Linux Kernel 2.6.21 Released

diegocgteleline.es writes "Linus Torvalds has released Linux 2.6.21 after months of development. This release improves the virtualization with VMI, a paravirtualization interface that will be used by Vmware. KVM does get initial paravirtualization support along with live migration and host suspend/resume support. 2.6.21 also gets a tickless idle loop mechanism called 'Dynticks', built in top of 'clockevents', another feature that unifies the timer handling and brings true high-resolution timers. Other features are: bigger kernel parameter-line, support for the PA SEMI PWRficient CPU and for the Cell-based 'celleb' Toshiba architecture, NFS IPv6 support, IPv4 IPv6 IPSEC tunneling, UFS2 write, kprobes for PPC32, kexec and oprofile for ARM, public key encryption for ecryptfs, Fcrypt and Camilla cipher algorithms, NAT port randomization, audit lockdown mode, some new drivers and many other small improvements."

Damnit!

[FunWithKnives]

And I just upgraded to 2.6.20-15! (Kubuntu Feisty Fawn)

who the h3!! marked that flamebait?

[1800maxim]

Must be Linus!

Re:

[arth1]

Same here -- I just went to 2.6.20, mostly because 2.6.17 no longer was supported by my distro, and 2.6.18 and 19 are somewhat buggy compared to .17 and .20. I'll wait and see before I go to .21 -- there's really nothing in there that I /need/. Others may have different needs and wants, of course.
At this point, I'm mostly interested in bugfixes, and not features. YMMV.

(Anyhow, why did someone mod your post "Flamebait"? I can't see anything in it that's inflammatory at all?)

Published?

[CastrTroy]

What's with the headline? Who publishes OS kernels? I guess it could be grammatically correct and all that, but it sounds a little weird to me.

Re:Published?

[diegocgteleline.es]

It's me who sent the headline. "Publicar" (to publish) is what people usually uses for those cases in spanish. So there you've the answer for your question :)

(I also planned to add the number of months of development (almost 3, 80 days), but I forgot it)

Re:

[Bacon Bits]

Hey, I always thought "release" sounded a bit like urination, but you don't hear me complaining.

Bloat?

[ArcherB]

Is it just me, or are all these options that are added with every new release going to result in a bloated kernel? It seems like every release adds new stuff, but I never see anything outdated taken away.

Yes, I know that you can recompile and remove what you don't need, but most "non-uber-geek" users are not going to be able to handle that, and most distros are going to include a kernel with the kitchen sink compiled in.

Re:Bloat?

[qbwiz]

most distros are going to include a kernel with the kitchen sink compiled in.

No, most distros are going to include a kernel with the kitchen sink compiled as modules, taking up a few megabytes on the hard drive, but never loaded.

Re:Bloat?

[arth1]

Even if you compile something as modules, it does take up memory and resources. Much less, but still not negligible. There's hooks for the modules, plus tests in other parts on whether a module is loaded or not, in addition to much larger symbol tables.
And, of course, there's many parts that can not be made into modules at all, but have to be part of the kernel. And that makes a HUGE difference.

Is the difference really that big? Well, the machine I'm currently on has a bzipped kernel that's around 1.5 MB in size plus a 820 kB map. The alternative boot to a commercial distro (no name, no shame) has a bzipped kernel that's around 2.1 MB, plus a 2.3 MB initrd, plus a 1.2 MB System map.

The difference might not be staggering, but it's there, and the kernel is growing with each revision. Here's how the System.map has grown for the last few revision on this laptop, with no new options added:


-rw-r--r-- 1 root root 754620 Nov 30 18:32 System.map-2.6.17-gentoo-r8
-rw-r--r-- 1 root root 768275 Dec 28 15:57 System.map-2.6.18-gentoo-r6
-rw-r--r-- 1 root root 809157 Mar 26 04:28 System.map-2.6.19-gentoo-r5
-rw-r--r-- 1 root root 839371 Apr 25 22:45 System.map-2.6.20-gentoo-r6


That's an 11% increase without adding anything. Similar for the kernel itself (although that's harder to compare directly, due to the bzip2 compression). While not alarming, it's a trend towards feeping creaturitis that I think bears watching.

Re:

[xenocide2]

Pray tell, how did you eliminate the possibility of existing components growing, in order to conclude modularity itself is the problem?

Re:Bloat?

[Anonymous Coward]

2.6 and recent kernels in general are aimed at *modern computers*. Modern computers that have a *lot* more features and devices than ever before. There is absolutely no comparison to kernels back in the floppy days. If you're still using floppies you might as well be using an old kernel for whatever weathered machine you're talking about. And a microkernel's not going to reduce the overall amount of code, just push it around.

Not to mention that people who are using Linux for embedded work can still rip out a ton of stuff. You can completely omit module support and all sorts of things that contribute to so-called "bloat." The kernel is still and always has been very flexible about compiling in features, and all the code's there to mess with...

Small, fast, full of features; pick two. I think.

Re:

[InsaneGeek]

Are people really seriously concerned about this?? Maybe its me but I dont boot my linux (or windows) systems daily, or even monthly. The flexibility that modules give me, more than outways shaving 3 seconds off the boot time...

Re:

[John Courtland]

I use linux as my desktop OS on my laptop. I shut it off all the time, so startup time does matter to me. I realize I'm an exception, but it does matter.

Linux hibernate support

[canadiangoose]

Linux has amazing hibernation support. I use some software called Suspend2 on my laptop, and it works like a charm. It was a little difficult to get it going the first time, but now that it's installed I find it very fast and reliable. It doesn't even require any special support from the hardware!

Re:Bloat?

[Lxy]

I've noticed that each time I compile a new kernel, something has been moved to [deprecated] status that was still live in the last release. All the deprecated stuff is not compiled in by default, keeping the resulting bzImage size manageable.

Most distros compile everything as modules, which generally keeps the overall size of the kernel down. Sure, bzImage grows over time (not just because of new features, but typically new patches == more lines of code), but not significantly from release to release.

Most "non-uber-geek" users don't care what's in their kernel, and if they did, they'd learn to compile it themselves. Compiling kernels has gotten easier over the years. Chances are, if you care enough about how your kernel is compiled, you'll have the skills needed to do it yourself.

Re:

[ehrichweiss]

Are you referring to the 2.4 days as actual compile time or time it took you to work out what to include in the kernel config? I don't think it ever took more than 8 hours to configure and compile a new kernel on my end even if the machine doing the compiling was 1/10th the speed of, then, current computers.

Plus, I don't find it THAT hard to configure the new kernels but I take my distro's config file and remove anything I know I don't need rather than starting with a blank-slate config and THEN trying t

Re:

[init100]

Are you referring to the 2.4 days as actual compile time or time it took you to work out what to include in the kernel config?

He probably refers to the days of Linux 2.4.x, not the time taken to configure and build the kernel.

Re:

[Qzukk]

and most distros are going to include a kernel with the kitchen sink compiled in.

Actually, they use kernels with everything compiled as modules, and a separate initrd/initramfs to deal with loading the drivers required at boot time.

Re:

[HomelessInLaJolla]

I had begun making a similar observation to myself towards the 2.4.20+ line and definitely when 2.6 came out. It correlated nicely with the media increase in Linux coverage, the size and comprehensive nature of desktop environs like KDE and Gnome, the size of Xorg/Xfree86, and the increasing popular emphasis on web applications and the tion of HTTP and file-sharing protocol network usage over nearly anything else (spam excepted). I've almost decided that the computer programming age, as an affordabl

Re:

[Grishnakh]

I've almost decided that the computer programming age, as an affordable hobby for the non-specialist, is nearing the end of its lifetime. In a few more years you'll have the option of working with entirely standardized/commoditized/completely controlled (corporate DRM style) equipment or, if that doesn't appeal to you, then you'll have to go off a polar deep end and spend absolute bricktons of time and money assembling a system using a soldering iron, a breadboard, and specialty chips ordered from remote cl

Re:Bloat?

[shish]

I don't see this becoming a reality unless the US Government mandates it through legislation

The common sense of the US Government is the only thing standing between us and DRM hell? Oh dear :(

Re:

[Grishnakh]

I seriously doubt removing the extra modules has any effect at all on performance. These modules aren't even loaded unless your machine actually uses them. You save a bunch of HD space (but not much compared to the typical 100+GB size of drives these days), but that's about it.

You probably get a small performance increase due to using optimizations specific to your machine, like compiling specifically for your P4 or Core 2 or X2 processor, and maybe leaving out some things you don't need. But overall, I

Meh

[1010110010]

I haven't been able to get anything past 2.6.17 to boot successfully, I think they seriously hosed the ATA shit.

Re:Meh

[FudRucker]

personally i hate using an initrd.img and prefer to build ext2 & ext3 support right in the kernel making initrd unnecessary, if you compile file system support as a module you will need an initrd.img too so insetead of selecting an "M" select "*" you could try that...

P.S. i never use reiserfs so i can not say if this works with reiserfs or not...

Re:Meh

[tinkertim]

personally i hate using an initrd.img and prefer to build ext2 & ext3 support right in the kernel making initrd unnecessary, if you compile file system support as a module you will need an initrd.img too so insetead of selecting an "M" select "*" you could try that...

Its not just the file system you need, its the ability to spin the drive containing said file system too :) Its legacy HW that's getting fuzzy , not file systems. Not really sure why you hate initrds so much?

The initrd does many more things than load drivers. What if you have an AoE based storage network with many disk-less stations needing to use an OCFS2 single system image? Initrd's can do neat things besides loading modules, have a look at linuxrc. You can bring network adapters to an up/link state, negotiate iscsi targets, download a boot config from a resource controller, all kinds of goodies. Complex networks need to do lots of things before pivot_root gets called, and we need complex networks.

piix hasn't been 'quite right' since 2.6.16.29 on most of the legacy servers using PATA (IDE) I still have up and working, many of us have been having a difficult time with it. But progress is progress, and this is good progress so I guess my move to all SAS will be sooner than later.

Re:

[swillden]

Does it stop on "waiting for root file system to become available" or some such?

Re:

[swillden]

Hmm. No suggestions then.

Have you tried asking on LKML?

Re:Meh

[elFarto the 2nd]

IIRC after 2.6.17 the SATA stuff changed quite a bit (it changed from the old SCSI based stuff, to libata), and requires turning the new options on.

Regards
elFarto

Re:

[Ant P.]

libata supposedly has a driver for my old IDE-only chipset, but I've never got that to work. Maybe it just doesn't work for some people.

Re:

[FudRucker]

change /dev/hd?? to /dev/sd?? (even with IDE)

it might work, No Guarantees though...

Re:

[giorgosts]

My feisty has a 35% chance of mounting correctly the swap and ntfs partitions. On other occasions it boots ok, most of the times displays error and I have to reboot. I have the ext3 and swap partitions on PATA disk and ntfs on a SATA. Anyone else experienced that?

I also notice the new feisty to be much faster, but when under load, desktop slows down considerably. On edgy, however hard you loaded the machine, there was always the extra power for sth else if you wanted.

Feisty looks feels like a windows machin

Re:

[frogstar_robot]

I've changed my fstab to use UUIDs rather than device names. Once I had everything correctly labeled for a kernel my machine would boot up correctly but updating the kernel could break things. UUIDs stay the same even though disk device labeling is shifting sand at this point.

But why?

[camperdave]

change /dev/hd?? to /dev/sd?? (even with IDE)

Why was this done, though? All my IDE drives changed from hd? to sd? and for a while I couldn't boot. I had to rewrite my entire fstab. Parallel ATA may be going the way of the dodo, but it's far from dead.

Quite possibly.

[jd]

Just tried the latest kernel and it hangs on trying to fire up the second ATA instance. Not even a kernel oops, nothing. That's true whether I use the vanilla kernel or Red Hat's RPM. Something is screwed up, and from the sounds of it, there's more than one of us experiencing a failure at the same point, so that would be the obvious suspect.

Re:

[Daniel Phillips]

Just tried the latest kernel and it hangs on trying to fire up the second ATA instance. Not even a kernel oops, nothing. That's true whether I use the vanilla kernel or Red Hat's RPM. Something is screwed up, and from the sounds of it, there's more than one of us experiencing a failure at the same point, so that would be the obvious suspect.

This problem needs to go to lkml, and cc Andrew.

KVM management?

[drinkypoo]

Speaking of KVM (slightly offtopic, but not totally) are there any worthwhile management utilities for it yet? I actually ended up giving up for a while on KVM entirely because the video device is horribly slow and VDE support is not reliable, and I'm using vmware server, but I did have to give it a try. I'd love to use KVM (since I have supported hardware and it's Free software, and I'd love to minimize my use of the closed stuff) but beyond those problems (which will hopefully both be fixed relatively soon) there is simply no decent management software unless you're on redhate. Either virt-manager or libvirt is badly broken and won't work properly otherwise. UNLESS... has anyone out there gotten it working on debian/Ubuntu yet? I tried for a while, but I'm just not a good enough programmer and the programs ain't done yet.

Re:

[Anthony Liguori]

I actually ended up giving up for a while on KVM entirely because the video device is horribly slow and VDE support is not reliable

How do you mean the video device is "horribly slow". Also, what do you mean by VDE (as in, vde.sf.net?).

Re:

[drinkypoo]

How do you mean the video device is "horribly slow".

I mean that updates to the display (I was using the console interface, not a VNC) are, well, horribly slow. Some issues might be windows-specific, such as windows 2000 taking so long to turn the whole display desktop-colored (initializing video memory, I guess?) taking literally a minute or more.

And yes, by VDE I mean Virtual Distributed Ethernet. I tried both VDE and VDE2. With VDE2 it never worked properly. With vde, I could start kvm four times and h

Re:

[drinkypoo]

Uh, Synergy is a program used for sharing a keyboard and mouse across multiple machines. I use it daily, with a Linux host, and two clients; one Windows XP, one Mac OSX. It doesn't manage settings and run state for KVM virtual machines. Thanks for playing, though.

Re:

[Breakfast Pants]

KVM--Keyboard video and mouse.

Re:

[drinkypoo]

KVM--Keyboard video and mouse.

This is a conversation about Linux Kernel 2.6.21 and the context made it quite clear that I was talking about KVM, a feature added to the Linux kernel in 2.6.20 and with additional functionality in 2.6.21.

I didn't call the guy names or anything, if he had been paying attention he would have known what I was talking about, and he (?) :) posted an apology for not paying attention. Now it's your turn.

By the way, Synergy does NOT repeat NOT do the same thing as a KVM switch. It

Re:

[drinkypoo]

Seriously? For real? Just how much management can you need?

virt-manager is supposed to provide a vmware server console-like interface.

I need that much management if I'm going to leave vmware server.

Running shell scripts to run my virtual machines, and editing those scripts, and having those scripts create and destroy network interfaces etc, is an annoyance. I am capable of doing it, but I am lazy and would rather use someone else's tools. It's [ostensibly] only a matter of time before virt-manager works,

Tickless only for x86 now, still good news

[MarcQuadra]

I follow prerelease kernels and I've been waiting for this. I've found that running my VMWare hosts and guests with tickless, low-HZ, voluntary-preempted kernels is seriously reducing the overhead you get when you run more virtual CPUs than real ones in your box.

I can't wait for it to mature on PPC, MIPS, and x86_64! Right now it's 32-bit x86 only.

Re:

[JohnFluxx]

Can you explain this a bit more please for those of us who don't know what tickless means?

Re:Tickless only for x86 now, still good news

[AaronW]

It means that they were able to successfully remove the blood sucking parasites from the kernel.

Most kernels use a periodic system timer tick to do various housekeeping chores, like rescheduling tasks, sending packets, flushing files from the cache, etc. Usually this occurs at some periodic rate, i.e. every 1-10ms for Linux and every 10-15ms for Windows (according to this article [microsoft.com].

This is a bit wasteful of CPU resources, since the kernel might not need to do anything for quite a while, or it might want a high resolution timer with higher accuracy than normal system timer. For example, when the system is idle, the CPU still must wake up and process a timer interrupt for every timer tick, and if it's set to 1ms there are 1000 interrupts per second.

A tickless kernel instead only schedules the next tick for when it is needed, so if the system is idle and nothing needs to happen for 50ms, then the next tick will be scheduled 50ms later. On the other hand, if a timer needs to go off in 750 microseconds, the kernel can schedule the next interrupt to go off then, giving much higher accuracy.

Re:

[JohnFluxx]

Does this mean things like sleep(5) will sleep for 5ms for real now, without needing to change the scheduler etc?

Re:

[AaronW]

sleep(5) currently sleeps for 5 seconds, however, calls like nanosleep should have much greater accuracy with a tickless timer.

That's nice and all...

[rob1980]

... but does it run Linux?

Re:That's nice and all...

[diegocgteleline.es]

Since it supports para/virtualization yes, of course, it runs linux!

(Damn, finally there's an answer for this!)

Another solution to a timeless problem!

[iamacat]

Once again, it took many months of work to optimize an idle loop.

The list of changes can be found...

[diegocgteleline.es]

Here. [kernelnewbies.org]

Hotplug CPU support!

[matt me]

Hotplug CPU support? That must burn.

Re:Hotplug CPU support!

[harrkev]

Hotplug CPU support? That must burn

Nope. Works great. Let me demonstrate...fsdjlksd+++

Re:

[verbatim_verbose]

Actually, this has been possible with Xen paravirtualized guests for some time now.

At last!

[Dr. Stavros]

Glad to hear that it's been published. Where can I download the PDF? I heard that Darl dies near the end, but I want to read it for myself.

Sooner or later...

[heretic108]

Sooner or later, my /boot/grub/menu.lst will look like:

...
title Ubuntu, kernel 2.6.29-5-generic
root (hd0,0)
kernel /boot/vmlinuz-2.6.29-5-generic root=/dev/hda1 ro \
  coffee=cappucino,sugar=0,hourly \
  massageareas=head,neck,shoulders \
  catfeedingtimes=4_hours,15_grams
initrd /boot/initrd.img-2.6.29-5-generic
quiet
savedef ault
boot
...

Re:Sooner or later...

[ctr2sprt]

Yeah, the absurdly long kernel command lines in Linux really bug me. It's a symptom of the suckiness that is the PC BIOS, so I'm not really blaming the Linux people, but there are better solutions and have been for years. The FreeBSD loader [freebsd.org], for instance, is capable of loading the kernel and any modules required to bootstrap the system, reading configuration files, and running Forth (!) scripts. Such a loader would completely eliminate the need for initrds on nearly all systems[1] without sacrificing any power. You could also emulate Openboot or EFI - or more realistically a subset of them - using the PC BIOS to prepare for the future.

[1] initrd is a really awesome feature and it shouldn't go away. But it's massive overkill the way it's typically used, which is to load modules required to mount the root filesystem.

Re:

[ady1]

massageareas=head,neck,shoulders

In this revision, they've changed added a additional feature. Now the line is like:

massageareas=head[0],neck[0],shoulder[0-1]

Massage the wrong array member and the kernel will panic.

Re:

[SeaFox]

catfeedingtimes=4_hours,15_grams

you forgot the
type=feline_supplement,id=25

Dynaticks: picture

[Tribbin]

These dynaticks capabilities will certainly blow you off your socks!

http://www.spymall.com/catalog/images/bombclock3.j pg [spymall.com]

Mactel MBP C2D

[JumboMessiah]

As an owner of a Macbook Pro, I've been waiting for this to get released. The Dynticks integration will (hopefully) help lower power consumption and heat output. Though this will help reduce heat and power on all platforms, those running Linux on a MBP C2D know it's hard to keep the fans from spinning up from relatively little activity.

Next up is to get ATI to actually support any power saving features in fglrx on the MBP C2D and give the mAdWiFi [madwifi.org] guys more time to work out the features on the Atheros AR5008.

OSX, right now, still has a significant advantage in keeping heat and power consumption down. Even though, I imagine some will testify that even OSX is having a hard time with it...

Here's to testing out 2.6.21 tonight :)

Re:

[diegocgteleline.es]

The Dynticks integration will (hopefully) help lower power consumption and heat output.

Dynticks stil doesn't set the CPU to a power-savvy mode (it will in the future, but not in 2.6.21), so power consumption is more or less the same than without dynticks.

My own wishlist

[mangu]

I'd like to have some form of force feedback running. Yes, I know, this is a device driver issue, not a kernel one, but the drivers come with the kernel. My Microsoft Sidewinder wheel is the only reason why I still keep a Windows partition at home.

Does anybody there know a wheel that has good force feedback features in Linux? If I had one I would start contributing code to some project like torcs [sourceforge.net] or rars [sourceforge.net].

Re:

[battjt]

I'm a Linux guy since 0.12. I really, really, really want a Linux laptop that works. I settled for a Unix laptop that works.

The MacBook was $1300 and features DVD RW, firewire, USB, bluetooth, wifi, a video camera, audio (record and play), a 3d graphics card and a 150 minute battery (for the way I use it). The coolest features are the little things that Apple did right: the magnetic power cord, the simple, sturdy case, the pulsing standby indicator, etc. I install the GNU bits that I need, but most of

Cool, but...

[asninn]

That's cool, but is this really news that's Slashdot-worthy? Sites like LWN and KernelTrap have already reported this, and anyone who's interested in Linux development is pretty much guaranteed to follow the former at least, I think (and most likely the latter as well).

Re:

[styrotech]

I dunno, back in the day it seemed that Linux kernel development stories and release announcements were the mainstay of Slashdot news.

Re:Cool, but...

[npsimons]

That's cool, but is this really news that's Slashdot-worthy? Sites like LWN and KernelTrap have already reported this, and anyone who's interested in Linux development is pretty much guaranteed to follow the former at least, I think (and most likely the latter as well).

Considering that slashdot was (note the past tense) first and foremost a Linux/all things geeky site, I'd say this article is very slashdot-worthy. Not to mention that we get a fawning mac fan boy article every time Steve Jobs so much as farts. At least the Apple section can be turned off. Wish I could do the same with Microsoft and Windows articles.

Re:

[pembo13]

here here !!

Re:

[Sancho]

I don't follow the kernel releases anywhere but on Slashdot, honestly. Though I'm typically interested in the new features, Slashdot usually has more information that I care about, rather than the detailed changelog that has extremely low-level changes. Generally, I'm interested in stability, security, and feature enhancements, and the summary+comments tends to hit all of the important ones in any given release.

Re:

[Kjella]

Slashdot used to post almost every point release when I started on slashdot, but then again there were often really important kernel changes. In recent years kernel releases have been more like "Works good like before, maybe a little better, a few more drivers, exotic functions X, Y, Z". Trip down memory lane perhaps? Maybe the virtualisation support was news enough to qualify? Fire up the good old flamewar on whether the 2.6 kernel is stable or not? Besides, with various posts to randoms blogs/slashvertise

eCryptfs public key

[omnirealm]

The public key support for eCryptfs can handle more than just public keys. It includes a communication mechanism with a user daemon that can be queried from the kernel on file open events. There is a pluggable key module interface accessible through that daemon. OpenSSL is currently implemented, but there is nothing stopping anyone from writing a module to use GnuPG or any other key management/encryption backend, all in userspace. The module just needs to accept a key signature, and it can perform encryption and decryption based on whatever that signature refers to.

In other news, eCryptfs has recently been given the go-ahead for inclusion into Fedora:

https://bugzilla.redhat.com/bugzilla/show_bug.cgi? id=218556 [redhat.com]

In the meantime, you can grab all the userspace stuff from the eCryptfs SourceForge site:

http://ecryptfs.sourceforge.net/ [sourceforge.net]

do I really care?

[nanosquid]

I think since 2.6 has come out, I have lost complete track of kernel versions--it simply doesn't matter anymore to me. I think at this point, people could come in and replace the Linux kernel with BSD, Darwin, or Solaris and I probably wouldn't notice. Kernels have become a commodity.

No, it's just the hardware

[mangu]

people could come in and replace the Linux kernel with BSD, Darwin, or Solaris and I probably wouldn't notice.

This means your CPU is much more powerful than what you really need. I used FreeBSD a bit in the 1990s, but switched to Linux because the kernel allowed me better fine tuning in the 486 CPU I had at the time.

Today the CPU is way over my needs too, but I stick to Linux because, first, I have no need to switch and, second, Linux has better hardware support than the others you mentioned.

Right, brothers...

[Slashcrap]

But apart from virtualization with VMI, paravirtualization, live migration and host suspend/resume supportsupport for kvm, a tickless idle loop mechanism with unified high resolution timer handling, bigger kernel parameter-lines, support for the PA SEMI PWRficient CPU and for the Cell-based 'celleb' Toshiba architecture, NFS IPv6 support, IPv4 IPv6 IPSEC tunneling, UFS2 write, kprobes for PPC32, kexec and oprofile for ARM, public key encryption for ecryptfs, Fcrypt and Camilla cipher algorithms, NAT port ra

You joke,

[StarKruzr]

but I wonder if we're ever going to see 2.8 at this rate. The current kernel revision is MILES away in technology from 2.6.0. What will it take to move to 2.8, or (dare I say it?) 3.0? What qualifies as a major enough change?

Re:You joke,

[Anonymous Coward]

I wonder if we're ever going to see 2.8 at this rate

Linux 2.8 will compete with SunOS 6.0 as the best platform for running Duke Nukem Forever.

Re:

[einhverfr]

Linux 2.8 will compete with SunOS 6.0 as the best platform for running Duke Nukem Forever.

I dunno. Hurd will probably give them both a run for their money.

Anyway, isn't Duke Nukem Forever written in Perl 6?

Re:You joke,

[hypnagogue]

A change to the ABI.

Oh, sorry, I didn't realize it was a rhetorical question.

Change to the ABI

[StarKruzr]

Was that the case for 2.0 -> 2.2? 2.2 -> 2.4? etc.?

Re:

[Chirs]

The development model for 2.6 is very different from previous releases. See the "version numbering" section at http://en.wikipedia.org/wiki/Linux_kernel [wikipedia.org] for some details.

In general no, the ABI was not broken between previous releases (although the API has been broken numerous times).

Yep

[StarKruzr]

Read that article, but don't really understand the reasoning.

To me, it makes it appear as though nothing "really major" as changed in the kernel since 2.6.0, when, of course, nothing could be further from the truth.

Language

[StarKruzr]

does not an arbiter of a versioning system make. :p

More things have changed between 2.6.0 and 2.6.21 than changed between 2.0 and 2.2.

How's that?

Re:

[Vulva R. Thompson, P]

Well, the upcoming sister-raping feature could qualify as "major".

Details: [kernelnewbies.org]
"The features are tested in the -mm tree, but be warned, it can crash your machine, eat your data (unlikely but not impossible) or rape your sister (just because it has never happened it doesn't means you're safe):"

Re:

[frogstar_robot]

Yeah but you still have to force your sister onto the Sidewinder. If I'm going to pay good money for Windows then I simply must have better automation than that!

Re:

[number1scatterbrain]

I know that's you, Ballmer. Anonymous, my ass!

Re:

[justinlindh]

Stop it. This isn't the GameFAQs forum, and nobody cares if they're the first post here. If you don't have anything to contribute, then don't post.

On topic:
All of this built-in virtualization stuff sounds great. How long, on average, does it take the Ubuntu repositories to receive new kernels?

Re:

[Zonk (troll)]

Six months.

Feisty will likely be using a patched 2.6.20 until Gusty's released in October.

Re:

[MoOsEb0y]

6 digit (l)user :P

Re:Does it still crash after 49.7 days??

[diegocgteleline.es]

You're confusing Linux with this Windows 95/98 [microsoft.com]. However, this problem [microsoft.com] or this another problem [microsoft.com] are even more funnier

Re:

[TheThiefMaster]

I'll accept your funny problems and raise you another:
AMD K6/2 too fast for Windows 95: News [com.com] MS Support [microsoft.com]

Re:

[Gospodin]

Definitely not a Linux problem:

[webadmin@garden-ghost logs]$ uptime
5:05pm up 53 days, 5:54, 1 user, load average: 0.00, 0.00, 0.00

Re:

[Gospodin]

Hey, it's not my fault that some admin kicked out the power cable 53 days ago!

Re:Does it still crash after 49.7 days??

[Tom]

It doesn't and never did. However, the uptime clock wraps around after 497 days. Took me two hours of finding out why the box rebooted (and then why there was no indication of the reboot in the logs) one day to research that. That same box has since looped the clock a second time. So I can say for sure it stays up for more than 50 days. :-)

Re:Does it still crash after 49.7 days??

[Cro Magnon]

497 day wrap around? You should switch to Windows. I'm sure no such problem has ever been reported on that OS.

Re:

[Tack]

It doesn't and never did. However, the uptime clock wraps around after 497 days.

I guess that one got fixed at some point:

[root@blade1 ~]$ uptime
18:00:25 up 622 days, 23:00, 1 user, load average: 0.17, 0.22, 0.29
[root@blade1 ~]$ uname -a
Linux blade1.[redacted] 2.6.9-11.ELsmp #1 SMP Wed Jun 8 17:54:20 CDT 2005 i686 i686 i386 GNU/Linux

Re:

[AaronW]

As far as I know, Linux never had a 49.7 day problem, but it did have a problem at 497 days. I have a machine at home running the 2.4.20 kernel and every 497 days my uptime restarts, but it hasn't crashed. It's gone through 2 rollovers so far and has been up for over 3.72 years. It will hit its next rollover around September. I really need to build a new server... I just don't know if it will be as reliable as this one has been.

Re:

[AaronW]

As I said, it is running 2.4.20. The rollover problem was fixed in 2.4.21 so 2.6 should not have this issue.

Re:

[TeknoHog]

Support schmupport. There are lots of things in computing that Just Work (TM) without being 'supported' officially by the manufacturer. Conversely, 'support' doesn't guarantee very much, it's just a meaningless buzzword [iki.fi]. Then again, I41 welcome our new PWRficient overlords, especially when you can get laptops with them.

Re:

[Wesley Felter]

PowerPCs tend to be less than 100% compatible with each other, mostly because there is not a huge legacy of OSes that must be carried forward (e.g. people expect the latest and greatest x86s to boot DOS and Win95, but nobody exepcts some random old kernel to run on a new embedded PPC). Also, the PWRficient is a SoC with several new on-chip devices.

Re:PWRficient support?

[Grishnakh]

It doesn't work that way outside of x86-land. As another responder said, the PWRficient isn't just a CPU, it's a SoC (System on Chip). To compare to an x86 system, that would be like having a low-power CPU, north bridge, south bridge, SATA controller, ethernet controller (but not PHY), memory controller, I2C interface, USB controller, interrupt controller, etc. all wrapped up on one chip. This is quite common in the embedded world, where most PPCs are used these days (I'm working on one myself in my day job).

Since each SoC is totally different, except maybe for the CPU core, porting Linux (or any OS) to it is a little more difficult than just compiling it and loading it. Check out the arch/ppc or arch/arm directories for examples of all the different chips supported. While the work certainly isn't comparable to, say, porting Linux to an entirely different CPU architecture, it does require several new files with custom code to support things like the way interrupts are assigned to the specific functions on the SoC.

Worse, sometimes new drivers need to be written for certain on-chip peripherals, because some bonehead empire-building managers at the chipmaker wanted to justify a higher budget for their department by, instead of just re-using an existing USB controller or Ethernet controller design and plopping that onto the chip, putting together a whole team and spending months creating a new controller because it might improve performance by a whopping 5%. My last company, which made a lot of ARM-based chips, was especially guilty of this.