Linus Torvalds released the 2.6-test6 development kernel with a number of exciting changes. Most notably this includes some impressive performance enhancements written by an Australian doctor, Con Kolivas. Additionally, it includes Al Viro’s “32-bit dev_t” patch, bumping up the number of device types that can be allocated by the Linux kernel.
Linux 2.6.0-test6 Development Kernel
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
straight from http://kerneltrap.org/node/view/465 …
I’m a little embarrassed people get me confused for a kernel hacker, as my real profession is very remote from IT. I’m a doctor; a specialist in anaesthesia.
Just as I finaly broke down and installed test5.
Ah well, here we go again.
Lots of the -MM patches have been encorporated into “Test6”. Does anyone know if Morton will be making an -mm patchset for test6?
If he is, i’ll wait for that.
> Lots of the -MM patches have been encorporated into
> “Test6”. Does anyone know if Morton will be making
> an -mm patchset for test6?
>
> If he is, i’ll wait for that.
He’s been releasing -mm patchsets since 2.5.17[ http://kerneltrap.org/node/view/875 ]. I’m sure he’ll release a test6-mm patchset.
Id just like to know how long until we actually get a release version! Oooooh boy, I can’t wait… =))
The real question is how many kids will I have by the time 2.6 is in Debian stable?…
I’m running 2.6-test5-love4 at the moment. It’s rock solid, all of the hardware that I’ve tested works with it and it definitely feels faster than 2.4. Very impressive indeed
Don’t you mean
a) grand kids, and
b) Debian testing
> The real question is how many kids will I have by the time 2.6 is in Debian stable?..
make-kpkg?
Do you want to know how long I had too wait before I’d a system completely based on Linux kernel 2.6-rc4 and gcc 3.3, gnome 2.4?
One day (building), saturday a week ago. This doesn’t include Ximian OpenOffice 1.1-rc5 which added a another day (building) all on its own (T-bird 1 Ghz).
Con have been working on smoothing out the interactivity patches ever since 2.6.64 when they were introduced and people began to complain about skipping XMMS etc, he knows enough code to do his thing, amazing what you learn by just looking at others code.
Only thing holding me back from switching to 2.6.0-test on my main workstation is the fact that my volume-up key doesn’t work with hotkeys anymore (due to the input layer changes I assume).
As soon as I find new function key software with an OSD I’ll definetly have nothing but 2.6 on my Linux machines.
Depends on how stable the next test kernels are – less time mucking around with errata means more time to get busy with your wife.
Just use xbindkeys and osd_cat:
http://hocwp.free.fr/xbindkeys/xbindkeys.html
with a little work you could have it showing the percent as you change the volume. The plus is that it works in X.. not just kde.. or Gnome…
thanks, that was useful!
incase osd_cat does not do it.. maybe this:
http://freshmeat.net/projects/osdsh/?topic_id=861%2C113%2C1…
I just hope that it will handle my keyboard (USB) and mouse (USB) well. I am definitely longing for a distro which supports the new kernel from the installation on, because I tested about 5 distros with the 2.4 kernels and the only one that handled them correctly (after passing “nousb” to it) was Knoppix.
<Keeps fingers crossed/>
i have a btc usb keyboard+oem logitech usb mouse, a logitech wireless usb keyboard+mouse, an anonymous 3 button usb mouse without wheel, an old logitech mouseman plus usb, and i have seen a bunch of other usb mices.
they all have always worked flawlessly with every single kernel since i started using linux.
what kind keyboard and mice do you have????
I just tried to install 2.6.0-test6 on my Slackware 9.1 box. When I installed Slack I chose ReiserFS for all my partitions.
I’ve compiled test6 and it boots *for a while* but I get a problem stating that it doesn’t recognize the FS on my root partition, even though I’ve compiled in ReiserFS support (built-in). Anyone have suggestions?
Also, I’ve tried to install ATI’s official drivers for Slack witj 2.4.22. But they will not install due to dependencies problems. Is the kernel too new? Should I go with 2.4.21 perhaps? Any thoughts?
Thanks!
I cannot connect to my ADSL provider using both my redhat 9 distro and a kernel, other than those provided by redhat.
I’ve already tried to find out whether the redhat kernel has a certain ppp patch but I couldn’t find one.
The error message is something like “NET: event not supported” (or something similar.), from /etc/hotplug/net.agent
(What does hotplugging have to do with DSL ??)
The problem is the same with both 2.4.x kernels and 2.6.0-testx kernels.
Is anybody out there with this distro, a 2.6.0-test6 kernel and a ADSL-internet connection who mad it to connect to the internet ?? I really spent hours and a lot of questions in many newsgroups/forums, but nobody could help !? Hope you can, since it’s the only reason for me not to use this great piece of software!
greets Boris
It is a kit (comes with cpu cabinet, keyboard, mouse, sound boxes) from http://www.dr-hank.com.br/.
🙂
Well all my connections are or by a regular modem or by cable, perhaps, your adsl modem is nothing being properly recognized.
I have heard about problems with some of them (specially USB ADSL modems).
ISDN is still not working in 2.6. There is some pending rewrite or something.
Check to see that you have PPPoE configured in your kernel. Then your going to need a PPPoE client… roaring penguin or something like that…
If you are using an initrd that may be that parition that is failing mount. To get Redhat to boot with 2.6 I had to change Default RAM disk size to 8192 from 4096 because modutils is now too big to fit on the old size.
“Check to see that you have PPPoE configured in your kernel. Then your going to need a PPPoE client… roaring penguin or something like that…”
Everything is configured properly, since it works with the redhat kernels. I think redhat does not use the kernel pppoe driver (the module is NOT loaded, when being connect). Booting another kernel results in not being able to connect. What patches did redhat apply to their kernel, if there are any ?
greets Boris
He’s only 32. He’s a physician. He’s married. He doesn’t know how to program in C, yet he’s making/applying/submitting Linux kernel patches?!
Ahhh, of course,.. this all adds up now. He’s obviously made a number of clones of himself. One’s studying medicine (and presumably seeing patients), one’s the kernel hacker, and one’s at home spending quality time with the wife and family.
Sometimes when someone hands you an apple and it looks like an apple.. and it tastes like an apple.. you just have to trust that it’s an apple.
He has since taught himself C, recently too, and is now contributing patches – wow. He’s an amazing fella to be doing all of that….
Of course he can code… but he claims not very well, sometimes it takes quite a bit of knowledge make uncooperative/outdated patches apply and not break each other.
Con’s main contibution to the 2.6 kernel is the 01int interactivity tuning. Although, I am not sure if that has been included.
“Con’s main contibution to the 2.6 kernel is the 01int interactivity tuning. Although, I am not sure if that has been included.”
FYI, O1int has been included in 2.6.0-test6
> If you are using an initrd that may be that parition that is
> failing mount. To get Redhat to boot with 2.6 I had to
> change Default RAM disk size to 8192 from 4096 because
> modutils is now too big to fit on the old size.
Actually, Slackware doesn’t use an initrd by default. It is all done with a lot of nifty BSD-like scripts. What he should do is check that he has specified ReiserFS support in the kernel, preferably as built in (a ‘*’ symbol in menuconfig). He should also make sure that he is passing a correct “root=/dev/hdaX” parameter to the kernel. This can be checked by typing “mount | grep reiserfs” and using the device value appearing in the far left hand column which matches “/boot or /”.
If someone could point me in the right direction of a doc that will help me compile and install 2.6test that would be great. I’ve looked around and have not been able find what i’m looking for (google search<linux>, linuxquestions.org)though I’m sue that’s due to my overlooking somthing. I’m running slackware 9.1 if that makes any difference thanks.
Newbie jumping in head first
The article entitled “Feature: HowTo Upgrade To The 2.6 Kernel” at kerneltrap.org should tell you what you need to know.
http://kerneltrap.org/node/view/799
Otherwise, you can checkout the Kernel-Howto:
http://www.tldp.org/HOWTO/Kernel-HOWTO.html
Well I just hope Andrew Morton doesn’t have too much of a say in this, I tried his mm sources and ended up in segefault city. It was disasterous. I’m suprised he has been selectecd by Linus as the new 2.6 maintainer. His last effort was awful. (Sorry, but for me it was).
Q
Andrew Morton’s patchset was very stable for me, as usual YMMV, segfaults are mostly due to defective hardware or user induced errors. IMHO Andrew is t
he best man for the job.
So, because you happened to have problems with a BETA-kernel, you decided that Andrew Morton sucks and he should NOT be in charge of 2.6-kernel? uh-huh….
Well I just hope Andrew Morton doesn’t have too much of a say in this, I tried his mm sources and ended up in segefault city. It was disasterous. I’m suprised he has been selectecd by Linus as the new 2.6 maintainer. His last effort was awful. (Sorry, but for me it was).
Funny that you are beating a dead horse in regards to Andrew Morton whilst completely forgetting/ignoring that the -mm branch isn’t rock solid or stable. It was never meant to be and never will be. The -mm branch is Andrew Mortons fork which allows him to test out bleeding edge additions, which are then tested then merged back into the main kernel tree.
You are the one with the big problem, no one else. If 99% people get the clue that the -mm branch is unstable and bleeding edge, and you, the lone whiner of osnews.com don’t, who do you think is correct?
You say something doesn’t work and they have a fit. Well it didn’t for me and I’m sorry about that – but really it has nothing to do with me.
My harware isn’t faulty, I tried it out on 3 of my machines at home all by different manufacturers and it always segfaulted at the same point. As for user induced error, all I did was type emerge mm-sources. I use Gentoo and it really is that simple. So where I wonder is the error in that?
You know what, scr*w you. I made a legitamate comment that if this was the way the kernel was going to eventually go then at this point it was not immediately encouraging.
I hate this elitist in denial ‘how could anything possibly go wrong, you must be the dumb ass’, elitist crap.
It’s attitudes like that that put people off switching to Linux in the first place. I have been using Linux for 14 months now – and I swear if I keep hearing it, it’s going to put me off one day too.
Q
2.4.20 recognized none of my 27 different hotkeys (not all individual keys, but just 2nd action on F1-F12), all keys gave Unknown scancode but still xbindkey could use all of them.
2.6.0-test4 failed to recognize 6 hotkeys, and gave the same scancode to two different keys. I couldn’t assign keycodes to the recognized scancodes with setkeycodes. Also xbindkeys was not able to recognize the unknown scancodes.
Sweet. Helps alot. Thanks.