Slashdot Mirror
New Testing Version Of Linux 2.6
James A. A. Joyce writes "It's all up now at the kernel archives. Get the full 2.6.0-test2 or a patch, whichever suits you. We need to test those new kernels! Hop to it!"
← Back to Stories (view on slashdot.org)
Being a LKML lurker, here are a few of the new features.
t y/patches/Module/
0 .3/0793.html
1 .3/1267.html
4 .1/0832.html
3 610918825614&w=2
3 553654329827&w=2
3 498293902006&w=2
In-kernel Module Loader and Unified parameter support: http://www.kernel.org/pub/linux/kernel/people/rus
Nanosecond Time Patch: http://www.ussg.iu.edu/hypermail/linux/kernel/021
Fbdev Rewrite: http://www.uwsg.iu.edu/hypermail/linux/kernel/011
Linux Trace Trollkit (LTT): http://www.uwsg.iu.edu/hypermail/linux/kernel/020
statfs64: http://marc.theaimsgroup.com/?l=linux-kernel&m=10
POSIX Timer API: http://marc.theaimsgroup.com/?l=linux-kernel&m=10
Shared Pagetable support: http://marc.theaimsgroup.com/?l=linux-kernel&m=10
Hotplug CPU Removal Support and Kernel Probes
Bethanie: Whore...
Fan Whore
http://marc.theaimsgroup.com/?l=linux-kernel&m=105 932590109238&w=2
And you're fired.
Use Debian, make-kpkg is your friend!
Its not a kernel issue as it already supports USB. You'd have to have a service(daemon) running that would scan the USB ports on your computer for any changes and then have it load a program to download the pictures off your computer.
In linux libertas
I am running an AMD 1600+ and a Geforce 4200. I have run 2.4.18bf, 2.4.20, 2.4.21, and 2.6-test1 without problems.
"To any truly impartial person, it would be obvious that I am right."
Compared to 2.4.X these are the problems i've seen. No cpufreq scaling, alsa doesn't work too good and preempt doesn't work properly (mouse stutters).
Hail to the king, baby!
I have no information regarding running it on "stable".
Bruce
Bruce Perens.
I just built and booted it on my Sun Ultra 60. CONFIG_PREEMPT seems broken on SPARC64, but after I removed a redefined symbol in ffb_drv.c, everything else seems to work.
/dev/sunmouse::BusMouse to /dev/psaux::PS/2 in XF86Config, but the middle mouse button doesn't work now. Gonna figure out that later when I get a chance.
One interesting note is that the sun/type5 keymap/keyboard settings for the console/XF86Config don't work anymore. Had to revert to the kernel keymap for the console, and xfree86/pc101 for X11. Also had to change
I use Linux for a very multipurpose server... SNAT, pppoe, tokenring+ethernet+atm+arcnet+slip, apache+php+perl, postgresql, Nvidia and hordes of other stuff I cant think of now. The test1 crashed for me as I was configuring the networking portions.
For setups like me, I couldnt test Linux beta versions. The server is really not mission-critical but I believe 2.6 will keep crashing for me till version say 2.6.15 or something. I'll be trying to add my contributions to the community but not at such an alpha stage.
"Give orange me give eat orange me eat orange give me eat orange give me you." -Nim Chimpsky
The Very Verbose Debian 3.0 Installation Walkthrough
This will give you a minimal but extremely stable install. If you want newer stuff you can shift to "testing" or "unstable". Unless your are doing cutting edge stuff, this install is bulletproof!
"I say we take off, nuke the site from orbit. It's the only way to be sure."
2.4.xx is the current stable release 2.5.0 = the current relase of 2.4.xx at the time they started the development series The last 2.5.xx became 2.6.0-test1 which sooner or later becomes 2.6.0
If religous zealots don't believe in Evolution, then why are they so worried about bird flu?
It worked under a completely unpatched Debian Woody DVD-R install, so I would say yes. I had some problem with VT support, but I think that was a configuration issue with the kernel rather than a dependancy problem.
You could use modutils and, say, gphoto2 along with a script to do it for you.
Since there's likely a specific USB driver, just insert a post-install line (post-install drivername command ) and when you plug it in, the module handler will start things going for you.
Since you're the one writing the script (which doesn't really have to be more than a gphoto command telling it where to put the pictures), you can make it as robust as you like... for instance, set up a file with camera ID to username mappings, and drop stuff in ~user/.photos-incoming or something.
If it's a camera that Linux sees as USB storage, the same thing pretty much holds, but you'll have to just run cp and put them where you'd like... no need to compile gphoto.
To bring this back on topic, I can't wait to check out 2.6, but I'm still in the middle of configuring test1. ;-(
Don't think of it as a flame---it's more like an argument that does 3d6 fire damage
err, wouldn't that be the 'b' and 'B's ? 57 KBytes at 512 Kbits/sec ?
Grab kernel 2.6.0-test2 via Bittorrent here
The scheduler in 2.6 is still having some problems with interactive programs; XMMS skips frequently when switching desktops or running a CPU intensive program in the background. Ingo Molnar put together a patch which seems to fix this nicely, I highly recommend patching the kernel before using it if you're running linux on the desktop. the patch is available at http://people.redhat.com/mingo/O(1)-scheduler/sche d-2.6.0-test1-G6
Never try anything new on your production machines.
Have a test server configured exactly (or as close as you can get) like your production server.
Always test new software on the test machine before putting it on the production machine.
This is important whether you're running Linux or Windows or whatever. Even when you're testing new versions or updates of apache or PHP or perl or postgresql or whatever.
Running untested (by you) code on production servers is guaranteed to bite you eventually.
Just save yourself the lost time and headaches and get a test box.
Your time and data are worth far more than the cost on a test box.
James A.A. Joyce. I should have been keeping closer track I guess, but I was still trying to do 2.6.0-test1 with the -mm2 and the -int7 patches. Thx for the info, I'll get back to you on it!
FWIW, I read the kernel ML daily and I can say that the kernel team strongly desires end user feedback in the form of commentary and bug reports (not flames!)
They're looking for all the possible "oddball" cases (AKA "corner cases") and also for extreme loads such as huge files/filesystems, throughput, mem and net bandwidth, fs integrity, etc.
Regular users should participate; Linus has said that his focus is still on desktop usability and responsiveness. Anyone else who is interested should make sure to enable the -preempt option as well -- amazing performance!
Damn... now I need to try -test2.
C|N>K
That is because you you didn't enable virtual terminals in your config. Make sure you have the options CONFIG_INPUT=y, CONFIG_VT=ys, CONFIG_VGA_CONSOLE=y and CONFIG_VT_CONSOLE=y.
Untar kernel sources
/path/to/kernel/sources
cd
make menuconfig
make bzImage modules modules_install
cp arc/[your arch]/boot/bzImage [appropriate place]
modify lilo/grub config, run lilo if required
get the new module-init-tools and install them (debian can apt-get them I believe, gentoo just emerge them)
reboot, enjoy.
Make sure you have Character Devices->SUpport for console on Virtual Terminal enabled.. If you don't, you wont see anything.. But you kernel actaully was booting.
Slashdot is like Playboy: I read it for the articles
I tried running xosview, but it hung. gdb shows:
/usr/bin/xosview
/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.3/libstdc++ .so.5 /usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.3/libstdc++ .so.5 /lib/libc.so.6
bdonlan@bd-home-comp bdonlan $ gdb `which xosview`
GNU gdb 5.3
Copyright 2002 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB. Type "show warranty" for details.
This GDB was configured as "i686-pc-linux-gnu"...(no debugging symbols found)...
(gdb) run
Starting program:
(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...
[time passes...]
^C
Program received signal SIGINT, Interrupt.
0x4017af39 in std::basic_istream >& std::operator>> >(std::basic_istream >&, char*) () from
(gdb) bt
#0 0x4017af39 in std::basic_istream >& std::operator>> >(std::basic_istream >&, char*) () from
#1 0x0805d6fb in strcpy ()
#2 0x0805d195 in strcpy ()
#3 0x080561d4 in strcpy ()
#4 0x08054949 in strcpy ()
#5 0x08055a9a in strcpy ()
#6 0x402217a7 in __libc_start_main () from
top shows it comsuming a lot of cpu time while it's hung.
No problems on debian stable... I'd just moved the backup server to 2.6.0-test1 for "testing" and two days later found out I had one day to backup all of our systems PRONTO for a poweroutage for the entire weekend.
Video is a little twitchy if you're running an ATI board and a KVM switch. (It wigs out on first switch.) Other than that, it's fast and rock solid. USB-2.0 works with an external drive.
I wrote 9 DVD's with it and copied 30 G's over to a usb drive without a problem. 4 of the 9 DVD's were backups from over a network.
Can't wait to throw test2 on the box.
Oh, one wierd clock error, but it didn't crash anything out.
-=fshalor
Yes, it was fixed: AGP GART pages are removed from the linear mapping with change_page_attr.
http://www.kernel.org/pub/linux/kernel/v2.6/Change Log-2.6.0-test2
Turns out that in the "Input Device Support" section, you must have "Input devices" compiled in. I had it set to module, and unfortunately that doesn't cut it.
Once you've set that, enter the "Character devices" section. At the top there will now be a "Virtual terminal" item which you can select. Once you've done that, you'll get the familiar scrolling of kernel messages.
I didn't get any compile errors, but then I went into the top level makefile (2.6.0-test1/Makefile) and changed all "gcc" and "g++" entries (3 total) to gcc-2.95/g++-2.95 which could account for something. You can also run "make V=1" to get the old, verbose output so you might get a better idea of what's causing the problem.
The dependencies thing I'm still working out myself. I suspect it's why the sound's not working yet.
In all, I'm quite pleased with 2.6.0-test1's performance on my system.
Cooper
ALSA works great, as far as I can tell, although I am doing everything thru the OSS compatibility modules. Preempt is awesome. If you're having problems with it, I'm guessing you're using Debian, and you need to un-nice your X server.
"Champagne for my real friends - and real pain for my sham friends!" http://ericblade.postalboard.com/