Linux 2.6.0-test11 Kernel Released

An anonymous reader writes "Linus Torvalds has released his final 2.6.0-test kernel, calling it the 'Beaver In Detox'. Following this release, Linus says that 2.6 development will be led by Andrew Morton. The kernel's name refers in jest to the previous release, which Linus had named "Stoned Beaver". It contains a fix for the aic7xxx driver, proper error handling in do_fork(), some firewire fixes, and correction of a few skbuff leakage points. Download it from a kernel.org mirror."

Comments

For those who don't RTFA...

[HoldmyCauls]

No word on a resolution for the problem with pre-empt, at least none that can be seen with a search for "preempt" in the page. :-P

Newbie comment

[Murmer]

If you're a new to all this, and you're a little intimidated by all the options you get from running "make menuconfig" or "make xconfig", here's a tip - if you're running Knoppix, or have one of your current kernels around in /usr/src, you can pull over your old configuration files from /usr/src/kernel and just copy them in the new kernel's source directory, if you're willing to compile the source.

Then run xconfig, make sure you've got console support compiled in (this point is key) and off you go. make, make install.

It's worth saying again, no matter how many times it's been said, that Linus and his band of merry coders are doing incredible work. On older machines, like my humble P2/266 laptop, a recompile actually feels faster, a lot faster in fact. If you're running older hardware, it's worth it to give these test kernels a try for that alone, if nothing else.

Re:Newbie comment

[FueledByRamen]

Actually, what you really should do, instead of "make xconfig" after pulling over the old config file, is "make oldconfig". It's a version of the (original, text-based) config that only asks you about configuration items that weren't defined in the file - like if a new feature (and thus a configuration item) is present in the newer kernel. BTW, the name for the kernel config file is .config (it's invisible - note the dot)

The pre-empt problem:

[ciaran_o_riordan]

A full discussion of the pre-empt problem can be found here:
http://kerneltrap.org/node/view/1679

This is probably what has been happening to me.
I tried up to -test9, but after a few (2-10) hours, my cpu and disk access would shoot up and RAM usage would slowly increase. I dropped to a tty, typed the username "root", but the cpu load was so hight that login timed out before I was even prompted for a password!
So I had to kill the power on my machine while I could hear that my disk was fully active. I was not a happy camper.

Re:The pre-empt problem:

[Daniel Serodio]

If you really, absolutely must kill the power, use the "Magic SysRq" key! Alt+PrintScreen+S will sync the filesystems, and Alt+PrintScreen+U will remount them read-only.

See /usr/src/linux/Documentation/sysrq.txt.

Re:The pre-empt problem:

[jesdynf]

That's /if/ you compile it in, and if my memory servers there's some cautions against doing that needlessly.

Re:The pre-empt problem:

[SpaceLifeForm]

It's under the hacking submenu, 'Magic SysRq key'.
Also known as CONFIG_MAGIC_SYSRQ in your .config.

Re:More Information

[golan]

Also check out this to see stats about the different 2.6.0-testX versions compile warnings and errors by John Cherry from the OSDL.

Re:Question

[Nucleon500]

You wouldn't think so, but there is a good reason for it. To make a sweeping generalization, Slashdotters want to see Linux succeed, and we have the technical knowledge to compile kernels and make informative bug reports if needs be. Publishing this story on Slashdot will entice more people to test the new kernel, thus ensuring that the 2.6.0 release will be Bug-free(tm).

Re:More Information

[squiggleslash]

Also, if you're at all unsure of the process used to develop the 2.6.0-testX kernel, there's a handy diagram here...

The importance of publishing test releases

[Anonymous Coward]

When you put 1 milion monkeys with keyboards in a room, open source occurs (taken from the Bazaar and the Cathedral), or a result of similar quality.

Since most open source development is not done by professionals nor professional organizations, there is no formal QA process. As such, Linux is in dire need of testing. Especially with all the unproven claims of stability made in its regard.

It is therefore very important that as wide a variety of people as possible download it and hose their systems so that bugs get reported in newsgroups where people will get laughed at for not reading every bit of manual out there.

I think that the people posting complaints about this story were either ignorant of the open source development process, or just trolling.

Thank you.

Re:The importance of publishing test releases

[Hieronymus Howard]

Since most open source development is not done by professionals nor professional organizations, there is no formal QA process. As such, Linux is in dire need of testing. Especially with all the unproven claims of stability made in its regard.

Have you looked at the email addresses of the kernel contributors. Most of them seem to be @intel.com, @ibm.com, @hp.com, @redhat.com, @sgi.com etc. etc.

Maybe ten years ago they were teenagers, but mostly they are now professional developers who are paid to work on the kernel by their employers.

HH
--

Is fdomain driver still actively maintained?

[pedro]

My only need for scsi is for my venerable old hp scanjet 4 and a cd burner, so this ancient future domain isa card works just dandy.
Unfortunately, the driver segfaults under 2.6.0-x. It worked a treat with 2.4.
Anyone else had this issue? Should I make a formal bug report? Is it a hassle to do so?

Re:Is fdomain driver still actively maintained?

[Anonymous Coward]

It's a known issue that many older drivers are busted. You could report it, but they'll probably just tell you to fix it yourself and submit a patch.

Re:non-patched distro kernel

[Azar]

Slackware is a completely vanilla kernel. Most others (but not all) tweak the kernel one way or another.

Torrent

[teoruiz]

As usual, I set a torrent here.

Re:non-patched distro kernel

[Pop69]

You can download and install a vanilla kernel in Gentoo and the Slackware kernel is unpatched out of the box apart from the occasional security issue like ptrace.

As you say, most other distros are patched to within an inch of their life.

How to report kernel bugs

[Anonymous Coward]

Go to bugzilla.kernel.org, create an account and post all the info you have about the bug. Don't be afraid of posting huge blocks of text (and if you think it's too large, you can add it to the bug report as an attachment). Post dmesg output, kernel configuration files, the output of programs like lspci -v or cat /proc/interrupts, and everything else you can think is useful.

Since what you getting is probably a panic or a oops, if it does not have already the symbolic function names, run it through ksymoops to add the missing data, and post the output.