Linux 2.6.32 has been released. New features include virtualization memory de-duplicacion, a rewrite of the writeback code which provides noticeable performance speedups, many important Btrfs improvements and speedups, ATI R600/R700 3D and KMS support and other graphic improvements, a CFQ low latency mode, tracing improvements including a “perf timechart” tool that tries to be a better bootchart, soft limits in the memory controller, support for the S+Core architecture, support for Intel Moorestown and many other improvements and new drivers. You can also read the full changelog.
this is great:
make localmodconfig
and a least 3d support for my ati now i can use moovida or xbmc ( i not decide yet )
An in-depth discussion of the new features in Linux kernel 2.6.32 has been posted on the H Open Source news site.
http://www.h-online.com/open/features/What-s-new-in-Linux-2-6-32-87…
FTA:
So for now, of the distributions recently released, only Fedora 12 might allow people to try this out as yet.
I’m waiting for Arch Linux (rolling release) to indicate that they have the new kernel and appropriate versions of Libdrm and Mesa 3D before I try this myself. A distribution like Arch will probably be the quickest way to get it running for people who aren’t confident to compile it all for themselves. Other than that, I would suggest that this set of capabilities won’t be available to the wider Linux-using public until Ubuntu/Kubuntu Lucid Lynx (LTS) is out in April next year.
I suspect not. Recently, Fedora luminaries, to my great surprise, have gone *way* out of their ways to convince me that Fedora isn’t really very cutting edge. Fedora releases every six months, like most other distros. Spring, Fall… Spring, Fall. And so it includes pretty much the same old stuff as the others. I was wrong to call it cutting edge, I guess. Stop perpetuating this myth!
Edited 2009-12-04 00:00 UTC
http://www.h-online.com/open/features/What-s-new-in-Fedora-12-86383…
What can I say?
It is not enabled by default, but apparently it is available for testing in Fedora 12.
No. “They” were refuting your assertion of it being “bleeding edge”. Please don’t change the words you used to mean something entirely different. That is disingenuous.
Gentoo has had .32 practically since the zero-minute…
It’s lacking in mesa and libdrm however, but that is easily solved by installing the x11 overlay and picking the git ebuilds from there, trivial as well.
I’m really looking forward to getting to test KMS on my ATI Radeon 3850’s. It’ll be awesome if it works. This is good news!
I unfortunately won’t benefit much from Linux kernel updates anymore as I removed Linux from my desktop and laptop, only having Linux on my server. I’ve had lots of different small issues which have nagged my nerves, but the one which finally broke the camel’s back was upgrading from 2.9.29 to 2.9.31 on my laptop and losing all power-management features (CPU frequency scaling is gone on my laptop, couldn’t find any way of re-enabling it, and suspend/sleep/hibernate all crash the system)…
It is a great OS and all, but losing such important features in an update just is a serious no-go.
Out of curiosity, was this an upgrade of vanilla kernels? Or of distro kernels? And if a distro kernel, which one?
The original link is to lkml.org
http://lkml.org/lkml/2009/12/3/11
This site is a reflector of the Linux Kernel Mailing List. Hence LKML.
The Linux Kernel Mailing List is where the Linux kernel development is done. This is every Linux kernel. The font of all linux kernels, if you will.
As I understand it, some time before release distributions will declare a “kernel freeze” and take a “drop” of the LKML kernel and compile it with their preferences for some given supported architecture (or architectures), and refine, polish and debug it from there.
For example, the Lucid Lynx release schedule is here:
https://wiki.ubuntu.com/LucidReleaseSchedule
Kernel freeze is on March 11, 2010. On that date the kernel version for Lucid Lynx will be decided.
PS: apologies, sbergman27, but I didn’t realise at first that your post was not addressed at the original article, but rather at WereCatf’s post. That was silly of me, I should have clicked. D’oh.
However, the information above might still be of interest to some.
Edited 2009-12-04 02:29 UTC
And your post has exactly what to do with my question to WereCatf?
kernel.org is the source of the original alpha quality Linux kernels, which have not gone through anyone’s proper QA channels. They pass the kernel dev’s “Works for Me” QA criteria.
I’m sure you will disagree violently. But a totally disorganized mailing list + a totally ignored kernel bugzilla doesn’t quality as a proper QA channel.
Profuse apologies. I only realised later that your post was a reply to WereCaft, and not a comment on the original article.
Silly me.
Distributions will normally only change the kernel version when they update the entire distribution.
Did you take it upon yourself to update the kernel, or did you update your whole distribution?
Either way, why not simply revert to the earlier kernel?
2.9.29? 2.9.31? Where did you get these from? The kernel just released in this announcement by lkml.org is only 2.6.32.
Edited 2009-12-04 02:24 UTC
Except Fedora. Fedora switches kernels on you willy nilly.
I think we can safely consider that to be a typo, lemur. Don’t nitpick. WereCatf is at least as well versed in these things as we are.
Does it now? That I didn’t know.
A “rolling release” dsitribution such as Arch will also do this, but you are warned about it all and you use Arch knowing that the kernel could go south on you.
However, if the kernel does go south on you using Arch, then you boot with the previous kernel and revert the kernel version. Surely if Fedora switches kernel versions, there is a similar mechanism provided to go back a version?
I haven’t come across any distro myself that upgrades a kernel version without the ability to revert it. Maybe Fedora is an exception, I don’t know.
But nevertheless if someone is using a cutting edge distro and accepts a kernel version update, surely they should know the risk that that entails? It is quite possible to have a regression. The options are these (for those who like to play with such fire):
(1) revert the kernel & carry on,
(2) revert the entire distribution & carry on, or
(3) change distro to something a bit more stable.
Ditching Linux entirely doesn’t seem a very sensible choice to make for someone who apparently liked to play with cutting edge kernels.
Hence the questions (and the disbelief they imply) about the versions.
Edited 2009-12-04 02:43 UTC
Yes, it does. Over the course of a Fedora release, they push out major new kernel versions at random times.
Well, except that I just came through a bout of Rahul insisting that Fedora was as conservative as any distro out there, and not bleeding edge at all. I guess he didn’t feel the need to mention that the kernel could go south on you at any time.
And I’ve had that happen.
But why should WereCatf not use something else if Linux is letting her down? You think she owes Linux something? Because she doesn’t, you know.
I’m pretty sure, at this point, that Linux’s overenthusiastic fans will turn out to be its eventual downfall. Which saddens me. Because I am a fan.
Edited 2009-12-04 02:55 UTC
“Well, except that I just came through a bout of Rahul insisting that Fedora was as conservative as any distro out there, and not bleeding edge at all”
Nope. You are just lying again. Your claim was simply that Fedora was the only distribution officially advertising itself as bleeding edge and I asked you for a reference which you failed to provide because your claim was false.
Now to come to subject matter, Fedora is hardly the only distribution to provide kernel revisions as updates. Many other distributions do that as well with some choosing to use separate streams for that. Like anything else, it is a trade off with its own advantages (better hardware support for instance) vs disadvantages (potential regressions for instance). The good thing with kernel updates are that you can always boot back to the older kernel. With Fedora 13, this kind of roll back can hopefully extended to the rest of the updates as well
https://fedoraproject.org/wiki/Features/SystemRollbackWithBtrfs
This is the reason why Red Hat kernel maintainers working (three of them full time) on Fedora pick and choose what to push out as updates. It is hardly random.
Uhuh. I made a statement, and you saw that your could turn the discussion into this blind alley where we could argue whether the Fedora project officially uses the word “bleeding” in their marketing materials. And you jumped on it. I’m not falling for it. I won’t accuse you of lying on that point. But there is certainly an element of deception in that.
Is Fedora popularly known as being “bleeding edge”? Yes, it is. Deny it, Rahul.
Fedora jumps you from 2.6.23 to 2.6.24 to 2.6.25 to 2.6.26, etc. all in the same, supposedly stable, release. Not many distros do that. They stick with one kernel version. Of the major distros, in fact, Fedora is pretty much the only distro with that cavalier attitude. So you and lemur can just cease and desist with the “everyone does that” mantra. Because it is not true. Although I should make clear that I realize that Fedora has its reasons for doing so.
Fedora is somewhat reckless. And that’s perfectly OK, as long as you admit it and users go in well informed. Fedora can be a lot of fun if you don’t have a lot of responsibility riding on it. I’d still be using it at home if I could afford to be so happy go lucky.
What I so disapprove of in you lately, Rahul, is that you are not willing to be honest about what Fedora is. Because there is nothing particularly wrong with what Fedora is. Fedora is good to the extent that its advocates are honest. Fedora does a good job of getting new stuff ready for the greater Linux distro community at large.
And my number one complaint about Fedora is not that it is bleeding edge, cutting edge, leading edge, whatever your marketing sensibilities are comfortable with, Rahul… but that it’s reps seem to want to deny that it is what it is when it is convenient for them to do so, and then show up to accept any credit when credit is being handed out.
Fedora is, in general, less reliable than other distros. Which is fine, considering what Fedora’s goals are. (And being Red Hat’s perpetual beta is one goal that Fedora can never get completely away from.) You can’t have cutting edge features and cutting edge reliability both at the same time.
So all I ask is that Fedora reps be honest about what Fedora is.
And in the interim, I will continue to be honest about what Fedora is.
Edited 2009-12-04 03:55 UTC
No. You have become a Ubuntu fanboy trolling about Fedora and continuously lying about it. You are dishonest person and I will not engage you in pointless discussions. Just don’t make false claims about what I said and I don’t have to bother calling you out on it.
No. The Ubuntu guys have done a very good job with Ubuntu. Ubuntu has been a success for Linux in its particular playing position.
The Fedora guys have done a very good job with Fedora. Fedora has been a success for Linux in its particular playing position.
Where we seem to disagree is that Ubuntu and Fedora have different playing positions. And would you really want it any other way?
No. We have no disagreement on that. What we have disagreement on is about you frequently make factually inaccurate claims and never apologizing when others point it out to you. That type of attitude doesn’t help you. It has nothing to do with distributions.
The statement claimed it was bleeding edge that was the aim of the releases made by the project (it is not) when your statement was shown to be BS you changed the arguement into what the “perception” of it is. You can play word games if you like but no one is buying it.
……. skipping masses of opinion……..
Twist away, if you change your language enough times maybe one version will be seen as accurate.
That must be why more ubuntu updates have caused the system to become unbootable for all users – in less releases and less kernel updates at that. Damn logic and factual data, that’s fine go on, keep ignoring reality.
I would reccomend not changing your language every time your arguement is shown to be wrong next time as part of this “honesty”.
Indeed she doesn’t owe Linux anything, and she of course can and should use whatever she pleases.
My beef would be that she offered a comment on a public forum that a kernel regression was her reason for abandonning her use of Linux, when such a regression (while possible) is very easy to revert. It shouldn’t be a problem at all for anyone used to running bleeding edge distributions which update the kernel from time to time. For anyone not used to reverting kernels, they shouldn’t be running bleeding edge distributions.
So I take it back … I believe WereCatf does owe Linux just one tiny little thing. Honesty. That is all. Nothing else. Just don’t publically rag it in an area where it shouldn’t be (and indeed isn’t) a problem.
Heaven knows there are more than enough people more than willing to bag it mercilessly for things that it really does have small problems with.
Desktop Linux failed her enough that she gave up on it. And you are still trying to put the blame upon her. As if that will do any good. It’s really no wonder that we’re getting nowhere fast with the Linux desktop. I actually can begin to understand why some here use pejorative terms like “freetard”.
Our beloved Linux f–ked up for WereCatf. Apologize. Don’t attack, Lemur.
Edited 2009-12-04 04:15 UTC
P**s O** with your holier than thou lecturing. Who died and put you in charge?
I have a valid point, and I stick by it.
You have an all-consuming obsession, and you stick by it. There’s a difference.
Edited 2009-12-04 04:36 UTC
http://www.google.com.au/search?hl=en&source=hp&q=define%3A+hub…
http://www.google.com.au/search?hl=en&source=hp&q=define%3A+hol…
There isn’t much difference between either of those and you.
Technically its was Mandriva that failed her, there is no way to know from the description whether the regression was in the kernel itself or in Mandrivia’s setup. The few kernel-based regressions I’ve seen were always cleared up pretty quickly, so I’d just revert to the old version and use that till the new version was fixed.
Otherwise, I agree, there’s no point in blaming the user here. Use whatever you’re most comfortable with.
Please stop whining, Linux kernel has its massive problems once in a while even in so-called stable distros. Sometimes e.g. a driver stops working and this problems last for several minor kernel updates – we call this regression and this regression is a well known factor among Linux devs.
The latest kernel _is_ stable, at least according the opionion of the majority of Linux devs. Sometimes I had to build the latest kernel just to fight some bugs introduced in the “stable” kernel of a Debian release.
Distributions will normally only change the kernel version when they update the entire distribution.
Did you take it upon yourself to update the kernel, or did you update your whole distribution?
Mandriva does update kernels every now and then, haven’t paid attention to kernel version numbers though, but yes, I upgraded the whole system from 2009.1 to 2010.0. I did try with the earlier kernel and power-management worked again, it’s just the newer one which broke it. And no, I didn’t bother trying to compile a kernel of my own from the sources, I got better things to do.
Oh, and now that I remember to mention it, I did try both vanilla and Mandriva-patched ones.
2.9.29? 2.9.31? Where did you get these from? The kernel just released in this announcement by lkml.org is only 2.6.32.
Yeah, sorry, typing when I’m tired isn’t always such a smart thing to do.
Edited 2009-12-04 03:59 UTC
How about downgrading until the problem is solved?
How about downgrading until the problem is solved?
Yes, I could have done that, but I decided that I don’t like simple updates suddenly ruining something so important as power-management on a laptop and moved away. There were other issues too, as I mentioned, but you could say it was the final nail in the coffin.
But what you have admitted is that your purported “final nail in the coffin” wasn’t actually a problem at all.
You are of course perfectly entitled to use whatever you want, but that is not the point. I utterly fail to see the point in your ragging Linux in a public forum over something that is easily worked around.
If you want stable … run a stable distribution. Or not … run whatever you want. But please … don’t try to pretend that there is a problem when there isn’t one.
Why should he ‘work around it’?
I believe it’s she. But look it the other way around. You install Windows, it gets infested or gets slow or whatever, what do you do? You run away screaming that you shouldn’t need to do anything? Find another OS that is perfect and guarantees no problems at all? Buy a Mac and pray it doesn’t have any problem ever?
Safe rollback/downgrade? Too bad ZFS is not ported to Linux yet. That would have solved your problem.
Each time I upgrade OpenSolaris, a new ZFS snapshot is automatically taken. They all show up in GRUB, and I can choose which snapshot I want to boot into. If the new upgrade is buggy, I just reboot into the previous one and delete the new snapshot. This takes on minute or so, and my system is exactly like it was before the upgrade. No bits are changed. This is complete and safe rollback in one minute. This have save countless hours, and I can have several branches in GRUB. One unstable where I try out different packages, and one stable branch. Or several branches.
(ZFS snapshots just starts to write everything to a new place on the drive, leaving old bits intact. Hence, each ZFS snapshots doesnt use much space, maybe a few KB)
ZFS has been ported to Linux (FUSE) but it isn’t going to be useful on Linux because ZFS is licensed under CDDL and that is incompatible with GPL.
Fortunately, Btrfs works fine as a replacement.
That’s not completely fair. I don’t really think that you can reasonably say that a kernel regression is the user’s fault.
You can’t really blame someone for not sticking with Linux if it doesn’t work for them — or if making it work for them consistently requires more effort than they are willing to invest.
And I don’t think she’s really bad mouthing Linux, or ragging on it. She’s just saying it doesn’t work for her, and why. Which is a reasonable thing to do, in context.
This is so brilliant.
You are so proud that Linux can be installed to your parents’ computer, but once something goes wrong, you have to be all pro.
You are so eager to advocate Linux on public forums, but once someone is even remotely critical, it is suddenly “ragging Linux in public forum”.
Edited 2009-12-04 18:46 UTC
Dude, if you are going to completely misrepresent the words of other poster, don’t cut and copy those words in your message. Just a friendly advice.
I totally get this. Same thing happened to me.
I didn’t upgrade the kernel for kicks and grins: the one I was using didn’t work with mixed USB 1.1 and 2.0 devices on a hub. New one did. But the new one broke my hardware, driverless RAID card.
And obviously, I could have reverted and worked around the hub issue until everything works (it all does, hub and RAID in the last kernel I tried) but like you it was just one more thing on top of many. For you it was the last straw, a term with which lemur2 doesn’t seem to be familiar.
For me, the loss of KDE3.5 was that last straw, though the kernel fun was one of the straws before that. If lemur2 wishes, he can read a rudimentary list of my complaints here: http://www.osnews.com/thread?363587
After using Linux and only Linux for several years, it slowly became obvious that it wasn’t for everyone. That was fine, because it was for me. I could change stuff I didn’t like, an opportunity of which I actually availed myself. But eventually the time required to keep things running with changing hardware, and having my DE yanked out from under me drove me off and I don’t miss it.
Hardware and programs work now. And I am free from the horrible problems so many people have on the evil platform by virtue of not being a computer retard, so I’m just not seeing a downside. Yeah some vigilance is required, but “not acting like an idiot” is far less time consuming than “fix random thing that broke” over and over.
My favorite random thing that broke was lynx not working anymore one day. I had let my root partition get a little too full (more than 95%) and as a result, several files in /etc/alternatives were corrupted (they read as multi gigabyte in size) and the link to lynx was one of them. No browser, thanks to my filesystem getting close to full. It wasn’t a one time deal either. If I ever forgot to empty caches and delete files, bam: corruption! I’d get to follow it up with scanning the partition for multi gigabyte files to delete and replace by reinstalling their .debs. Again, but one example of many. One can give me all the reasons in the world why that shouldn’t happen or how to fix it, but one would be ignoring “of many“.
So I totally see where you are coming from, and am really sick of Lemur2’s advocacy. Not because he is advocating something different and alien. Not because I have a stake in seeing some platform succeed or fail (I don’t tie my ego to how well my favorite software is doing in the marketplace). It’s because I have been there, have been that guy, and know that the reason he’s doing it is because he’s got a lot of time invested in his choice and feels compelled to defend it; not because it really is a flawless OS or easier to use or stabler (when we are talking about upgrades) or ideal for everyone, but because he has to justify his own wasted time by trying to get other people to waste theirs. It’s a crap move.
Selfish too, if he’s anything like I was. The more people use Linux, the more software and hardware will support it. So get more people on board so your life is easier! Never mind the hell you put them through by tossing them into an unfamiliar environment where their programs and wifi card don’t work. Of course it isn’t Linux’s fault it isn’t supported, but that doesn’t change the fact that it’s not, and that converts are going to have a rough time of it. But sure, get them all in through ceaseless promotion so someday your OS will be better supported and more pleasant for you to use! Great plan I thought, until it became obvious how dickish it is.
Not being one anymore, it is painfully clear how annoying the overzealous advocates are. I’m sorry I was one. Hopefully Lemur2 does himself a favor and tones it down.
What’s that Lemur2? I’m participating in some internet meme to denigrate Linux and KDE for no reason? Sorry man, I’m speaking my mind due to severe disappointment. If you’re going to tell me I’m doing it as part of some wider movement you can go screw yourself for calling me a liar.
If lemur2 wishes, he can read a rudimentary list of my complaints here:
Hmm, what kind of camel is that? A cigarret? Seriously, I have never heard of 2.6.31 failed to support CPU throttling, power management, suspend/resume etc. As I am currently using it with a very experimental nouveau driver for my NVIDIA card on a powerbook pro, and every thing works just fine.
If this really happens to you, sorry, but my suggestion is that do not pretend to be a pro if you do not have the expertice. Just wait for your distro to come up with a package, or, and even better, avoid using rawhide or similar experimental things. Just stay with a vanilla stable version of a distro.
I do understand her.
This is exactly the reason, why even though I work maninly on Linux and other UNIX systems, I run mostly Windows on my laptop.
I use GNU/Linux since 1994. The first time I installed it did not even recognize my IDE CD-ROM drive, forcing me to use floppy based install.
Since then I have used most of the well known distributions in different types of hardware.
But still, here we are in 2009 and still lots of things require tweaking to get it working properly, or break out of the sudden.
Yes, Windows does have its own share of problems, no need to discuss them. But it works.
Apparently, other OSs are allowed to have issues and problems, except for Linux.
She did, didn’t she? I thought she said she tried both self-compiled kernels and the packages provided by her distro. I got the impression that the update that originally borked her system was an auto-update from her distribution’s update system. I didn’t get the impression she was doing anything unreasonable, that one could reasonably have expected to be dangerous. And, like Stevie said, WereCatf is a reasonably knowledgable user. I don’t think it’s likely that user ignorance is the culprit.
I have Slackware 13 64-bit on my desktop at home, uses an AMD Phenom 9850 Black edition processor; Slackware 13 uses the 2.6.29 kernel. I do not have CPU scaling either (I haven’t even tried suspend/resume, and I probably never will: I never expect that to work). I have put some effort (and documentation reading time) into getting it to work, with no luck. I’m far short of a kernel hacker, but I possess a reasonable level of technical skill — certainly enough, I think, that I should be able to follow thru any reasonable procedure for getting CPU scaling to work, if it where possible.
I thought the same with my system for a long time (back in the 2.5.x days), until, by accident, I suspended the system (2.6.31-rc9), and… it worked. (Wait a sec, how long has *that* been working?!)
Hardware support takes longer for Linux because most of the hardware makers don’t actively support it, but the support does seem to eventually get in, amazingly.
Laptops are the worst simply because they have the most proprietary junk in them, compared to desktop machines made with more standard (and thus supported) components.
That was one of the reasons that kept me using XP on my older laptop.
The only OS that could properly support the laptop was the XP OEM version that came with it. Every other OS version, even Windows, had problems with the laptop.
The fan would just work at 100% non stop.
for me booting with the kernel boot option acpi_osi=”Linux” would “fix” the issue of the fan going all the time. But this doesn’t work for my current kernel (2.6.30-2-686 (Debian)).
So my fan is on all the time on my laptop and it’s annoying.
In the case of suspend/resume and laptops, the problem is ACPI.
http://www.linux.com/news/hardware/laptops/8253-how-to-suspend-and-…
“Since most manufacturers tend to add proprietary extensions to their implementations of ACPI, there’s a slight chance that your laptop might need some additional steps in order to suspend or hibernate. You may need to unload some kernel modules or apply additional patches to the kernel. Since there are so many different laptops, it’s hard to offer concrete advice.”
http://rdist.root.org/2008/10/17/all-about-acpi/
“ACPI could be seen as an attempt by Microsoft and Intel to take control over areas formerly owned by Phoenix/AMI and the numerous Taiwan integrators that actually build systems for companies like Dell. They had arguably good reasons for doing so, including the support cost of working around so many BIOS bugs. The conspiracy theorists may think that this was an attempt by Bill Gates to undermine Linux, but the pain that ACPI caused has been also borne by Microsoft. Gates never really got his proprietary extensions to ACPI, but Windows did enjoy a privileged position of OEMs using it as the validation suite for their BIOS implementations. Because there was no open interoperability testing, the spec was so complex, and information about the low-level BIOS and hardware were tied up in NDAs, open source kernels suffered many compatibility problems as new PCs appeared. The process definitely should have been more open and seems to be getting better, largely due to the efforts of Intel with Linux.”
In the end, it is surprising that suspend and resume works as well as it does on Linux.
Unless something managed to screw it up again (I haven’t tested it yet), this will be the first time you’ll be able to get both 7.1 LPCM audio over HDMI and hardware accelerated H.264/VC1 video playback using a stable, vanilla Linux kernel (albeit with the proprietary NVidia driver). HTPC owners, rejoice!