balrog]

I've been curious about what the -mm patches actually do. I know that they introduce some experimental stuff into the kernel, but that's about all I could get from Google about them. Do they improve preformance? Implement new features?

Re:-mm patches?

[numark]

-mm patches are patches to the kernel source by a guy named Andrew Morton. Basically, his patches are more of a "testing ground" for new features that, while useful to some, may be not up to the point of risking including them into a production kernel that is used by businesses who need stable kernels. The features are therefore put into Morton's patches so they can be tested by people who want to take the risk, and some of these patches may eventually migrate to the standard kernel after testing.

Re:-mm patches?

[Anonymous Coward]

According to Gentoo's Kernel Guide:

The mm-sources are based on the development-sources and contain Andrew Morton's patch set. It assembles several other patches, like ext2/3 Extended Attributes and Access Control Lists, Page Table Sharing, the Orlov Allocator, non-linear mapping behaviour, etc into one patch set.

If you really want to live on the edge and you think development-sources are for wussies, then try out mm-sources.

But that might be out-of-date, because the announcement for the patches to 2.6.6 looks like it says competely different things.

As for interpretation of all this (e.g. do I want these patches?), beats the fuck out of me.

Re:-mm patches?

[Anonymous Coward]

mmm, patches... (someone had to)

great!

[Chuck Bucket]

now we'll see all sorts of posts like, "Damn, I jst got and install 2.6.6 now!"

blech. oh well, at least it's giving ppl something to do.

CVSgb

Re:Beware the Toll

[Anonymous Coward]

No, they are modded down as troll because news of the latest kernel releases is WHAT SLASHDOT IS SUPPOSED TO BE ABOUT. If you don't like it, don't read it.

Re:Beware the Toll

[4of12]

If MS announced that they were releasing an update to XP this whole site would be full of posts like

...just like the ones you're getting here about Linux 2.6.7.

Namely, "are people here running their production servers with good uptime, various weird applications, exotic hardware, etc.?"

When XP SP 2 comes out, you can be damn sure people are going to ask the exact same questions about reliability, incompatibility, etc.. At MyCorp we've got staff that do nothing but stress test the latest Windows patches and releases for compatibility with our current environment. We do that because:

• it's very important to our business that people aren't down.
• past history suggests that blind upgrades are for fools rushing in...shit happens.

Slackware 9.1 patched

[mosschops]

The Slackware 9.1 kernel has already been patched and released, still using the 2.4.26 version number.

uname -a now shows:

Linux mosschops 2.4.26 #6 Mon Jun 14 21:07:47 PDT 2004 i686 unknown unknown GNU/Linux

(just a change of date)

You can get it here. Note: there's a 100 user limit on the server, and not all the mirrors have been updated yet.

most destructive?

[dpilot]

Naaah, IMHO the memremap exploits were worse.

This one is 'only' a local DOS. Even if, as others say, crashed time is money, it could be much worse. At least you don't get 0wn3d, and you have a way to get back up by kicking users off, temporarily.

Drifting the topic, slightly...

This exploit, as well as the mremap ones, were derived from intimate examination of the source. So far, most of the Windows exploits have really been using 'features' for nefarious ends, not exploits of bugs. The recent Windows worms exploit a true bug in the security system, but I've heard that this one was developed from access to the source that leaked.

The Linux source has been out and discussed for over a decade, with plenty of time to find truly deep bugs. With the leak of WinNT/2k source, one hole was revealed fairly quickly. As people REALLY study that source, what else is going to emerge? (And how much code was really rewritten for XP vs reused?) Note that this isn't just a function of the source leak. As Microsoft shows more with Shared Source, more people will have the kind of access needed for this type of exploit.

Re:most destructive?

[gunpowder]

This exploit, as well as the mremap ones, were derived from intimate examination of the source.

Some might disagree. This bug was discovered here and was later posted and analysed on the lkml.

Re:most destructive?

[spitzak]

No, this new bug was discovered in exactly the same way such bugs are discovered on Windows. Open source did not make it any less secure in any way.

Here is the bug report. Please examine it and tell me where they looked at Linux source code to discover the exploit:

I'm on a vacation and doing some hobby project and came across a minor fault
when using assembler inlines. Result is the my linux console totally freezes
(Can't even CAD the console, but ALT SysRQ still works). This problem must be a
race since it doesn't trigger when system is loaded, or processes is straced.
I'm not able to update my versions for some weeks due to my vacation, so
problems might be fixes. I find this both to be a bug in gcc, that again
triggers a kernel bug.

I need confirmation if this is an issue with gcc here (looks like the frstor is
off by 4 compared to the address fsave gets).

first reply:

It locks my computer yes, but when I compiled using -S flag and looked at the
assembler output, the fsave and frstore is not given the same offset in the
stackframe. fsave is -124(%ebp) while frstor is -128(%ebp).

last reply:

Thanks for the respons. Seems to be harder to get up the eyes of the kernel
people, since the floating point exceptions inside that signal handler fucks up
the current linux-kernels. Haven't got any respons from any of the kernel
developers yet. A bit scary I think. But a big thanks to you atleast for given
this some attention.

Re:most destructive?

[dpilot]

Yep. I read it after the first reply told me about it.

But it's just possible that Open Source led to the quick resolution, because there wasn't a week or two of finger-pointing before getting down to working on the real problem.

Re:most destructive?

[spitzak]

The first patches were not really solutions, they simply tested if this exploit was being tried and did an abort(), which means that slight variations would still crash it. However I wonder how many closed-source patches are of this nature? Probably a lot of them!

I can't tell from the reports whether the underlying cause has been identified and fixed yet. It sounds like it, but I am unsure.

MM patches question

[macdaddy]

Could someone tell me a bit about the MM patches? I've used Morton's patches for some time now but I never understood when the guts of his patches made it into an actual vanilla kernel release. Does anyone know? For example the last MM patch as of right now is for 2.6.7-rc3. Does that mean the vanilla 2.6.7 now contains all of MM before that? I never have quite understood that.

Re:MM patches question

[rytier]

just wait a while, Andrew will eventually update the patch as well :-)

I for one...

[98jonesd]

I for one welcome our new kernel running overlords.

Sorry.

It was obviously going to be released today.

[skeptikos]

After all, last night I downloaded and compiled 2.6.6. Darn.

NTFS changes

[bender647]

I see changelog entries to the NTFS driver. You can now safely overwrite existing Win XP files so long as you don't change the filesize? This could be a BIG timesaver to those of us who chainload linux from the NT bootloader.

2.6 kernels-Kernel Panic

[Anonymous Coward]

Well under Mandrake 10.2. 2.6.6.x is the latest that works correctly. 2.6.7.x and later, promptly kernel panic right after the bootloader.

Re:2.6 kernels-Kernel Panic

[Anonymous Coward]

You probably forget to run mkinitrd(8) before trying to boot the first time.

mkinitrd creates a file which contains a very small ext2 file system. This little file system contains certain modules specific to your computer's configuration which are needed at boot time. During the boot process the kernel loads this little file system to a ram disk from which are then loaded the device drivers required to complete the boot process.

If the proper modules aren't present then your kernel will panic. That is what is causing your problem.

You're new here, aren't you?

[Lethyos]

Really, please try to chill out. First of all, people asking "why is this important" are trolls. Considering that there is plenty of material to read on the matter (release notes, change logs, recent vulnerability alerts), it should be obvious to the inquisitive reader why a kernel release is important. It's a stupid question to ask insofar as you expect other posters to simply regurgitate what is easily accessible from the story content.

Re:Well I'll be...

[RLiegh]

But are all kernels fixed? I'm not seeing anything about them releasing a fix for 2.4

Kernel.org is insane.

[caluml]

Does anyone else find kernel.org quite insane? 1007 processes, pumping 250Mb/1000Mb, with uptimes (before they upgraded to 2.6) over a year.
Me wants root@kernel.org. Just to say I had it. :)

On another note - vsftpd is supposed to be really good, but it's not as flexible as proftpd, is it.

Re:Kernel.org is insane.

[caluml]

Replying to my post - I'd love to see munin stats for that machine.

Re:Kernel.org is insane.

[xgamer04]

Yes, it is QUITE insane. They must be running some uber-l33t hardware for that. Especially when a new kernel gets /.'ed.

Re:Kernel.org is insane.

[Anonymous Coward]

They must be running FreeBSD.

Re:Kernel.org is insane.

[diz]

Speaking as root@kernel.org, I think other sites sometimes over-design. As the main page mentions, we use a Compaq (now HP) DL380 G2 generously donated to us, and it's not as beefy as you might think.

It's just a dual Pentium III system. vsftpd was the biggest help towards lowering our load (it was exceeding 300 easily), as well as moving from our ancient distro to a modern one with proper threading and moving from prefork to worker in apache 2, and moving to apache2 for sendfile support to begin with.

Don't hold your breath on this one, but I hope to get graphs and stats posted regularly as well as site changelogs in the not to distant future -- you know useless stuff like what the maximum transfer rate with a single client was, historic rates by class, etc.

Re:English: Linux Today has human redable changelo

[DF5JT]

Did those who modded this informative actually visit the link?

"Summary of changes from v2.6.7-rc3 to v2.6.7"

I know very few people running rc3, so the only relevant changelog is the one from the previously released version, i.e. 2.6.6. A link to that one can be found directly on kernel.org and it's very much worth reading the extensive changelog, since there are numerous surprises that cannot be found in the changelog from rc3.

The extensive involvement of xfreedesktop.org contributors is worth noticing. Seems like they are very much involved in kernel development and I wonder where that leaves those with BSD or other non-Linux systems.

As if you aren't a troll?

[Anonymous Coward]

It's at least debatable whether he's a troll or not. It's 100% obvious to everyone that you are a troll.

Re:As if you aren't a troll?

[pebs]

It's at least debatable whether he's a troll or not. It's 100% obvious to everyone that you are a troll.

Troll? No.. I made an honest comment on the pointlessness of the repetitive (and recursive) bitching on Slashdot about the same things in every thread (while adding to the bitching myself, yes). Was I nice about it? No.

You'd think people would just quit reading Slashdot if they thought it sucked so bad.

Do both

[amightywind]

The kernel make/installation process smartly separates new versions. Select the kernel version (or runlevel) when the Grub screen comes up. I boot variants of 2.4 and 2.6 without issues on Gentoo. That way I keep a functioning system and can still impress the chicks with a cutting edge kernel running in VGA mode. You might even set up a parallel installation on another disk

why mod down

[vijaya_chandra]

the parent!?!

It *is* funny ( and is partially even the reality )

No software is(/can be) perfect and linux is nothing but a software.
There are a zillion devices out there and the drivers for some of them can/would be unstable. Would you deny the fact that the stable versions are only *officially* stable until around the .5 or .6 point upgrades come out?

You might call me aged but am still on 2.4.26. The slight apprehension about minor things going wrong and quirks popping out here and there still daunts quite a few people like me from considering the initial builds of the stable kernels as really stable.

No! I haven't booted in windows for the last 12 months except for playing quake

Re:why mod down

[Anonymous Coward]

You might call me aged but am still on 2.4.26.

[please insert obligatory abacus joke, complete with "whipper snappers", "knee-deep in snow" and "uphill, both ways"]

(Obligatory) Damn you!!

[vijaya_chandra]

How is a full compile of the kernel done and how long would it take on a 3GHz,756RAM computer?

Linux is for slower systems like mine ( a modest p3 with 256 megs, celeron 500MHz with 128 megs )
With a system like yours you should now be running windows with a trillion apps in the system tray so that there's atleast some equality between the haves and have-nots of this miserable world

I hope you are atleast using some bulky shiny desktop environment like the gnome or kde

Supermount

[Bralkein]

I just thought I would post a brief message about supermount. If anyone wants to upgrade to 2.6.7 and still use supermount, I don't think vanilla kernels have it in there (yet, I'm sure it'll get in there sooner or later). I'm pretty sure the Mandrake and Gentoo kernels have support for it (gentoo-dev-sources do, anyway), but I just looked at gentoo-dev-sources and it is at version 2.6.5, dunno about Mandrake, but I'm sure it will take a few days for all the distros to catch up.

If you want to upgrade for security reasons, but you also want supermount in your kernel (as I do), this guy seems to have a patch for 2.6.7, which might come in handy if you don't want to wait for your distro to catch up. I am going to use this patch myself, but I cannot guarantee that it won't bone your system so to speak. The patch is not just supermount, it looks like it has some other stuff in it too, so decide for yourself!

Seeing as how I'm posting this, I may as well give a little background for those not "in the know". Supermount is a sort of filesystem, you mount your CD-ROM and floppy drives (or even USB sticks) with it, and it will automatically mount and unmount the media when you insert or remove it, kind of like on Windows. Personally, I think it is great, and it is hard to live without it now I have it.

You can learn more about it at the project website. Jeez, if it turns out the vanilla kernel does have supermount after all, I am going to look a right idiot... *presses Submit*

Re:Supermount

[gid]

I wish it would get included in the main kernel already. It's such a desirable thing, I can't believe it hasn't made it in yet.

From the FAQ:

Q: Will supermount be included in standard kernel?

A: Frankly speaking, I do not know. Version for 2.4 kernel is still more of a hack so I would not even try to ask for it. I still do not consider 2.6 version to be ready for inclusion in mainline - although it would definitely make some things easier. If anyone thinks supermount should be part of standard kernel - feel free to ask on LKML.

Re:Supermount

[Anonymous Coward]

In 2.6.x, can't you use some combination of the hotplug daemon and udev to get similar behaviour (arbitrary user-space mount code run on a particular kernel event)?

Re:Supermount

[qtothemax]

Supermount is not included in the vanilla kernel. If anyone is compiling a kernel for Mandrake, or any other distro that gives errors when trying to mount non-root filesystems, you need to apply the supermount patch. Just in case the parent poster's link doesn't work, I got mine here, (don't let the .cx scare you) and it definately works. There is no 2.6.7 patch on that site as of yet, but i used the 2.6.6 patch on the 2.6.7 source, only applying the supermount part of the patch (this patch does more then just supermount) and it compiled and worked fine. Another point to note, I am using Mandrake 10, and coulden't get my filesystems to mount even with supermount because I didn't select the loopback module under block devices. Make sure you select this as a module if you are using supermount.

What happened to "many eyes?"

[bonch]

This embarrassing FP exception hole has been around since 2.4. Yet I still expect to hear how much the "many eyes" of Open Source magically eradicate all bugs in all situations.

Re:What happened to "many eyes?"

[catenos]

This embarrassing FP exception hole has been around since 2.4. Yet I still expect to hear how much the "many eyes" of Open Source magically eradicate all bugs in all situations.
--
People routinely mod me down for voicing an opposing opinion instead of just replying in disagreement.

Well, I would mod you "Troll", if I had mod points. That's not because you voice an opposing opinion, but because you make up stuff.

I don't know whom you refer to by "how much the many eyes of Open Source magically eradicate all bugs in all situations." [emphasize mine], but it's surely nobody seriously involved in the FS/OSS community.

I have never read anyone making such a statement (even from fanboys here on Slashdot) except when being implied by people like you.

Re:What happened to "many eyes?"

[Anonymous Coward]

YHBT. YHL. HAND.

Love,
bonch (aka Overly Critical Guy)

Re:What happened to "many eyes?"

[catenos]

I would mod you "Troll"
YHBT. YHL. HAND.

I have been trolled? Nice attempt, try again.

In case anyone wonders why I replied to a troll: This is Slashdot, which resulted in the trolling post already getting modded as "Interesting", i.e. to some mods you have to explain that it's a troll and why.

[Sorry for the off-topic]

If Microsoft had been churning out *quality*

[vijaya_chandra]

software (or atleast patches) like the linux kernel developers do, I would have modded you +100 Insightful

But sadly that isn't the case

I am not sure if I am allowed to post this publicly but the following was part of the mail for XP SP2 from betamail@ms.


Known Issues
Build 2142 is not considered Beta quality and as such it does contain some issues we expect some people will enounter during usage.

1. Media files that are encoded with the DIVX codec will fail to play. If you open an explorer window in a directory that contains these files, explorer may fail. Viewing web pages with DIVX encoded files embedded in them may cause Internet Explorer to fail.


This is insane. Why rush out a build ( even if it is officially non-beta-quality ) if there's a problem like this? One would be extremely p*ed off if their explorer suddenly crashes because of some video in some folder, considering the popularity of the format now a days. Or should I be happy that they have atleast warned us of known issues

Re:If Microsoft had been churning out *quality*

[Anonymous Coward]

The problem is, Linux developers do this too. Not to pick on Fedora, but they released an entire... uh... whatever the increments between Fedora Cores are, without Firewire support.

Can you tell me when you are planning

[vijaya_chandra]

to move to 2.6.7?

so that I can decide whether to jump on to 2.6.7 today or wait for the next one

windowmaker anyway would be

[vijaya_chandra]

fast irrespective of the kernel

what other *interactive* software do(/can) you run on the p133??

Re:windowmaker anyway would be

[MartinG]

xfce and firebird mostly. sometimes abiword.

other stuff i use such as vim is always blisteringly fast regardless, but xfce and firebird just felt faster.

Is it just me...

[Anonymous Coward]

Every time I see an acronym with a J in it I think it must be some Java API.
JFS = ? Java File System?
Java in the kernel? Say it isn't so!

As a chick...

[posu256]

As a real flesh and blood chick, the fact that you would even spend time thinking about this... Chicks don't care what kernel you run - they want you to stay away from the computer and do things with them, darn it! Its fine to spend brain cycles on upgrading your OS, but it better be the lowest priority process in the task list. If you enjoy your machine's companionship more than your chick's, then you don't deserve a GF.

Re:As a chick...

[Nutria]

Chicks don't care blah, blah blah...

Methinks you dropped your sense of humor that day you chose Womens Studies as your college major.

But then again, since "regular" chicks really do care about superficial things like what kind of clothes that guys wear(*), why shouldn't a guy geek think that geek chicks care about superficial stuff like uptime?

(*) Be reasonable: I'm talking about basic stuff like whether the clothes are clean.

Re:As a chick...

[posu256]

Actually, my major was Electrical Engineering. I would think one of the defining traits of being a geek is avoiding the unecessary. When it comes to an OS, the overwhelming philosopy is "If its not a necessary feature, get rid of it." I'm sorry if I find the SO's uptime or kernel version a completely unecessary relationship feature. Things like actually enjoying the other person's company in the absence of any technological gadget would be a good start.

You got me.

[halivar]

That's a pretty cogent reason to upgrade from 2.6.3, as I have been having probs with sound. Still, if you already have 2.6.5, 2.6.7 may be an unnecessary upgrade.

Now, once they put out a fix for that 20-line shell exploit, I imagine everyone will have a good reason to upgrade.

Uptime

[bstadil]

Ask her, I will bet uptime comes out on top (sorry) ;-)

Holy Crap!

[xgamer04]

A 760 KB changelog!

Re:Holy Crap!

[Lennie]

Maybe partly because of the new format:

From: Andy Whitcroft

[..some describtive text..]

Signed-off-by: Andy Whitcroft
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds

The good thing that came out of all the SCO-crap.

Doesn't compile for me?

[martin_b1sh0p]

Hmm...I downloaded 2.6.7 and did a make oldconfig, said N to most of the new stuff (all but one if I remember correctly) and during compile I got:

drivers/scsi/cpqfcTScontrol.c:610:2: #error This is too much stack
drivers/scsi/cpqfcTScontrol.c:722:2: #error This is too much stack

Oh well, I'll stick with 2.6.6 for now...I'm not running any "lame free-shell access" accounts on my box anyway :-)

Re:WHAT?!?!?!?!

[Anonymous Coward]

Blech...
You're one of those Windows guys who feels threatened every time someone mentions Linux, right?

Re:Beware the Toll

[Anonymous Coward]

There you go, modded you down as Troll. And guess what? Before the others! Sore loser...

Incorrect, I was modded down Flamebait.

Size matter

[Anonymous Coward]

Sure size matter

Just my 10",
_Art

Re:English: Linux Today has human redable changelo

[sineltor]

Human readable eh?

Shirley Ma:

* [IPV6]: Initialize pmtu/advmss in ndisc dst entries

*MUCH* better :]

No.

[Inoshiro]

I still can't sync my iPod. It takes longer to break, but sbp2 is still fucked for fw HDs.

Re:English: Linux Today has human redable changelo

[gujo-odori]

Little did we realize what a stunning revelation would be included in this changelog:

The most notable change may be the one-liner that should fix the embarrassing FP exception problem.

Personally, I had no idea that it was possible to get a First Post in the Linux kernel, let alone that it could cause an exception.