Submitted by Linus Torvalds has released version 2.6.25 of the Linux kernel. “It’s been long promised, but there it is now. Special thanks to Ingo who found and fixed a nasty-looking regression that turned out to not be a regression at all, but an old bug that just had not been triggering as reliably before.”As usual, KernelNewbies provides a readable changelog that we can all understand. They sum up the most important changes in 2.6.25 as follows:
“Summary: 2.6.25 includes support of a new architecture (MN10300/AM33) and the widely used Orion SoCs, a new interface for more accurate measurement of process memory usage, a ‘memory resource controller’ for controlling the memory usage of groups of processes, realtime group scheduling, a tool for measuring high latencies called latencytop, ACPI thermal regulation, timer event notifications through file descriptors, an alternative MAC security framework called SMACK, a ext4 update, BRK and PIE-executable address space randomization, RCU preemption support, FIFO spinlocks in x86, EFI support in x86-64, a new network protocol called CAN, initial ATI r500 DRI/DRM support, improved device support and many other small improvements.”
Head on over there for a much more detailed list of all the new features. If you’re really die-hard, you can check the original, unformatted changelog too. As always, this release will find its way to your distribution of choice, but if you cannot wait, you can download the sources from Kernel.org and build it yourself.
Congrats for the release, Linux the best ever as always.
Changelog that more people can understand
http://kernelnewbies.org/Linux_2_6_25
Err, I linked to that. Check the “read more”. OSNews is moving to more elaborate newsitems (check the previous 3 items too to see what I mean).
I was actually thinking about that, when reading through the latest news items. They were a lot more detailed than “usual”. I like it!
Then perhaps you should make the “full” story as found in “read more” appear instead of the summary when viewing comments. This would mimic Slashdot style, as well as the style of livejournal and other blogs (wherein “below the fold”, as it were, also includes listing comments) and I imagine would be less surprising behavior.
IMO it’s a bit confusing that when I want to see all the comments I can’t see the “read more” section
My Rant. I am not an expert but if my sound card is not working, I have to I have to compile and insert module in kernel or patch it. Not good for joe average user. Once kernel is installed there should be no need to do anything with it. But even on user friendly distros like *buntu Joe user encounter kernel tweaking. When was last time I heard about compiling windows kernel to set up my soundcard or digital camera??
Mr Linus is doing great job for linux community. But at some time he should realize there is lot of modern hardware exists besides his keyboard and compiler.
Mr Linus should share some responsibility for failure of linux systems compatibility on modern hardwares and come out with vision to integrate s/w and h/w.
Erhm, no.
If your soundcard is not working, then in ~99% of the cases, its because the user simply has not unmuted it and turned volume up.
in the last 1% of the cases, problems can be various things, but to assume that even 50% of those things are kernel related would be wrong, given that lots of configuration happens with alsa in userspace..
I’d go so far as to say that only 0.25% of all sound related issues are due to kernel space..
i’d also go as far as to say, that of all the PC’s in the world today, which are not broken(not counting OS here), linux will work with _ALOT_ more of them, and directly out of the box, than is even possible to get working with XP. And if we count in vista aswell, well, thats not really much of a fight, now is it?
(yes, i have personally seen MANY older computers which has integrated audio that does not work for 2k/xp, only 98)
and then ofcourse also comes all the pieces of embedded devices out there.. ooh, i think winblows, in any version, is pretty much beaten.
Why do people insist on doing juvenile spellings of companies they don’t like? It makes your arguments that much less compelling…
As a grown up, please explain how a misspelt name makes an argument less compelling. We’re talking logic and facts here, right? So how exactly a name that you can still recognize as referring to a specific entity changes facts and logic?
(Yeah, disregarding minor form issues and focusing on facts and logic takes some growing up, that’s for sure.)
Because people take you a lot less seriously when you do do that? Same as people here giving really good rebuttals to someone but ruining it by ending it with “idiot” or the like.
gah, moron. ;P
Because people take you a lot less seriously
Exactly. A person, not an argument. “Winblows” is lame and the one who uses it may be childish, but you don’t refute an argument by pointing it out.
fine, you win
I suck at arguing. *sigh*
Yes.. describing something by what it actually is, makes me SO much more juvenile and unserious, than say.. being the CEO of the worlds largest company, and throwing chairs yelling: “IM GOING TO f–kING KILL GOOGLE!!”
yes, i totally see your logic.. or perhaps it makes me more serious than a mega corporation failing to properly implement transparent png for more than a decade, while my simple mobile phone can display transparent png’s as icon for mobile java applications.. Yes.
ooo… nice rebuttal.
All I have to say is COUNTER RIPOSTE!
While I did not find the lone reference to “winblows” to be all that notable and would not have pointed it out myself… since the topic has come up, I’ll briefly weigh in. Terms like “Winblows”, “Microsoft”, and “Microsoft” (the OSNews Censor kicked in on the last two terms. Please feel free to use your imagination.) do diminish one’s credibility a little, or a lot, depending upon the point of view of the listener. If the objective is to be as persuasive as possible, it makes sense to avoid such terms, as there is really no up-side to using them. Few people are going to find a post more persuasive due to its use of derogatory slang terms, but many people will find it less so.
At any rate, that’s my pragmatic, nonjudgemental slant on the issue.
Edit: I find the OSNews policy of censoring nonprofane terms from users’ posts to be far more offensive and annoying than the use of the actual use of the terms by users. These terms do add a certain connotative element to the post, and censoring them removes useful information.
Edited 2008-04-19 15:35 UTC
yes well, i guess we know osnews’s official policy on censoring and freedom of speech etc… I really had not imagined them being chinese, but well..
I shouldn’t bite, but I will.
1. If you’re using a standard distro like ubuntu or fedora/redhat or such, then the likelihood is that your package manager will be able to find a pre-compiled kernel module for your hardware. So in 99.9% of cases, no compilation will be needed. In the Windows world, drivers are usually kernel modules too. [This is changing slowly, I know]
2. In the Windows world, it is the hardware manufacturer’s responsibility to ensure that they distribute drivers. In the Linux world, it is usually left up to the community to maintain a working set of drivers. This will not change until Linux gains a significant market share [well, actually, it is changing as we speak]. So there are always going to be little problems, and issues.
3.
IIRC, Linux is the Operating System that supports the most Hardware. It beats Windows hands down, which considering that a lot of Linux driver development is done without full manufacturer support, is pretty remarkable.
4. So you reach a situation where your OS installation does not support X hardware.
If you were with Windows, or OSX, then this would be then end of the line for you. There is no realistic likelihood of being able to write your own device driver for it. The Windows DDK costs a LOT of money. And you’re not likely to find community drivers.
With Linux however, you can search for kernel modules, recompile the kernel with extra support, write your own driver, call for a community effort to get your HW working.
The fact is that with Linux, you have far MORE options than with a closed OS. People shouldn’t grumble when they try to do something that isn’t possible on any other system and find themselved outside their comfort zone.
Dear god, I get marked down to -4 after ranting over someone whining over the cost of a Mac, and this Nobel prize winner not only maintains his points but some numskull adds some to it!
Personally, the above sounds like nothing more than a variation of a FreeBSD rant or the usual rant put out on Slashdot of ‘Linux is dying”.
Oh, as for sound, I’ve yet to find a sound card not supported, and as for the digital camera, that has nothing to do with the kernel as it is sorted out by a user space library like libptp – assuming you has enough brain cells to ensure that the digital camera you purchased conforms to either PTP specifications or can mount as a mass storage device.
Will be this kernel come in the next version of Ubuntu?
Regards
No, Hardy has kernel v2.6.24.
My another rant. Very popular distros like Knoppix, Ubuntu and PCLOS have Highly modified and patched up kernels. Try to replace ubuntu kernel with plain simple kernel from linus. It will Not breakdown system but suddenly most of your peripherals will stop working.
Thats why is one peripheral like printer works perfectly in Ubuntu will not work on PCLOS on exactly same machine.
Mr Linus will not take responsibility if my printer will not work with his latest kernel and HP will not take reponsibility for if printer is not working by changing distro, keeping same kernel version. No one will take responsibility but developers will blame joe end user..
At least you can blame Bill Gates if something is not working in windows. Thats why 90% people use MS.
Here is something more basic, when you purchase your hardware, check whether it is Linux/*BSD or Solaris compatible; so if you do, at a later date, wan to dabble with one of those alternative operating systems, you know your system is compatible with the operating system.
The kernel, linux, is part of an OS, ubuntu.
So if you are going to use an OS, ubuntu, use the packages it provides.
So it has no sense saying “Try to replace ubuntu kernel with plain simple kernel from linus.”.
Maybe you could help me, but I don’t see highly modified “patches”.
http://changelogs.ubuntu.com/changelogs/pool/main/l/linux-meta/linu…
Clone the git repo from
http://kernel.ubuntu.com/git
Do a git diff between this one and kernel.org upstream kernel.
No it isn’t but thanks for playing.
Your twisted, irrational logic is awesome.
No. If the device works correctly under ubuntu’s kernel, it will work correctly under a custom one. When it doesn’t, that means the custom kernel was configured incorrectly. There’s no mythical make_device_work() call in ubuntu, it’s all the same stuff made to appear different through configurations and patches.
[/q]
I bet he would if he had anything to do with that printer, but HP wrote the module and ubuntu configured your kernel. To assume responsibility of that would make less than any sense.
1st: You can blame whoever you want. I blame Luke Wilson for having read this. He has about as much to do with this post as Bill Gates has to a broken HP Windows driver.
2nd: I’m sorry it sounds that it sounds mean, but when a user configures something wrong, it is their problem, not the developers. That’s just where the liability falls. Now if the problem is a bug, while developers like to be informed of this, liability still falls under the user in the OSS world; that’s the trade-off for gaining the level of control of a program that the GPL/BSD/MIT/… license grants.
3rd: I was under the impression that the 90% population you reference are using Windows because it happened to be on the computer when they bought it, not because they evaluated the distribution structure of two separate platforms.
All those support issues have an effect. Kernel alone will not make Desktop Linux happen – does not matter how stable or how high the uptime is…
This is what RedHat said in regards to the Global Desktop they were going release my are no longer going to.
——–
“We are focused on infrastructure software for the enterprise market, and to that market we are offering the Red Hat Enterprise Linux Desktop,” said Michael Chen, vice president of corporate marketing at Red Hat.
Red Hat said that the Red Hat Global Desktop (RHGD), originally announced last year, was delayed because of business issues, although Red Hat had hoped to deliver RGHD in a few months, it said in the post. The RGHD is designed exclusively for small, reseller supplied, deployments in emerging markets, like Brazil, Russia, India, and China, and will be supplied by a number of Intel channel partners, the blog entry said.
In a reference to Microsoft, Red Hat said that the desktop market suffers from having one dominant vendor, and some people still perceive that today’s Linux desktops simply don’t provide a practical alternative. However, a growing number of technically savvy users and companies have discovered that today’s Linux desktop is indeed a practical alternative, it added.
“Building a sustainable business around the Linux desktop is tough, and history is littered with example efforts that have either failed outright, are stalled or are run as charities,” according to the post. But there’s good news too. Technical developments that have become available over the past year or two are accelerating the spread of the Linux desktop, it added.
—–
Am I the only one excited about Smack? its like SELinux without all the shit you don’t need and have no chance in hell of configuring correctly :p
Last time I tried MAC on linux it was SELinux and it was pain. I think it perhaps is time to try it out again.
Naaah….. Smack is baad…. mmmkay?!
Am I the only one excited about Smack? its like SELinux without all the shit you don’t need and have no chance in hell of configuring correctly :p
Last time I tried MAC on linux it was SELinux and it was pain. I think it perhaps is time to try it out again.
Try grsecurity (http://www.grsecurity.net/). It’s much easier to get going and some people think it is better than SELinux.
Edited 2008-04-17 22:39 UTC
This is all great and all … but how many new bugs did they introduce?
In the words of the great* donald rumsfeld, that is not knowable.
*not serious in case the total unseriousness does not cross the tubes
Edited 2008-04-17 19:27 UTC
Very nice. Linux 2.6.25 compiles with gcc 4.3.0. Linux 2.6.24 didn’t.
A C kernel compiles in a C compiler? Somehow I fail to be impressed by that.
I have nothing against GCC extensions when they make sense, but at least they should stick to whatever they come up with. They should not be breaking source compatibility with each new version.
One feature which has quietly made an appearance in 2.6.25, and which should really be at the top of the list of things to be ecstatic about is TASK_KILLABLE. No longer will admins have to deal with unkillable processes, stuck in a ‘D’ state, because the NFS server they were doing IO from became inaccessible, or the tape drive went south in the middle of an operation, etc. I’ve long held a deep resentment toward those annoying and embarrassing little reboot-forcers that refuse to die on command, and have plagued me for my entire 20 year Unix/Linux career. If 2.6.25 brought us nothing else, I would still be overjoyed about its release.