Linux 2.6.5 is Released

lamont116 writes "Featuring a 367.6 KB changelog, the next Linux kernel is now ready for action. As always, enjoy!"

Performance...

[Denny]

I saw a fairly good set of benchmarks a couple of days ago showing that the 2.6.x series is doing a lot better than 2.4.x as a SQL server and as a fileserver, with minimal losses in some other areas. It looked pretty impressive for something that's still on fairly early versions, so I was planning on swapping over this weekend... although I guess I'll wait for the Debian package to update to .5 now.

Re:Performance...

[Denny]

Here's the article [2cpu.com]

Re:Linux Changelog Email Publishing

[Kourino]

It's only a problem because people post the damn thing to Slashdot, really. I mean, all these people already get spam because they post to LKML, so having the slight added exposure in the Changelog probably really isn't a big deal compared to that.

(Actually, the last time I posted to LKML, I didn't get spam, so the stated problem may be even less than you might think.)

Re:A day?

[dolo666]

> Uhm, a day? I'd say about 20 minutes + adding the various email addresses to a db...

Yes, but creating code to allow users to add themselves and perform email redirects using the form might take longer than 20minutes. For real shit-stick code, how about 4min? For industry ready code, I never take less than a day on any app (just my ethics, sirrah).

Stable?

[hanssprudel]

I love the speed increases that the 2.6 kernel has achieved on the desktop (and for things like media: mplayer never bugs out with that charming "YOUR COMPUTER IS TOO SLOW" message anymore). However, I don't know if it can be considered even remotely stable. Since switching, my uptime has been a Windows like joke.

For example:

- The conversion to ALSA works great, but the modules for OSS compatibility segfault whenever an app tries to use them. Segfaults in the kernel are fun! There is pretty much nothing to do but reboot after that.

- Firewire and sbp2 support is completely broken. Ironically this has, I believe, more from "experimental" in 2.4 to a normal feature, yet it worked fine before and now doesn't work at all (the linux1394 forums forums reflect that I am not alone in this). Trying to copy data to sbp2 drives segfaults, hangs, and worse. Beware of connecting to 2.6 if you have a firewire drive with data you hold dear...

I'm sure there is more, but I am forced to return to the land of 2.4 most of the time. Now, I'm not complaining about the quality: if I want working 1394 drivers I ought to write some or shut up about it, but I am questioning to what extent 2.6 should have been released, if even after four releases basic things are completely broken...

Not the same as -rc3

[Kourino]

Linus' announcement on LKML has a "Summary of changes from v2.6.5-rc3 to v2.6.5", so no, this is not the same as -rc3.

Re:Where can i browse online through the source?

[SkiddyRowe]

You can browse the 'old' source Here [linux.no] but it doesn't have up to the 2.6.5.

The highest it goes is 2.6.1. But it has all the architectures.

Re:Instructions for 2.4 to 2.6 upgrades for Luddit

[Anonymous Coward]

debian stable (woody) isn't 2.6 ready, since it lacks the package module-init-tools in its sources, you should try sarge, it's fairly stable and fairly up to date

802.11g support

[egrinake]

Drivers for the prism54 [prism54.org] chipset have finally been merged, which means that the vanilla kernel now has support for 802.11g (54 mbit) wireless lan. The prism54 chipset is used in whole bunch [prism54.org] of 802.11g cards.

Re:Instructions for 2.4 to 2.6 upgrades for Luddit

[mick29]

Not that hard, actually.

Just get the packet "module-init-tools" from www.backports.org installed, if you plan to use modules.

The configuration dialogs are heavily restructured, but you'll find your stuff I guess. "make-kpkg" works fine.

If you have some rather peculiar stuff running, like LVM, there are some undocumented pitfalls, though.

Re:Instructions for 2.4 to 2.6 upgrades for Luddit

[neodymium]

With debian and apt-get, thats easy. Just choose from one of the many sites offering testing or unstable backports for the stable distribution (I recommend www.backports.org), add their site to /etc/apt/sources.list, do apt-get update && apt-get dist-upgrade, and you are finished...

Re:Alsa with Intel8x0 ?

[cheesybagel]

NForce card here, using the 2.6.3 kernel from Fedora 2 test 2. I do get freaky sample rates from the drivers ocasionally and everything plays speeded up or down, but "Sound & Video -> Volume Control" works well. The problem is the PCM volume by default is 0 and if that is 0, the master volume control doesn't do anything.

Re:Installation?

[Pros_n_Cons]

http://fedora.artoo.net/faq/#Kernel26 Or upgrade to FC2 test 2 like I did to help squash out bugs.

Re:Scheduler?

[Anonymous Coward]

I had this kind of problem with the kernel 2.6.3-3mdk. I update recently to the kernel 2.6.3-7mdk and the problem went away.

Re:Alsa with Intel8x0 ?

[FyRE666]

I've seen a lot of Intel8x0 fixes in the changelog...

You don't have to upgrade your kernel to install the latest ALSA drivers. Just download the source from the ALSA site [alsa-project.org], build and install it. I never use the ALSA drivers in the 2.6.x kernels (they never seem to work correctly for me, if at all). I never have problems with the official source versions though.

BTW, if your card is working ok with OSS emulation, what's the problem?

Re:Yeah !!!

[rastakid]

Will this fix bugs in KDE?

No, since this is only the Linux kernel, it only fixes kernel bugs. However, if KDE didn't function properly because of bugs in the kernel, this could be solved by this new kernel release.

Re:version.h

[Kourino]

Uhh, version.h is a generated file. It'll get rebuilt if you (e.g.) make menuconfig. (Actually, 'make prepare0' will work, as will ... well, 'make include/linux/version.h' ^_^ )

Re:Stable?

[rudmer]

- The conversion to ALSA works great, but the modules for OSS compatibility segfault whenever an app tries to use them. Segfaults in the kernel are fun! There is pretty much nothing to do but reboot after that.

Well OSS (Open Sound System) is flagged depricated, so it is going to fade out. As you seem to have problems with it why just not do it completely and also drop the OSS compatability.
I have dropped it months ago and never had problems!

Re:Stable?

[l-ascorbic]

Funny, I never managed to get 1394 working until I did move to 2.6. It didn't work at first, but disabling eth1394 got sbp2 working. I can now use my Mac, HFS+ formatted iPod with gtkpod. I never got HFS+ working with 2.4. On the other hand, I haven't managed to get ALSA working at all yet. The biggest problem I had with the upgrade was with USB (for my mouse). In the end I dumped uhci in favour of usbmouse. This is all with a VIA Epia-M Mini-ITX board.

Re:Is Slashdot a kernel.org announcement board?

[Anonymous Coward]

because people want to discuss the new kernel silly

Re:Installation?

[qtp]

I'm not a RedHat user, but I did find this document [redhat.com] for building a custom kernel on RH9.
It's not specific to Ferdora, but it should be enough to get you through it. It looks pretty generic and is very similar to what I did before I began using Debian's kernel-package [sourceforge.net] system to manage my custom kernels (which BTW, is one of the best things about using Debian, especially if you are using more than one box).

You may want to familiarize your self with the

lspci

command, and possible with the

/proc/pci

file (use

cat /proc/pci

in an xterm to read it) in order to be sure that you're not leaving out support for some of your hardware. Read the help file for anything you don't know about, and don't hesitate to read the device specific files for items that you might need (if they have them).

Be sure to use the Fedora users list [redhat.com] for questions that you can't figure out from the docs. I'm sure there's lots of helpful folk there. If it's anything like debian-user [debian.org] you'll have no problems getting prompt and helpful answers. (If anyone gives you an RTFM, follow the provided link if there is one, if they did not provide one, then list the docs you've already consulted and ask if there's something you've missed and where it might be). Using the lists is not scary and, despite the *ss-hats who think otherwise, RTFM is not an insult. You'll learn much faster if you read the material yourself and ask questions afterward.

Be sure to set up your

/etc/lilo.conf

to boot from more to one kernel (if you're using grub you'll need help from sonmeone else for this). Move your old kernel into the new place (usually, I use

/boot/vmlinuz.old

for this) so you'll have a working alternative if you screw things up.

Be patient, take your time, check everything twice. It take's quite a bit of time to do this the first few times, but once you know your way aroiund the kernel config you realize that it's realy not all that difficult.

Firewire problems.

[ijuma82]

I had some issues with firewire in kernel 2.6 (getting total lockups when under heavy activity). I solved them by using the latest code in the subversion tree of http://www.linux1394.org/ with kernel 2.6.5-rc2. Since then I have upgraded to 2.6.5-rc3 and now 2.6.5 without any problems.

So, my advice is, try the latest kernel with the latest subversion tree from the linux1394.org website. Hopefully, that will solve your problems. 8)

Re:Stable?

[Ignominious Cow Herd]

Not Informative.

The OSS drivers are marked deprecated.
The OSS compatibility modules are not deprecated. They are what allows OSS-based apps to run in an ALSA environment.

He is not having problems with the OSS drivers.

Re:Scheduler?

[Anonymous Coward]

Try this.

hdparm -d1 /dev/hda

Re:Linux Changelog Email Publishing

[SCHecklerX]

Your only friends and family must not use windows/outlook/msie then. I get tons of spam, mainly because of people I know getting infected with windows viruses (and I'm in their address books).

Re:the PPC64 work is looking real nice

[zdzichu]

PPC64 will generally get more attention now, since Linus' main workstation is PPC64 since about two months now.

Mini ITX & i2c

[Cytlid]

For those of you (like me) who might have an EPIA-V mini-itx board and want to use lm_senors, the chipset (vt8231) doesn't look like it's been ported to 2.6 yet. :( grep i2c ChangeLog came up with quite a bit, but no port for that chip yet...

Re:version.h

[LordHunter317]

IF you're on Debian and tried compiling your kernel with kernel-package_8.084, this is just one of hte problems I saw. For whatever reason, that version of kernel-package is extermely broken and not working correctly.

Bit Torrent

[peripatetic_bum]

Why dont they set up a bittorent link, I have no idea anyway here it is http://mung.net/~dude/linux-2.6.5.torrent [mung.net]

Laptop Mode

[cerberusss]

It would be great if they would include this laptop mode [xs4all.nl] patch, like they did in 2.4. It really prolongs battery life on my laptop, not to mention that with quick spindown times (using hdparm) it kinda solves the heat problem [dell.com] on my Dell D600 laptop.

Re:VM/swapd

[Anonymous Coward]

Most of them are more likely to be "adjustement" patches. There was a lot of changes and they were debugged properly in the development series, but only real world (publishing 2.6 so everyone starts using it) allows you to see what you've really done. So, most of those are just adjustements. For example, some of those patches are just tweaks wich allow low-memory machines to behave as 2.4 under some loads. Just tweaks. And then, there're those "blk_congestion_wait races" that you will only hit with 32 CPUs and a specific benchmark stressing the docens of disks, then you'll hit the bug. But that is not likely to happen in a P4 with 256 of ram. And that's one of the reasons why 2.6 is _so_ great and it's so stable: It has been tested by IBM, OSDL, SGI etc people, with docens of CPUs and disks that can find bugs that would take years to reproduce in a UP o dual machine, or they would just not happen. Which means that those machines have already found 99,99 % of the critical bugs that could affect a UP computer.

Re:Stable?

[slamb]

I love the speed increases that the 2.6 kernel has achieved on the desktop (and for things like media: mplayer never bugs out with that charming "YOUR COMPUTER IS TOO SLOW" message anymore). However, I don't know if it can be considered even remotely stable. Since switching, my uptime has been a Windows like joke.

[...]

- Firewire and sbp2 support is completely broken. Ironically this has, I believe, more from "experimental" in 2.4 to a normal feature, yet it worked fine before and now doesn't work at all (the linux1394 forums forums reflect that I am not alone in this). Trying to copy data to sbp2 drives segfaults, hangs, and worse. Beware of connecting to 2.6 if you have a firewire drive with data you hold dear...

It's important to keep some perspective. Usually whenever anyone says something is full of bugs, they mean that they keep running into the same bug over and over. If you're having problems with Firewire, very likely you're running into one bug in your driver repeatedly. The other people complaining may have the same chipset and the same problem.

My point is that you can't make any generalizations to the entire kernel series (or even subsystem, like 1394) being more or less stable just because you encounter a single bug that you didn't used to. Look more closely at the oopses and your system logs, see where it's happening, file a good bug report. They'll probably have it fixed in a couple releases.

People use "stable" or "unstable" to mean a lot of different things:

1. If they're changing the APIs constantly or not.
2. If the core of the system doesn't crash and performs well under a variety of loads
3. If their system doesn't crash and performs well under their load

...and #3 really needs to be qualified with "for me" or "with this exact hardware, doing this". Because otherwise, you're saying the whole series sucks because of a single bug. And very likely, a bug in a driver. When I read kernel traffic [kerneltraffic.org], lwn [lwn.net], or kernel trap [kerneltrap.org], I frequently see mention of fixing some unsafe coding practice...in the core kernel. Drivers are left for their maintainers to update. Some do so quickly and well. Some don't.

Ideally, a system would be so rock-solid that you would never run into even one stability or performance bug. But I don't think that's much more realistic for Linux 2.4 than it is for 2.6.

(This message is not just aimed at you. I see this a lot.)

Re:Again

[diegocgteleline.es]

So, do it yourself. You can't expect that the kernel developers are going to hide the guts of the development to other developers. As said do it yourself, or read www.lwn.net:

The 2.6.5 kernel is out
[Kernel] Posted Apr 4, 2004 15:01 UTC (Sun) by corbet

Linus has announced the release of the 2.6.5 kernel. Changes since -rc3 include another ALSA update, some architecture updates, and various fixes. Big changes since 2.6.4 include the netpoll interface, a change to the DMA API (covered a few weeks ago), some SELinux improvements, the hotplug CPU patch set, and more; see the long-format changelog for the details.

Simple enought?

Debian with 2.6.4 and it works like a charm.

[Bob_Robertson]

Don't delay, there's no reason to. There are enough differences between 2.4.x and 2.6.x that if you will want to upgrade, do it sooner rather than later. It took me a while to get everything hammered out.

TLS libraries, for instance, as well as pcmcia and alsa changes. 2.6 is also doing substantially more hardware discovery, which caused me a bit of trouble on my annoying Vaio laptop until I figured out what kernel modules to remove from the tree so they couldn't be installed.

As you already know, installing a later tertiary kernel version is very easy under Debian.

Bob-

Wrong device driver for X? Using /dev/input/mice?

[Anonymous Coward]

Hi.

I'm working on an X driver for a different tablet, and it's based partially on the driver for the wacom tablet.

Internally, XFree86 seems to know about 2 kinds of pointers: Absolute devices, and Relative devices. Absolute devices include CAD Digitizers and tablets. Relative devices include mice, trackballs, and joysticks.

Linux, on the other hand copies all events from pointer-like devices (but not joysticks, afaik) to the relative device interface of /dev/input mice, in addition to giving each device a /dev/input/eventN and /dev/input/mouseN interface.

To use an absolute device under X you must:

1. Use an XFree86 device driver that knows about Absolute events (wacom, in your case), connected to the correct device file (one of /dev/input/eventN).
2. Make sure any other X pointers are using specific mouse devices (/dev/input/mouseN), not /dev/input/mice. If you do not do this, every tablet event will happen twice, and the cursor will jump around.

You probably also want to make sure that the tablet is not sending core events (no options CorePointer, SendCoreEvents, or AwaysCore on that inputdevice in /etc/X11/XF86Config). That way in applications like the GIMP you can use the tablet for drawing while using the mouse for menu selections etc.

Any questions? [e-mail]

Re:Laptop Mode

[Kourino]

It looks like the laptop mode patch is in Andrew Morton's -mm patch series. Since -mm has the leading edge of six different maintainer trees for merging with Linus, and contains lots of other general fixes, running -mm may not be a bad idea, since fixes often appear in -mm before they do in mainline. (Of course, sometimes things break in -mm that aren't broken yet in mainline too ... )

Nothing to do with "new make parameters"

[Kourino]

Hrm, looks like you meant to post that with "Plain Old Text" formatting ... :/

Anyway, make install can depend on the contents of /sbin/installkernel, which is distribution-specific. Everything's compiling, but installkernel is trying to make an initrd that contains the ata_piix module, which you didn't build as a module (either you built it into the kernel or aren't using it at all). So, either fix your installkernel settings (check docs on how to do that), or build ata_piix as a module. (Basically, this is a configuration error -_^ )

This has nothing to do with "new make parameters".