Linux Gains Two New Virtualization Solutions

An anonymous reader writes "The upcoming 2.6.23 kernel has gained two new virtualization solutions. According to KernelTrap, both Xen and lguest have been merged into the mainline kernel. These two virtualization solutions join the already merged KVM, offering Linux multiple ways to run multiple virtual machines each running their own OS."

Comments

So, will it run Windows?

[The_Fire_Horse]

just asking...

Re:So, will it run Windows?

[realdodgeman]

KVM (have been in the kernel since 2.6.20) already runs windows.

Re:So, will it run Windows?

[init100]

You mean Lguest? FTA:

Lguest doesn't do full virtualization: it only runs a Linux kernel with lguest support.

So the answer is no, Lguest does not run Windows. Xen runs Windows, but only if you have a VT-capable processor. Like Lguest, Xen can run Linux without a VT-capable processor.

Re:So, will it run Windows?

[zlatko]

Absolutely, running Windows XP on Linux is both easy to setup and performs quite well. I'm quite amazed with kvm technology for both reasons. This is not to say that Xen is bad, but it seems so much harder to setup, that I haven't even tried. kvm is dead simple.

Re:So, will it run Windows?

[baadger]

For those wondering the best (IMHO) FOSS solution for those wanting to run Windows on Linux without a VT-capable processor is Virtualbox

Re:So, will it run Windows?

[Anonymous Coward]

If only it didn't require QT-libraries for installation... :(

Re:So, will it run Windows?

[Anonymous Coward]

Its 2007, how much hard drive space do you have? Seriously, why are people still saying that program would be great if I didn't need QT, or GTK, or whatever. How stupid...I guess you'll just be missing out

Multiple ways to run Multiple OSs

[Anonymous Coward]

erm.......why?

Re:Multiple ways to run Multiple OSs

[tgatliff]

Competition is a wonderful thing!! I suspect three solutions probably will quickly end the vmware / XEN disagreements that went on for so long... :-)

Re:Multiple ways to run Multiple OSs

[Iphtashu Fitz]

A number of reasons. One is to be able to run different linux distros on the same machine for testing purposes. Another is to set up two completely different environments that run tasks at different times.

I used to work for a search engine company (not Google) that has thousands of linux servers. After doing a bit of research they discovered that the vast majority of these machines are idle for a good amount of time. Rather than buy new servers they simply installed Xen and intellegently divided up the physical hardware to perform their different tasks. Now instead of separate physical servers to do web spidering, data analysis, log processing, etc. they've combined these tasks onto the same physical hardware but kept them as individual virtual servers.

Re:Multiple ways to run Multiple OSs

[Anonymous Coward]

I don't think the parent poster was asking why somebody would use virtualization, but rather asking why there are so many different programs you can use for virtualization.

Re:Multiple ways to run Multiple OSs

[init100]

Why what? Why multiple virtualization solutions? Because each solution has its own advantages and disadvantages. Use the solution that fits your needs best.

Re:Multiple ways to run Multiple OSs

[Doctor O]

Um, you know, that is basically what virtualization is meant for, and it's common practice in *every* company with more than 5-10 servers. I'm a bit surprised this makes for a "+4, Informative" on a technical site like the dot.

I wonder why all the Linux priests here on the dot don't get that Virtualization is the magic key to drive Linux adoption among the typical Windows luser crowd. Since VMware Server was free, I've converted almost *everybody* I help with computers to kubuntu with VMware running a XP VM for the few things that people "need to keep" like say Quicken or their exotic label printers' proprietary software or whatnot. I notice that people don't panic because Windows is "still around", but when I give them a short introduction to their new system, I use Linux software all the time, and present Windows at the end with the comment "oh and if you need your label printer, just fire up this windows button (VMware link on desktop) and off you go.

Just snapshot their Windows VM after configuring it, and when it unavoidably fubars, you just login remotely via ssh with X forwarding, launch the VMware console and revert the VM to the snapshot.

This takes around half an hour longer than giving them Windows-only, but it will save you a *ton* of grief even short-term. I have relatives who fuck up their Windows installations at least biweekly. I've shown them the 'revert' button and put the snapshot on "protected", so they can never fuck up beyond all repair anymore. None of them has managed to do *any* harm to the k|ubuntu I gave them, and there are real boneheads among them.

Re:Multiple ways to run Multiple OSs

[Anonymous Coward]

Are they going to be able to play their directx games that way?

Re:Multiple ways to run Multiple OSs

[mrsteveman1]

How is someone who has "converted almost *everybody* I help with computers to kubuntu with VMware running a XP VM" NOT a Linux Priest?

That's probably the worst thing you could possibly do to a home user, if someone wants Linux they can decide for themselves, pushing it on people even if you help them resolve problems, is a bad idea, one that i hope doesn't need to be explained to you.

Re:Multiple ways to run Multiple OSs

[mrsteveman1]

That little snapshot trick you have taught them is also ripe for "where the fuck did my report go?"

Re:Multiple ways to run Multiple OSs

[Wolfrider]

--Duuuude... You just made my Friends list for that. :)

Re:Multiple ways to run Multiple OSs

[Wolfrider]

--That's what BACKUPS are for... With Win2k / XP Pro, you can use ntbackup. Failing that, you can always install vsftpd on the host and FTP your files over.

Re:Multiple ways to run Multiple OSs

[Wolfrider]

--Install Nomachine NX on their Linux host, and your remote experience will be even better.

http://www.nomachine.com/

Re:Multiple ways to run Multiple OSs

[Doctor O]

No idea, probably not, but really I couldn't care less. All DX gamers I know are teens, and I'd *never* help one of those. They're young enough to either learn how to run a Windows machine or go dual-boot. I've explained the latter to some, but won't support it. I've got enough with the old folk already.

Re:Multiple ways to run Multiple OSs

[Doctor O]

Probably not, no idea, but I couldn't care less. All DX gamers I know are teens, and I don't help those beyond some pointers on what to google. They're young enough to either learn how to run a Windows machine or go dual-boot. I've helped some with the latter, but won't support it. I've got enough already with the older folks.

Re:Multiple ways to run Multiple OSs

[Doctor O]

No, I'm not a Linux priest at all. I'm just being pragmatic and going the way of the highest maintainability. Plus, I'm fed up with repairing stuff that broke because people keep downloading warez to perform basic stuff every other fricking OS has legally available for free.

And I think you're not considering my target audience enough. We're talking about people who have no clue whatsoever and just want to surf the web, write e-mail, rip their music CDs to MP3s and listen to them, and maybe some misc stuff like converting their vinyl to CDs or burning DVDs of the family holiday videos. They don't know how to do it at all, and I'll rather keep them warez-free and get a maintainable system as a bonus. For them it doesn't make a difference what they use, as they couldn't tell Windows from Linux anyways.

Re:Multiple ways to run Multiple OSs

[Doctor O]

Aw, c'mon, spare the strawman. That's a no-brainer. You don't keep their "My Documents" in the VM but link it to someplace in the Linux file system where it's regularly backed up. (You DO know that all Windows system directories can be easily symlinked or moved, yes?) Add in Subversion and you can even have something similar to the Time Machine feature the upcoming OSX 10.5 will have, even if it won't be as accessible for the lusers. (But as they always refer to me first when they lose something instead of even searching for it locally... you get the point.)

Re:Multiple ways to run Multiple OSs

[Doctor O]

I looked at the features and liked the management console, but apart from that, what does it give me over a ssh -CX user@host? Everyone I support has at least 6 MBits with 768kBits upstream, and that's *very* responsive with plain vanilla compressed X forwarding as offered by OpenSSH, available for Linux (bummer), OSX, the BSDs and even for Windows through Cygwin-X-on-USB-stick-no-installation-required-tha nk-you.

Mind you, I'm curious, not trolling.

And: My pet peeve is that the VMware Server console on the remote machine sometimes runs shitty through X forwarding, but you can always just tunnel a connection for a locally running VMWare Server console, that works like a charm.

Re:Multiple ways to run Multiple OSs

[Doctor O]

Oh, my second three-digit-fan. :) Reciprocal.

I just hope you didn't friend me as a Linux priest, because I'm not. I'm from the "use the best tool for the job" crowd and just getting *very* pragmatic with age. Personally, I use OSX at work and Linux/FreeBSD at home, but that's really only because they fit my usage patterns best. I was on Windows quite a long time, and still use it at work for, say, 10-20% of the time. I'm just getting sick and tired of it, because compared to the alternatives, it has really lost a lot of ground, and Vista is not making it better. Mac OS and Linux have improved significantly in the last 7 years while Windows has actually gotten worse. IMHO Microsoft has a serious problem here. Maybe Vista will *really* be a huge flop and get them working on a completely new OS, designed from the ground with security in mind, that would be great. I mean, with their monopoly, they also have the power to force the market to the new version, if they have the balls to take the flak just this once. Apple has done it with OSX, and there is no reason why Microsoft couldn't.

That was my Microsoft rant of the day. If you feel like dancing or getting naked, don't worry about me.

Re:Multiple ways to run Multiple OSs

[Doctor O]

Interesting. I accidently closed the tab directly after clicking "submit", but it did go through. Why does it take several seconds from submit to response page then?

Re:Multiple ways to run Multiple OSs

[Wolfrider]

--Try it on (1) remote install. Trust me. I'm all up on Linux and Vmware, and NX is the ONLY thing that makes Server remote console usable over Internet for me. Plus, you get a full terminal-server desktop along with the bargain. (I use Icewm or XFCE.) // Haven't had to use anything that slow in years, but I heard it also works 1/2way decent over DIALUP.

Re:Multiple ways to run Multiple OSs

[Wolfrider]

--'Sall good. I'm very much eclectic** myself.

** 1 : selecting what appears to be best in various doctrines, methods, or styles
2 : composed of elements drawn from various sources; also : HETEROGENEOUS
( Courtesy m-w.com )

Why?

[realdodgeman]

Wouldn't it be enough with one? Or maybe they could have merged all the features into one VM.

I think this will confuse users. Choice is good, yes, but 3 VMs in the kernel? Sounds like overkill.

Re:Why?

[QuantumG]

Yeah, like all those file systems the kernel supports. What's with that? You only need one. Man. Choice is good and all, but it sounds like overkill.

Don't get me started on buses.. PCI, USB, SCSI, IDE, how many do you need?!

Re:Why?

[realdodgeman]

In what way are hardware drivers similar to VM technologies?

More VMs does not necessarily give more OS support. I can understand the need for a VM like lguest, since it does not require a CPU with virtualization technology. But wouldn't it benefit the user more if this where intigrated into KVM instead? I don't know how possible this is, but it sure would do my choice a lot easier.

Re:Why?

[Anonymous Coward]

I'm wondering what's NOT going to be put in the kernel eventually. I mean what's next, MPlayer? At what point do we say enough is enough?

Re:Why?

[Stephen Williams]

There are already 47,000 or so filesystems in the kernel. Linux has always been about choices.

Just as with filesystems, what will probably happen is that distributions supporting virtualization will pick one. Unless the user selects "super-duper expert installation mode" or whatever, he/she will get the distro's default.

-Stephen

Re:Why?

[Anonymous Coward]

I'm wondering what's NOT going to be put in the kernel eventually. I mean what's next, MPlayer? At what point do we say enough is enough?

If you want a minimalistic kernel with only the bare minimum, the Windows NT microkernel is available for you.

Re:Why?

[QuantumG]

Which is why I mentioned file systems...

That said, you mentioned KVM.. KVM (for Kernel-based Virtual Machine) is a full virtualization solution for Linux on x86 hardware containing virtualization extensions (Intel VT or AMD-V). (from here). It *is* a hardware driver.

Re:Why?

[Anonymous Coward]

For one, they all fill different needs.

KVM allows you to virtualize any PC OS, as long as you have a VT CPU. lguest allows you to run another copy of Linux. Xen sits somewhere in the middle - you can run any Xen-compatible OS, not just Linux, but you can also run normal OSes if you have a VT CPU.

Xen is hardly lightweight. It's really suitable for servers, but it's too intrusive for general use. KVM and lguest, on the other hand, are pretty unintrusive, don't radically change the system, and can simply be used by regular applications. And their functionality doesn't really overlap.

Users will never see them anyway. Now they're part of the kernel, users will just see a program that makes use of them.

Re:Why?

[Anonymous Coward]

You forgot the school bus

Re:Why?

[evilbessie]

IDE is not a bus, don't confuse this with ATA (more recently SATA and PATA). IDE == Integrated Drive Electronics.

Re:Why?

[QuantumG]

bus error: driver not found.

Re:Why?

[MrNaz]

The other issue I have is with the inclusion of lguest. It is a highly immature piece of code that is not really usable in anything resembling a production environment.

Why is the Linux kernel being bloated with things that are clearly not going to be used by anyone other than tinkerers and hobbyists? It just gives weight to the Microsoft claims that Linux is for hobbyists. It's one thing for hobby tools to be bundled with distributions like Gentoo, but for there to be code like this directly in the kernel, well, it makes it hard to argue that Linux is a serious kernel for serious applications.

I think that your flippant comment may perhaps be intended to highlight the fact that Linux is not intended to be a microkernel, but nonetheless that does not mean that it should be bloated with everything under the sun. I think that the bloated mess that Firefox has become highlights the fact that just because a program is open source and starts good, does not mean that it can't become a bloated sack of fertilizer through poor technical decision making.

I'm just glad that there are other open source operating systems that have remained purist to their initial goals. While this leads to slower development, it also ensures that they won't one day turn around and realize they've traveled a decade in the wrong direction.

Re:Why?

[rocket22]

I think having different options will make virtualization stronger on the linux ground

Re:Why?

[larry bagina]

Yeah, like all those file systems the kernel supports.

It doesn't support ZFS.

Re:Why?

[Anonymous Coward]

users may never see them, but it increases the kernel footprint and is a potentially exploitable virus/trojan/rootkit vector.

Re:Why?

[drinkypoo]

In what way are hardware drivers similar to VM technologies?

in this situation the analogy is clear. As time went on, people discovered new designs for virtualization and decided to implement them. Each design has strengths and weaknesses that make them appropriate for different situations. The same is true of hardware buses; older buses tend to be cheaper to implement. There are exceptions, it's probably cheaper (or will soon be cheaper due to economies of scale) to implement PCI-Express at PCI bandwidth than it is to implement PCI itself. It's certainly cheaper to implement firewire than SCSI (in spite of this, there are practically no native firewire storage devices. But anyway.) (And firewire, which goes up to 800MHz which peaks at 100MB/sec, is superior in most ways to anything up to and including LVD SCSI, including speed, simplicity of cabling, etc etc) Can you tell I have an ax to grind?

But anyway, the point is that we have UML, which runs linux as a process; we have this new lguest, which runs linux as a module; we have xen which is full virtualization without a need for VT, we have kvm which is like xen but does need VT, we have vmware which is also pretty much like xen (and doesn't need VT, although I was under the impression newer versions of vmware would take advantage of it if present, for a speed boost.)

There's some other examples too, but these are enough to talk about right now. Suffice to say that each approach has advantages and disadvantages. But they're useful for different things!

For maximum separation, for example, you could have a Linux that ran servers inside of different UML processes. While exploits in UML would still be possible, this would stop a privilege escalation bug in one server from affecting another. I envision a tool that tracks dependencies and generates the UML filesystem images automatically. Syslogging is done through the virtual network, to the syslog on the core system. Want to test a package? A command to run it in a UML might be as simple as running fakeroot. (fakelinux?) You could do all of this with this new lguest system, instead of UML.

Meanwhile, you're still going to need a full virtualization solution to run non-linux operating systems under Linux (at least until a cobsd (see "colinux") comes out - I forgot about that one for a moment) so there's still a purpose for that.

Re:Why?

[fritsd]

I didn't realize that iguest is going to be turned on by default. Oh wait.. it probably isn't. If you ever did a

make xconfig

you can see that the very first option is "code maturity level options", and that there are hundreds of features which are by default NOT TURNED ON and therefore do not show up in "anything resembling a production environment". And I'm not talking about kernel modules here, but things like CONFIG_MATH_EMULATION (under "processor type and features" near the bottom of the page) which isn't necessary anymore since intel brought out the i486DX processor in 1989 or so.

Re:Why?

[master0ne]

there only "in" the kernel if you compile them in on your next kernel compile, or if your distro compiles them in at install.... if they wernt "in" the kernel, then linux would have no way of understanding those filesystems, or methods of virtualization... asfar as end user confusion, if your technical enough to NEED them, you'll understand what they do and how they work, if not, then having them "in the kernel (or even actually compiled into that spefic kernel) wont hurt anything, as the user will probably never use them. I agree too many choices only creat confusion and a segmented market, but in the case of File systems, and virtualization, the choices dont "compete"... there may be some overlap, but they each solve different problems and needs of different people....

Re:Why?

[Anonymous Coward]

ATA is just a new name for IDE. PATA is a backronym used to distinguish "old" ATA from Serial ATA. As I'm at it, ATAPI stands for "ATA Packet Interface" and is a sub-set of SCSI over ATA.

Re:Why?

[Chris Snook]

Xen has all the features that KVM and lguest have. That's the problem. Xen is extremely complex, and the patches to support it are very invasive. This is why KVM beat it getting merged. LWN infamously predicted Xen could get merged as early as 2.6.10, whereas lguest was only created a few months ago, weighing in at a mere 5000 lines of code.

Xen does some really cool things, but it has a lot of human overhead in terms of management and maintenance that the other two don't have. Now you get to pick the right tool for the job, which is how it should be.

Re:Why?

[Courageous]

KVM is doing paravirt also, FYI

Re:Why?

[init100]

we have xen which is full virtualization without a need for VT

Actually, Xen uses paravirtualization if VT is not available, and can only run operating systems with Xen guest support in those cases.

we have vmware which is also pretty much like xen (and doesn't need VT, although I was under the impression newer versions of vmware would take advantage of it if present, for a speed boost.)

VMware isn't like Xen, in that it can run unmodified guest operating systems without VT. You are correct in that VMware takes advantage of VT if available.

Re:Why?

[init100]

The electrical interface of IDE is certainly a bus, since it connects more than one device to each channel. On the other hand, SATA is not a bus, it is a point-to-point link, which connects exactly one device to each channel.

Re:Why?

[SirTalon42]

It'll only increase the kernel foot print IF you compile them into the kernel, which they won't be enabled by default.

Re:Why?

[init100]

Only if enabled in the distribution. It doesn't harm anyone to have it available in the kernel source tarball. And both KVM and Lguest are implemented as modules, so if you don't load them, they aren't there.

Re:Why?

[Ultra64]

Right, because it's just *impossible* to compile a kernel with only the features you need.

Re:Why?

[Anonymous Coward]

Not exactly. VMware takes advantage of VT if it's available AND it's better than VMware's software-based binary translation. In most cases it's not. (Of course "next-gen" VT / AMD-V might be a bit different.)

Re:Why?

[evilbessie]

However a SATA HDD is still an IDE HDD, and by the same token PCIe is not a bus, I was just correcting the use of IDE because this is _not_ a bus but something describing the fact that the drive controller and drive are in the same physical device. So thanks for the correction regarding SATA, but don't use IDE as it is just plain wrong (as are so many other people).

cf. http://en.wikipedia.org/wiki/Advanced_Technology_A ttachment

Re:Why?

[MrNaz]

Right, because it's a better idea to turn lots of stuff on and have users *recompile* to turn off stuff they want. *cough* Windows *cough*.

Re:Why?

[drinkypoo]

Actually, Xen uses paravirtualization if VT is not available, and can only run operating systems with Xen guest support in those cases.

Er, I think that's what I meant to say, but clearly I didn't; thank you - I sit corrected.

Re:Why?

[Haeleth]

Why is the Linux kernel being bloated with things that are clearly not going to be used by anyone other than tinkerers and hobbyists? It just gives weight to the Microsoft claims that Linux is for hobbyists.

Well, for one thing, Linux is for hobbyists. It is also for a wide variety of other users, but it would be dishonest to deny that hobbyists are an important part of the Linux audience.

It's one thing for hobby tools to be bundled with distributions like Gentoo, but for there to be code like this directly in the kernel, well, it makes it hard to argue that Linux is a serious kernel for serious applications.

Why? That's like saying that Windows is unsuitable for business use just because it comes with Freecell.

Re:Why?

[Wolfrider]

--Yah, I'm a Firewire-400 fan myself; even setup a simple P2P network with it. Search vmware forums for details.

--Only prob with FW800 is the physical plug incompatibility (which was a bonus for USB1.1 -> 2.0 users) and lack of widespread installs.

--Some ppl say FW is dying, but I don't believe it. All the movie-editor ppl and such are hardc0re FW fans. (USB2 is Teh Suxor with >1 drive on the bus.)

--On the plus side, you can get an external-drive carrier/enclosure for ~$50-60 that will do FW400, USB2, AND Sata... It roxxors with a SATA-250GB drive; just runs a tiny bit warm.

See:
http://www.google.com/search?q=coolmax+cd-311&ie=u tf-8&oe=utf-8&aq=t&rls=com.ubuntu:en-US:official&c lient=firefox-a

Re:Why?

[drinkypoo]

On the plus side, you can get an external-drive carrier/enclosure for ~$50-60 that will do FW400, USB2, AND Sata... It roxxors with a SATA-250GB drive; just runs a tiny bit warm.

I have purchased three, all for around that price or less. I have one metal gear box with a fan that holds my music and I have a 250GB and a 80GB volume in cheaper, crappier enclosures that I use for storage/scratch. All are USB2/IEEE1394. I pretty much never use USB2 at home.

I don't think having the SATA drive is necessary, although firewire is certainly faster than they are (even at 400Mbps.)

lguest doesn't need VT

[physicsnick]

More importantly, lguest apparently does not require a CPU with virtualization technology. This is exciting news for those of us running on older hardware.

As a cross-platform developer, I'm interested in installing Windows on a virtual machine instead of dual-booting, and the current virtualization technologies don't cut it for me; VMware player is proprietary and doesn't work with my wireless card, QEMU is just too darn slow, and everything else requires a VT CPU. I'm looking forward to trying out lguest.

Re:lguest doesn't need VT

[Anonymous Coward]

FYI, Xen hasn't required VT since the beginning either. The only problem was you needed a specially patched kernel because linus didn't like how xen implemented their hooks into the stock kernels. It looks like that has been resolved however.

Re:lguest doesn't need VT

[Anonymous Coward]

> QEMU is just too darn slow

What are you talking about kqemu is most definitely faster and better featured than lguest at this stage.

1. Install kqemu
   
2. modprobe kqemu
   
3. qemu --with-kernel-kqemu [...]
   
4. ???
   
5. Profit?
   

Re:lguest doesn't need VT

[Anonymous Coward]

>

There are two other options: Virtualbox, which has a GPL'd and a freeware proprietary version (and a very nice GUI), and kqemu which is the kernel module qemu accelerator. Both are faster than normal qemu alone. Also look into seamless virtualization where you use rdesktop to run apps...might need on of those terminal server hacks to remove the license restriction though.

Re:lguest doesn't need VT

[init100]

As a cross-platform developer, I'm interested in installing Windows on a virtual machine instead of dual-booting

Then Lguest is not for you, since it requires a guest kernel with Lguest support, which Windows most certainly don't have, or will ever have.

Re:lguest doesn't need VT

[flyingfsck]

Hmm, I have recently played with a variety of VM systems. VMware Server is slow but it works really well - and it is free. So if you want a production quality solution, use that. Qemu can be sped up enormously if you can get the kernel optimizations compiled and installed. However, Qemu (Virtualbox is based on it) doesn't work with everything - for example I found that MS Windows 2003 would crash after an hour or so. The other solutions such as Xen require special processors.

Yes, VMware Server and basic Qemu are slow, but they are nevertheless quite usable and you can speed things up by toning down the requirements in Windows - select classic mode and 16 bit colour with no animations for example, just like one would do for VNC. I did my tax return on a 1.6GHz laptop using Qemu (without the optimizations) and WinXP and it was OK.

I wasn't aware of Iguest, so I'm going to give it a try.

Re:lguest doesn't need VT

[flyingfsck]

BTW, Lguest only runs Linux on Linux, so it is similar to Xen.

KVM can run Linux on Linux or Windows on Linux, but it needs special hardware, just like Xen.

Soooooo, if you wish to do Windows on Linux and you don't have special hardware, then you are pretty much limited to VMware Server and Qemu (Virtualbox).

Re:lguest doesn't need VT

[Anonymous Coward]

BTW, Lguest only runs Linux on Linux, so it is similar to Xen.

Except that Xen can run FreeBSD, NetBSD, OpenBSD, OpenSolaris, Plan 9, and more, on Linux or NetBSD.

Could somebody clear this up for us?

[Tribbin]

What are the pro's for heaving two implementations of, seemingly, the same solution?

Re:Could somebody clear this up for us?

[GreyWolf3000]

The more people who use both solution, the quicker the kernel team can figure out which one works better, and go with that.

Re:Could somebody clear this up for us?

[QuantumG]

Actually, it doesn't work like that. What actually happens is that the code which is maintained poorly gets dropped. So if there are dedicated people working on KVM but no-one actually working on lguest, eventually something will change that results in lguest not working anymore. Eventually people will drop the broken code from their tree until someone fixes it. If no-one fixes it, then it'll never be picked up again. There's no "oh, lguest is actually faster than KVM, we should all work on that".. it's individuals making their own decisions on what to work on (be it that they find it interesting, or they find that bit of code more pretty, or they are paid by someone to work on it) and those individuals are responsible for what happens to that code.

As long as N solutions are maintained there will be N solutions in the kernel. A solution won't be dropped because it performs worse.. or any other "technical" reason.

Re:Could somebody clear this up for us?

[sekra]

It's not the same solution because lguest and KVM have different goals. While KVM is trying to use as much hardware virtualization support as possible to gain full speed, lguest is not using these functions to run on more hardware. XEN tries to do everything and is thus a bit more bloated, but also with more functionality. Choice is good, just take the solution which fits your requirements best.

Re:Could somebody clear this up for us?

[GreyWolf3000]

Good point...but I believe that, over time, the one that most users choose will end up being the most actively maintained.

Re:Could somebody clear this up for us?

[bfields]

Actually, it doesn't work like that. What actually happens is that the code which is maintained poorly gets dropped.

That's a pretty unfortunate situation if the unmaintained code is still actually used by someone. Even if another alternative has come along with a superset of the given features, if they provide different system interfaces--so if it would mean rewriting scripts or applications or retraining users--then the migration can be a pain. And you want people to be able to drop a new kernel into an old working system--otherwise it's hard for them to get security fixes, for example.

So userspace-visible stuff shouldn't really be going into the kernel unless everybody's pretty confident that it can be maintained indefinitely.

That said, yeah, if someone notices that filesystem FooFS has been completely broken for ages and nobody has even noticed, then that's a pretty good argument for dropping it. But even then it's not just because it's unmaintained, it's because at that point you're pretty sure nobody really gives a crap about it.

Re:Could somebody clear this up for us?

[Cygfrydd]

Except the Internet clearly demonstrates reverse Darwinism at work: survival of the most idiotic.

@yg

Re:Could somebody clear this up for us?

[Chris Snook]

These aren't even close to the same solution. KVM provides hardware-assisted virtualization, with Linux as the hypervisor. Lguest provides linux-in-linux paravirtualization (no hardware support), and is extremely lightweight (5000 lines of code, total), but lacks many advanced features. Xen provides both paravirtualization and full virtualization, runs under a custom hypervisor intended to run multiple different OSes (Linux, Solaris, Windows, etc.) simultaneously, and has a plethora of sophisticated features, such as live migration (and all the maintenance headache of the correspondingly huge codebase).

They each fill very different niches, so there are very good reasons for having all 3 in the kernel.

Re:Could somebody clear this up for us?

[MoxFulder]

Actually, it doesn't work like that. What actually happens is that the code which is maintained poorly gets dropped.

That's a pretty unfortunate situation if the unmaintained code is still actually used by someone.

(...)

That said, yeah, if someone notices that filesystem FooFS has been completely broken for ages and nobody has even noticed, then that's a pretty good argument for dropping it. But even then it's not just because it's unmaintained, it's because at that point you're pretty sure nobody really gives a crap about it.

The Linux kernel *almost never* drops support for any devices/filesystems unless (a) it's INCREDIBLY obsolete and NO ONE is using it, or (b) it's been superseded by something clearly better and there's a straightforward upgrade path.

For example, if you read the kernel changelog summaries on LWN.net, you'll see that support for IBM PC/XT hard disks was only dropped in the last couple years... although they have been obsolete since the late 80s and perhaps literally no one has used them for 5-10 years. And support for the original "ext" filesystem was removed a few months ago, despite the fact that it's been completely superseded by ext2--which was introduced in 1993.

As Greg Kroah-Hartman has pointed out, the kernel developers are perfectly willing to maintain a driver for which only a single piece of hardware exists in the whole world!

Re:Could somebody clear this up for us?

[init100]

Or they have different advantages for different use cases, which would mean that all of they stays there.

By the way, to merge means to take two things and make one thing of of them. You don't merge out something out of another thing. A better word would probably be split.

Re:Could somebody clear this up for us?

[Anonymous Coward]

That's how forward Darwinism works as well: cockroaches, bacteria, and dinosaurs. Mammalian smarts are unusually, and likely as short lived as any other animal with an extremely overdeveloped organ.

Re:Could somebody clear this up for us?

[ankura]

Additionally the Xen patchset adds only support for building Linux as a DomU (or guest)
under the Xen hypervisor. The kernel doesn't have any Xen hypervisor or the code for
Xen management domain Dom0.

I RTFA twice and thought to myself...

[PrimeWaveZ]

Wow, there are now three VM solutions built right into the kernel? What are they going to do next? Merge emacs?

Re:I RTFA twice and thought to myself...

[chabotc]

it might be worth remebering that the _kernel_ part of these VM solutions have been merged into the kernel, and not the userland tools (they are seperate packages). A VM needs certain kernel hooks for the hardware virtualization, hence the need for a kernel 'driver(s)', and the VM scheduling happens there too.

So the comparisment with emacs is very inaccurate, emacs is a userland tool, and doesn't have kernel modules :-)

Re:I RTFA twice and thought to myself...

[brunes69]

I once considered writing a kernel emacs accelerator module, but later decided it would be easier to just run Linux inside of emacs!

Re:I RTFA twice and thought to myself...

[Anonymous Coward]

What are they going to do next? Merge emacs?

Yes - with vi.

Re:I RTFA twice and thought to myself...

[timeOday]

three VM solutions built right into the kernel? What are they going to do next? Merge emacs?

Hey, that's not a bad idea.

Re:I RTFA twice and thought to myself...

[Anonymous Coward]

Oh, yes yes it is a bad idea.

But the real question is...

[Anonymous Coward]

...will it run Linux?

What about kqemu?

[Anonymous Coward]

It's full featured doesn't require CPU VT support and is widely used (ie: tested).

Is the linux kernel community going through a NIH stage?

Re:What about kqemu?

[QuantumG]

If kqemu want to integrate their kernel components into the kernel they can. It's not the Linux developers going out looking for things to add to the Linux kernel... or them developing their own solutions.. or anything like that. All of these technologies have been added to the kernel tree by the people who maintain them.

Re:What about kqemu?

[SirTalon42]

KQemu hasn't been GPL for very long, so it hasn't so there hasn't been very many people looking at it for very long. Also the developer of KQemu probably hasn't even ASKED it to be merged (or maybe he has, I don't read the LKML). It's also possible he submitted it to be merged and theres a few things they want him to work out first (Xen has been trying to be merged for a LONG time now).

Re:What about kqemu?

[smchris]

Actually, adding the complexity of all this virtualization into the kernel is a little scary to me.

I just upgraded my wife on Debian Sarge testing with a Win4Lin 2.4-27 kernel. Web designer who demanded PhotoShop, Illustrator, Flash, and IE but the Win98 Win4Lin base and apps were getting dated. Did a dist upgrade to Etch Stable and installed XP Pro on QEMU with kqemu. Was good. The kqemu performance was very adequate, net, samba share and got her apps working.

But then I upgraded to Etch testing. First, qemu itself had changed -- blog rumor says he changed the base virtualized hardware and my XP didn't start. Downgraded and pinned that. Couldn't use the 2.6-21 kernel with nvidia. I understand it has been a function issue within the paravirtualization. Could work around that an easy way -- in which case qemu wouldn't work again. Or I understand a person can hack some source. Also, I let a bochsbios upgrade slip in. Another package that killed qemu that I had to downgrade and pin.

So I ended up with sort of a Debian Etch testing but with several packages pinned back to stable that works fine but it has been a bit of a minefield. And it is my understanding that the virtualization built into the kernel is at the heart of the problem.

Re:What about kqemu?

[baadger]

KQEMU (and indeed QEMU) releases are relatively infrequent, the latest pair of releases was in February.

QEMU has so far been a solid foundation for a handful of other FOSS virtualisation solutions, KVM use a modified Qemu (Does anyone know if KVM support is going upstream into the next QEMU release?) and Virtualbox incorporated QEMU to establish full system emulation on top of their own hypervisor. If anything I'd like to see the Virtualbox OSE kernel module merged, which imo is far superior to raw QEMU+KQEMU at the moment.

QEMU is far from dead though, there seems to be quite a bit of activity regarding patches on the qemu-devel mailing list.

Re:What about kqemu?

[deimtee]

You run linux on your wife ? Cool!!

As a testament to my lack of knowledge...

[jimktrains]

...why should virtualization technology be incorporated into the kernel, and not kept outside, as a "3rd" party app? Shouldn't the kernel be essentially a library and some low level support (multi-tasking, handle certain interrupts, that sort of stuff)? I've never really even considered bash, or even ls as part of the kernel. Am I just really mistaken, or is the word kernel used more broadly than that?

Re:As a testament to my lack of knowledge...

[Anonymous Coward]

Because THAT IS NOT HOW LINUX IS DESIGNED.

Go read up on Monolithic vs Microkernel design and then you'll know.

FYI The vm stuff in the kernel is miniscule, if you want linux to do something that'll actually save you space, have them strip out all the broken arches from the kernel tree, or better yet everything other than x86, so we don't end up with a 300 meg source tree only 100 or less of which we actually use!

Re:As a testament to my lack of knowledge...

[QuantumG]

The hardware support for virtualization is in the kernel.

Just like the hardware support for webcams is in the kernel.

Re:As a testament to my lack of knowledge...

[jimktrains]

See, now, that would make sense. So it's not the entire virtualization programs, just hardware hooks and drivers, basically? Meaning that there still needs to be a separate program to take care of actuality running things and what not?

Re:As a testament to my lack of knowledge...

[jimktrains]

My question has nothing to do with monolithic vs microkernel. My question has to do with why are these programs being including with the kernel.

Only one hardware branch of the kernel gets compiled, and yes, I know I can choose not to compile many things into the kernel, and do so whenever I compile it.

See the post below you for an answer that was helpful. Compare that to your answer, and figure out how to answer a question instead of trying to belittle someone.

Re:As a testament to my lack of knowledge...

[QuantumG]

Yes. Thing is, bare x86 metal can do virtualization.. you just gotta be creative. There's a lot of ways to do it, utilizing different parts of the hardware. So there's some solutions that work great for some things and some solutions that work great for others. It's like having two drivers for the same bit of hardware and choosing which one to use based on how you're using the device.

Then there's para-virtualization.. modifying the kernel of the guest OS so you don't even need anything in the kernel. Well, sometimes kernel support can help para-virtualization :)

Re:As a testament to my lack of knowledge...

[jimktrains]

Thanks, that all makes sense now.

Re:As a testament to my lack of knowledge...

[Chris Snook]

You're thinking of a microkernel. Most modern operating systems have a monolithic virtual memory model, in that a large number of system services run in the kernel memory space, but they use dynamic linking to achieve a degree of modularity. That said, the Linux kernel internal API is fairly fluid, so any code that runs in kernelspace has to be maintained quite regularly to keep up with the changes. Merging your code into the main tree makes this much easier.

Bash and ls are still userspace. All of these virtualization implementations have userspace tools that control them, but they need some help in kernelspace to set up the virtual memory mappings, and that's the code that's been merged.

Re:As a testament to my lack of knowledge...

[TheRaven64]

Anything that uses VT or AMD-V needs to run in ring-0 (when not in VMX mode, at kernel boot, and then in a special VMX mode that can be entered from there), and so needs to be in the kernel. Adding virtualisation support is basically (warning: oversimplification) adding support for a new type of process. These have to be able to do things like respond to interrupts raised in the process (for example, to handle system calls). Solutions like VMWare get around this be re-writing interrupt instructions to jumps to a special handler, but paravirtualisation requires a trampoline in the kernel that reflects interrupts back to the virtualised kernel.

The situation for Xen is slightly different, since Xen is a hypervisor. While KVM lets you run virtual machines under Linux, a Xen system has Linux running atop Xen. The complication comes from the minimalist nature of Xen. It delegates a lot of functions, such as initialising new virtual machines and providing drivers for hardware, to the first VM to be started (currently, this has to be Linux, NetBSD, or Solaris), known as domain 0. This means that this guest needs some significant modifications to support interaction with Xen (as well as being modified to run atop a hypervisor, rather than on the bare metal).

Wireless card??? WTF?

[brunes69]

I don't have any idea what you mean by "VMWare Player doesn't work with my wireless card". VMWare doesn't know ANYTHING about your underlying networking hardware. All it uses is the IP stack.

Re:Wireless card??? WTF?

[stef0x77]

VMWare by default bridges your network interface into the VM. Wireless drivers have such poor support for network bridging that this almost never works. It especially doesn't work with WPA or any such.

If you NAT your VM network traffic, then things work (well sorta, with all the nastiness that NAT comes with).

Re:Wireless card??? WTF?

[physicsnick]

I have an Atheros chipset wireless card which requires binary drivers to work. It does not work with VMware.

This is the Ubuntu bug report (note the length and number of duplicates) which actually breaks apt on installation, but it's not Ubuntu specific; you can't configure it manually with this wireless card either. The only solution is to disable networking virtualization, which means I can't even have VMware use my wired connection unless I disable the wireless card entirely or physically remove it from my system.

Was I seriously modded down for that? Mods, what the hell?

Re:Wireless card??? WTF?

[brunes69]

Just use NAT instead of bridging in your vmware config

Re:Wireless card??? WTF?

[physicsnick]

Obviously I tried that, as have the dozens of other people who encountered this problem. It doesn't work either.

Re:Wireless card??? WTF?

[Anonymous Coward]

This is why Linux should get on board with systems like OpenBSD, where they've reverse engineered wireless drivers for such proprietary cards. Wireless on BSDs is so much less of a pain thanks to this kind of work. It just works! Imagine that! And you can configure it with ifconfig, instead of inventing an unnecessary program (iwconfig currently; previous versions of Linux had a whole mess of other tools... ugh.)

Re:Wireless card??? WTF?

[Anonymous Coward]

Which version of Player are you ranting about? The bug that you linked to talks about an old version of Player (1.0.1-4.) Have you tried using the new version? I'm using Player (2.0 build 45731,) and it works great on Ubuntu 6.10. (And yes, I do have an Atheros wifi card. It's a Toshiba Satellite Pro A100 and I'm using ndiswrapper around the Windows driver.)

I hate to say it, but I think your "problem" is really a lack of you RT'ing the FM.....

Re:legality

[JamesRose]

You do realise microsoft claims patent infringements about linux all the time, and as such, even running linux without any virtualisation software would be illegal in microsoft's eyes.

does anyone actually use a VM....

[LingNoi]

... on the desktop? I only have Ubuntu installed and I don't see why a VM is such a massive feature these days? Have I missed something amazing that I can do on these or is it simply for a cool "hey I can run a desktop on a desktop!"

I understand that application compatibility is a big deal but Linux has a zillion apps already.

I just don't get all the marketing surrounding it.

Re:does anyone actually use a VM....

[billbaggins]

It's a big help for software developers needing to support multiple platforms/versions. At my company we provide support for the past 5 or 6 versions of our software, so I have a VM for each version that I fire up when I need to check something or patch a bug. Lots easier than dealing with multiple physical machines.

Re:does anyone actually use a VM....

[Anonymous Coward]

I use them for development and testing, no way do I want to litter my main install with software I only want to try out. There are further potential advantages such as deployment, portability and COW images. Stick to /home/ling/beastiality-pics/ if you don't see any advantages.

Re:does anyone actually use a VM....

[ShieldW0lf]

If you still need access to Adobe products like Photoshop for print production, like my GF does, there's nothing available on Linux that will do the job.

Linux + Xen + W2K lets her leave the windows desktop and still use these tools.

Pretty straightforward.

Yes.

Re:does anyone actually use a VM....

[drinkypoo]

I only have Ubuntu installed and I don't see why a VM is such a massive feature these days?

I have vmware installed and use it on a regular basis. Here's what for:

• Windows emulation. Wine is great and good, but it doesn't run everything. Sometimes I want to run some Windows software not supported by Wine. Mostly this takes the form of various (non-3d) games. I have Windows 98 and Windows 2000 VMs. Also cellphone hacking can pretty much only be done under Windows (at least for Motorola) - it's possible to flash only like one format of software image under Linux, whereas I can handle about five on Windows.
• Linux testing. I can test a LiveCD in a virtual machine without even burning the ISO.
• Appliances. Excellent for testing/development. I made a Debian LAMP appliance, for example, with everything I needed to run Drupal. When you don't need it, it's turned off, and preventing potential security risks and avoiding using any resources (not than an Apache site not getting hits is using a lot of resources.)

I've talked about it elsewhere, but I also envision a system using UML (or now, lguest) to separate servers (or groups thereof) away from the main system to reduce security risks. It would let you use selinux with a fairly restrictive policy on your controlling system, and if one of the subsystems is compromised it could easily be discarded and rebuilt.

Re:does anyone actually use a VM....

[moco]

Why have virtualization on the desktop? Good question. Here are a few answers I can think of:

  * Software development, as it has been mentioned in this thread.
  * Testing "stuff", a sandbox to play in before messing with the system, "stuff" being other operating systems, applications, services.
  * security, the secure vm and the unsecure vm running on the same physical hardware.
  * Corporate environments, the user's machine is a vm that can be ran on any of the physical PCs on the network.

Re:does anyone actually use a VM....

[Chris Snook]

I do. It's delightfully convenient if you do development work, because you can run tests in something a lot more realistic than a chroot build directory. It's particularly nice if you're doing kernel work. For cluster testing, the only alternative involves $20k worth of hardware.

So, Joe user may not need this, but it's a major feature for the people who work on improving the Linux kernel. That alone justifies including these features.

Re:does anyone actually use a VM....

[raftpeople]

I understand the desire for VM's so this question really isn't about that, but why can't you have 5 or 6 versions of your software on 1 box? When I worked for an ERP company it was pretty common for our servers to have multiple versions of the software.

Re:does anyone actually use a VM....

[flyingfsck]

There are cases where virtualization is very handy. Instead of creating a 'do everything' server where the configuration of Apache, PHP and Perl can interfere with Amavisd-new, CLamAV and SpamAssassin - you upgrade the one and break the other - one can keep things separated by running two virtual servers - one for Apache and the other for email. Then you can even stop and start the one without influencing the other. Once you start playing with VMs you can build up a collection of basic installed systems for rapid deployment: Keep DVDs with images for Windows 2003 Server, Windows XP Pro, various versions of Linux, all pre-installed (registered/activated) and ready to go. You can even have complete solutions for email, groupware, web servers, CRM and what not as virtual appliances to deploy at a moment's notice. Some systems are extremely hard to install - this way you can do that once and keep a copy. As business requirements grow, you can transfer a virtual machine very easily from an old slow server to a new faster one, without having to reconfigure anything. Lastly, once you have your laptop set up and everything working properly, you can use a VM to experiment with something - even run multiple servers and clients at the same time, without fscking up your real system. I suggest that you go and try VMware Server. It is easy to set up and will make you discover a whole new world of productivity. Once comfortable with VMware, you can try the other systems which are still somewhat more difficult to get going.

Re:does anyone actually use a VM....

[init100]

Virtual machines can be very useful in the server room, especially those with advanced features such as live migration. A guest can be migrated between different hosts without shutting the guest down. It was previously quite common to run each service on its own computer, and still is, but now we can run these services on its own virtual machines instead, making each system fairly clean and only used for one service.

Several different virtual machines can run on one host, making it possible to aggregate many low-load services onto one host, removing the necessity to have a large amount of nearly idle computers. If you find that you need more computing power, just buy a new computer and configure it as a host. Then you can move guests there or create new guests for new services.

Re:does anyone actually use a VM....

[stevey]

Yes I use virtualisation a lot.

I run Debian's unstable branch (sid), but I still need to build packages against Sarge & Etch. With virtualisation I can have both those systems running with X, etc, installed upon them. Whilst you can get by with chroot()s having a fully running system, with networking etc, is a much nicer experience.

It is also very handy for testing deployments and system upgrades. Prior to upgrading several remote servers from Sarge -> Etch I was able to fully test this on a virtual install, to make sure that I would run into no problems.

There are many more reasons why you might want to use this kind of system on an average desktop, but these are enough for me.

Re:does anyone actually use a VM....

[Anonymous Coward]

You don't need a VM for such security. Plan 9 has a preaty neat solution for this problem. It is simple and actually solves many more problems at the same time. Per PID filesystems. If you come to the mailing list or #plan9 I am sure someone would be grateful to explain how it works and it solves more than this problem.

-- Ron

Solved it three different ways!

[symbolset]

Please review Robert Frost: "The Road Not Taken".

yes but

[Anonymous Coward]

Does it run on linux?

Does it run on linux?

Very fishy and intriguing

[jkrise]

"The happy theme of today's kvm is the significant performance improvements, brought to you by a growing team of developers. I've clocked kbuild at within 25% of native. This release also introduces support for 32-bit Windows Vista. " I can't understand why the Linux kernel development team had 'Windows Vista support' as one of the items on their agenda at all. Virtualisation as I understand it, is basically an abstraction of the hardware that is performed in software. Should not all operating systems be designed to work with standard instruction sets, interrupts, registers and memory?

Why should it be the job of a particular kernel or it's VM component to satisfy specific requirements of a specific version of another kernel (the Vista kernel?). Besides, how exactly did these developers get access to the Vista kernel specs? Should it not be the other way round - i.e. for closed-source Vista to be compatible and optimised for the open-source Linux kernel?

That Linus chose the GPL as a matter of convenience was well known, his antipathy to the FSF is also well chronicled; but this aligning to the interests of specific closed-source kernels from Microsoft is a dangerous new development.

Re:Very fishy and intriguing

[QuantumG]

The people who work on this stuff really wouldn't call themselves kernel developers, but ok, whatever. Associating any of the VM stuff with Linus is even more retarded.. what they do in their own modules is none of his fault or concern. Anyway, some people want to run Vista in a VM on Linux. These VM solutions don't try to virtualize every nook and cranny of the x86 hardware. Vista uses the system level x86 hardware in a slightly different way to XP. As such, it takes some changes to make Vista work.

Should it not be the other way round - i.e. for closed-source Vista to be compatible and optimised for the open-source Linux kernel? Yeeaaaaaahhhh.. ok. Whatever dude.

Re:Very fishy and intriguing

[Anonymous Coward]

The Vista support is in the user-space portion of KVM.

There's the KVM kernel part, which does only those parts which can not be done in user space. A user space application can not directly use the VM hardware, nor could it do virtualisation without that hardware. That's why VMWare requires a kernel module. Basically, the kernel component provides an abstracted interface to the hardware virtualisation capabilities, and exposes that interface to user space, hopefully in a secure way that prevents user-space programs from breaking the host OS. It provides little more than pure CPU virtualization.

The KVM user-space part is basically a modified version of Qemu. It's responsibility is to emulate all the rest of the hardware, from interrupt controllers, up through the motherboard, busses, peripherals, and I/O devices like video, sound, networking, keyboards, mice, access to the host's USB devices, disk drives, and all the other stuff. It doesn't do CPU emulation, because it uses the kernel component to virtualize the host CPU. Even the BIOS is contained here.

Vista wasn't working before because the user-space component didn't implement everything it needed to run properly. At the very least, there were some issues with ACPI, and probably a few other bits. This is now fixed, and Vista now works correctly.

Re:Very fishy and intriguing

[jkrise]

The people who work on this stuff really wouldn't call themselves kernel developers, but ok, whatever. Associating any of the VM stuff with Linus is even more retarded.. what they do in their own modules is none of his fault or concern. I find the announcement about these VMs is from Linus himself. Besides, it is Linus who decides which components get into the main kernel tree, so he is answerable for any decisions made.

Anyway, some people want to run Vista in a VM on Linux. These VM solutions don't try to virtualize every nook and cranny of the x86 hardware. Vista uses the system level x86 hardware in a slightly different way to XP. As such, it takes some changes to make Vista work. If Vista has any idiosyncracies, it should be the job of the overpaid, bloated development team in Redmond to iron out the kinks and make it standards-compliant. Why should it be a concern of the Linux kernel development team? Besides, how did these developers gain access to quirky behaviour of Vista?

Re:Very fishy and intriguing

[QuantumG]

I find the announcement about these VMs is from Linus himself. Besides, it is Linus who decides which components get into the main kernel tree, so he is answerable for any decisions made. Linus puts whatever he wants into his tree, yes. His tree is the defacto "main" kernel tree, yes.

If Vista has any idiosyncracies, it should be the job of the overpaid, bloated development team in Redmond to iron out the kinks and make it standards-compliant. Why should it be a concern of the Linux kernel development team? Besides, how did these developers gain access to quirky behaviour of Vista? What standards are you talking about exactly? The Intel x86 hardware documentation? I can assure you they are writing their code to those "standards" otherwise their code wouldn't work..

If anything the virtualization guys are the ones who are not implementing the "standards".. as not everything that will run on an x86 processor will run the same way under virtualization. That's simply because it's a lot of effort just to get the most common usage of x86 to virtualize.

Re:Very fishy and intriguing

[timeOday]

Virtualisation as I understand it, is basically an abstraction of the hardware that is performed in software. Should not all operating systems be designed to work with standard instruction sets, interrupts, registers and memory?

Ideally. My impression is it doesn't actally work very well without some "cheating" (optimization). For instance VMWare works a lot better if you use the special VMWare video driver on the guest instead of sticking with generic VESA or whatever. Also some timing related issues, which result in problems like the mouse moving at weird speeds, require peeking behind the curtain.

Parallels is supposedly going to have 3d acceleration soon. My guess is here again this will require specific guest support and some sort of snooping or rewriting of guest programs.

One optimization I think would be great is sharing read-only memory between guests. This could save tons of memory when you're hosting N guests with the same software versions on them.

GPU support question

[JustNiz]

So do any of these solutions support 3D graphics (nvidia) hardware?
The only reason I currently have a windows partition at all is for gaming.

Being able to run Windows 3D games in a VM would allow me to move to a Linux-only box and also give me a nice way of:
* managing the way windows keeps grabbing diskspace
* remove the need to go through reinstalling/reactivating windows every 6 months or so
* limiting the damage Windows virusses can do
* limiting all the phone-home comms with Microsoft that windows keeps doing

Re:GPU support question

[QuantumG]

No. But if/when there is ever an open source nvidia kernel driver with 3d support that isn't completely broken and is integrated into the kernel, you might see some people take an interest in virtualizing it.

Probably the first thing they'll do is make it so X running in a virtual machine can share the same DRM (Direct Rendering Module) as X running on the host. Of course, that's not much good to a Windows guest.

Re:GPU support question

[EvilRyry]

So do any of these solutions support 3D graphics (nvidia) hardware?
The only reason I currently have a windows partition at all is for gaming.

I recently read an article on the progress of just this. It sounds pretty cool and the initial results are impressive. This combined with the DX->OpenGL Wine code, that I'm sure will be open sourced from the makers of parallels (just had a slashdot story on this), makes for an exciting future for providing hardware acceleration to guest applications.

More information: http://www.cs.toronto.edu/~andreslc/vmgl/

Re:GPU support question

[Chris Snook]

Not very well. Xen with PCI pass-through might work here, but that requires having a dedicated graphics card for each OS. 3D video generally involves some amount of writing directly from userspace to hardware, without any kernel interaction after initial setup. This is difficult to do right in all cases with virtualization, but they are working on it.

Just buy Cedega and be done with it.

Re:GPU support question

[halovaa]

VMWare had some limited support for this: http://www.vmware.com/support/ws5/doc/ws_vidsound_ d3d.html The last time I tried it, it was probably not up to the standards of Wine/Cedega, but it had potential for running games that require Windows' quirks. It was only Directx 8 though (or DirectX 9 that only actually uses DirectX 8 features).

Re:GPU support question

[baadger]

Parallel's patches to Wine have already been released, apparently they weren't very exciting...

Re:legality

[Anonymous Coward]

You do realise microsoft claims patent infringements about linux all the time, and as such, even running linux without any virtualisation software would be illegal in microsoft's eyes.

You do realize slashdotties claim all sorts of stupidity all the time, and as such, it's a crime the way some gullible chumps think it means something.

Gaming applications?

[Anonymous Coward]

My kids want a new computer to play some of their games on. But my wife and I need a new computer as well. However, our computing needs are quite minor. We mainly browse the Web, and send email. To keep our data safe, we only use Linux. But I also do some Mozilla development which often touches a number of source files, leading to fairly hefty spurts of C++ compilation. So I still need a powerful system, but only for several minutes now and then.

Using this virtualization technology, would it be possible to simultaneously run Windows XP and Linux on the same system, and offer maximal performance for each? Namely, would Windows XP still be able to have sufficient access to the actual video hardware for gaming purposes?

Re:Gaming applications?

[init100]

Namely, would Windows XP still be able to have sufficient access to the actual video hardware for gaming purposes?

AFAIK, Direct3D support is highly experimental in VMware, and I haven't heard of it being available in any of Xen or KVM (Lguest can only run Linux guests, so Direct3D support is a moot point). So the answer is probably no.

Try running your games under Wine instead. It would probably be a safer bet, but it isn't guaranteed to work especially not without hitches. I've read it has improved a lot since I tried it 4-5 years ago, but it isn't 100% complete yet.

Re:Gaming applications?

[Anonymous Coward]

Install Windows on the hardware, run VMware Server on Windows, run your Linux instances as VMs and access them remotely with a thin client, remote X, VNC, etc. VMware Server is free (beer free,) and perfect for what you need.

http://www.vmware.com/products/server/

Re:Gaming applications?

[Anonymous Coward]

He is better off dual booting Linux and XP. He doesn't need to be running XP all the time, so need to complicate things with virtualization, and slow down his Linux work.

Re:Gaming applications?

[Anonymous Coward]

His question was: "Using this virtualization technology, would it be possible to simultaneously run Windows XP and Linux on the same system, and offer maximal performance for each? Namely, would Windows XP still be able to have sufficient access to the actual video hardware for gaming purposes?"

Dual booting wouldn't allow for simultaneous access...

User Mode Linux

[Anonymous Coward]

I would like just to mention User Mode Linux (http://user-mode-linux.sourceforge.net/), that was included in the kernel mainstream a lot of time ago (much before KVM, I remember).

Clarification of these technologies

[GiMP]

Each of Xen, KVM, lguest, and UML can be considered virtualization products but they are all vastly different. Below I describe each of these products in relation to their inclusion to the Linux kernel.

Xen - the Linux kernel supports code allowing it to be run as a guest underneath the Xen kernel, all through software. Linux's support for Xen does not make Linux a virtualization platform, only a GUEST for the Xen kernel which sits at Ring-0. (though a "dom0" Linux system can interact intimately with the Xen kernel, it actually sits at Ring-1). I should note that the Xen kernel also supports hardware virtualized domains, though this is unrelated to the patches to Linux.

KVM - the Linux kernel supports virtualization of guests through hardware extensions, this requires supported hardware. Linux becomes the Ring-0 kernel.

lguest - (my understanding is) an unmodified Linux kernel can act as a hyper-supervisor through loading Linux kernels as modules. Linux sits as both Ring-0 (supervisor) and Ring-1 (guests). This is experimental with limited features and only supports Linux guests.

UML - the Linux kernel becomes a userspace program. This allows Linux to run as an executable application/program. With UML, Linux can be compiled for a Linux or Microsoft Windows target. The executing OS sits at Ring-0 and the UML program sits at Ring-1. This has the advantage of requiring no modifications to the host OS and is very portable (you could email an entire Linux system to a friend without requiring anything installed to their system), but the disadvantage of poor performance.

From a high-level, the products UML, Xen, and lguest are actually very similar in function. They act as architectures to which Linux can be compiled in order to make it a guest OS of another Ring-0 kernel. These architectures provide the targets of a kernel module (lguest), a userspace program (UML), or a xen-domU guest (Xen). On the other hand, KML is the only patch that is intended to add support to Linux to act as a Ring-0 kernel on behalf of guest systems -- and even then, KML can be viewed more as a hardware driver for the processor extensions.

Re:Clarification of these technologies

[_Knots]

Slight corrections:

The UML program sits at ring-3 on X86 machines: it's just a normal user program using the ptrace() mechanism and extensions [except when the host has been patched with SKAS, but even here it's just a "normal user program". Rumor has it that SKAS might eventually make it into mainline, but it's time in 'real soon now' is starting to rival Duke Nukem Forever's.]. Rings 1 and 2 are odd, rarely used (IIRC there's the current virtualization craze and OS/2 as notable consumers) features of the x86, derived from MULTICS. For processors with only two (user & supervisor) modes, identify ring 0 with supervisor mode and the other rings with user mode.

It is a little odd to say that Linux "becomes" the Ring-0 kernel under KVM. It was already running in ring 0.

Re:Clarification of these technologies

[Per Wigren]

Yes, they are all very different but at the same time quite similar from a user's perspective. All of them (unless I've missed something) more or less emulate a whole machine. This means you have to mess with disk images or dedicated drives/partitions/LVs, allocate a fixed amount of RAM to the guest, among other things.

Personally I like the approach of OpenVZ and VServer better. The main OS and the guests all share the same kernel, share the RAM and their root filesystems can be just subdirectories of the host's filesystem. When inside the virtual server you don't realize that though. You only see your own processes and everything works as if it was a dedicated server. You can run iptables, reboot and just about everything you could normally do in XEN/KVM/VMWare. Including live migration of virtual servers to other physical hosts. chroot on steroids.

I really hope OpenVZ and/or VServer will be merged at some point. VServer seem to keep up with current kernel releases so that wouldn't be too hard to merge I guess. OpenVZ usually have a lag of something like half a year.

Re:Clarification of these technologies

[radarsat1]

Hope this isn't too far off topic...

Xen - the Linux kernel supports code allowing it to be run as a guest underneath the Xen kernel, all through software. Linux's support for Xen does not make Linux a virtualization platform, only a GUEST for the Xen kernel which sits at Ring-0. (though a "dom0" Linux system can interact intimately with the Xen kernel, it actually sits at Ring-1). I should note that the Xen kernel also supports hardware virtualized domains, though this is unrelated to the patches to Linux.

I'm not too familiar with all these virtualization solutions, but this struck me as being somewhat reminiscent of how RTLinux worked for Linux 2.4. I haven't really needed hard RT since the preemptive scheduling in 2.6, but I was wondering if some of this virtualization stuff has been used to implement hard real-time?

(In RTLinux, the Linux kernel ran at a lower priority than your RT code --- this is distinctly different from the soft RT that can be achieved with a high-priority user-mode process in 2.6, despite that fact that you can basically achieve millisecond timing that way.)

Re:Clarification of these technologies

[Anonymous Coward]

KVM - the Linux kernel supports virtualization of guests through hardware extensions, this requires supported hardware. Linux becomes the Ring-0 kernel.

Well... Ring 0, 1, 2 and 3 exists since around 20 years. The introduction of the newer CPU supporting hardware virt is sometimes referred as "ring -1" (as in "minus one").

An hardware virtualized guest under KVM (or Xen full virt) will have itself access to all ring 0 instructions... (but not to the ones allowing hardware virt, so you can't run Linux full virt under Linux full virt etc. At least not on current proc).

My point is that this old "ring 0" thinggy from the 386 is kinda incomplete nowadays.

Re:Clarification of these technologies

[GiMP]

> It is a little odd to say that Linux "becomes" the Ring-0 kernel under KVM. It was already running in ring 0.

But important to distinguish from Xen, where Linux only runs underneath ring 1.

Re:Clarification of these technologies

[GiMP]

> Personally I like the approach of OpenVZ and VServer better.

They are yet different solutions for different problems. Obviously there is much overlap in all of the virtualization solutions, but these glorified chroot solutions are certainly a different breed.

The advantages for OpenVZ and VServer are exactly their disadvantages. There can be advantages to having dedicated disk and ram resources, rather than shared resources as you have with OpenVZ and VServer. Furthermore, you're tied to having a single kernel for all your guests.

I think that the "shared resources" model of OpenVZ and VServer are a little scary to enterprises as it can make planning and troubleshooting more difficult (avoiding over utilization, determining bottlenecks, etc). For web hosting, I know that many hosts love the shared resources -- more accounts! Obviously, though, this is only another way for the host to oversell. With Xen, there is no such thing as over selling, except perhaps on cpu and disk performance (but not on capacity)! This is why my hosting company only offers Xen accounts, we give customers the satisfaction that do not oversell capacity -- we couldn't if we wanted to!

Still, I think that there are definate uses for VServer and OpenVZ. Obviously, they will remain a staple of web hosting firms for some time, even with Xen creeping into the low-end of the market. However, I do see enterprises using it too, but more for security hardening than for server consolidation. On the other hand, I see Xen more useful for server consolidation than for security hardening. (Where by security hardening, I'm referring to separating apps between virtual machines.) I see both being useful for getting rid of the "one app per server" model that has hit some enterprises, depending on the apps in quesiton. For instance, I think a larger mail server would do better underneath Xen than it would underneath OpenVZ, but a BIND/Named server might do better underneath OpenVZ/VServer. My opinion for this comes from the capacity overhead of running a virtual machine in Xen is already higher -- this overhead increases when used for very-small guests (such as many Bind installations); on the other hand, Xen is very good for large guests, needing much ram and disk, such as one running a very active mail server.

One interesting thing to note is that there is a Debian package containing patches for both Xen and VServer so there is no reason that they cannot be used together.

but

[thatskinnyguy]

But will it run on... nevermind!

Re:legality

[throup]

Despite being modded down to -1, I think this needs treating as a legitimate question:

Isnt it illegal to run windows with this? Googled it n microsoft seems to think so.. MelNews Illegal? That depends on your definition of legal... different nations have different laws.
Breach of software license? Possibly... if I recall correctly, the EULA for Vista forbids running in a virtualised environment. I believe it is perfectly legitimate to run XP this way as long as the license key has been purchased legally and is not currently in use in another installation (obviously with the exception of multi-user licenses). For other versions of Windows, it depends on the EULA but I think Vista is the only one to forbid it.

yes!!

[SolusSD]

I think all the xen users out there will agree with me when i say "yes!!!!!!!!!!!". I'm actually quite impressed, given what is involved in maintaining xen in the kernel, that this happened as soon as it did.

Re:legality

[init100]

if I recall correctly, the EULA for Vista forbids running in a virtualised environment.

At least the Enterprise and Ultimate editions are okay per the EULA to run in a virtualized environment, but I'm not sure about the rest. I faintly recall Microsoft being opposed to using the same copy of the cheaper editions as both the host and the guest OS. It is possible that it is legal to run a separately purchased copy as a guest in a virtual machine.

IANAL though, and I don't have Vista, so I cannot check its EULA.

MODS ON CRACK

[Just Some Guy]

That was a perfectly legitimate question (and one that I'd have asked, too). Right now, most people install VMWare to run Windows on their Linux hosts. I'd be quite pleased to be able to run it using standard, Free, built-in functionality.

Re:MODS ON CRACK

[andy_t_roo]

The question remains why run both windows and linux? I run windows for ease of game playing (yes i know most things can be convinced to work under linux, but it is less time to maintain a windows install than it is to get them working). The biggest problem with virtualisation technology is that there is minimal 3d acceleration within the guest. This leads me to run windows native, for the game performance when needed, and have linux running in a virtual machine. As linux can also be compiled paravertualised (in the newer kernels, .21 and newer i think) this leads to linux-in-windows being more efficent in many situations than windows-in-linux, both from the functionality-of-guest, and game-playing points of view. If anyone knows otherwise, please let me know. Andy

Re:MODS ON CRACK

[Anonymous Coward]

this leads to linux-in-windows being more efficent in many situations than windows-in-linux, both from the functionality-of-guest, and game-playing points of view

This is true, but it doesn't follow from that that Linux-in-Windows is better than running the two separately. For example, users who find Linux desktop environments more efficient to use than the Windows shell would be disadvantaged by running Linux-in-Windows. Users who primarily require graphical Linux applications would also gain benefits from running an environment that had native X11 with full hardware acceleration, rather than a Windows-based X server running clients on virtualised Linux.

I used to do pretty much what you say - I had Linux running in coLinux on a Windows desktop. But these days I run Linux natively, using Wine or virtualised Windows for cases where I need Windows software and performance isn't critical, and I dual-boot into Windows for 3D action games, and that seems to be the best setup for my personal usage patterns. (I play 3D games infrequently enough that rebooting to get to them isn't a major issue. Hard-core gamers will obviously not be satisfied with this!)

Re:MODS ON CRACK

[Just Some Guy]

The question remains why run both windows and linux? I run windows for ease of game playing (yes i know most things can be convinced to work under linux, but it is less time to maintain a windows install than it is to get them working).

My shortest answer: Quickbooks Pro doesn't run under Wine yet, and that's what my accountant uses. For business reasons, it's much easier and cheaper to run Windows in a VM for that lone non-graphic-intensive application than to try to get my accountant to accept something different.

Although Windows is mostly a game OS, it does still have a few important business apps.

Hardware requirements and microcode.

[Climate Shill]

Three choices, but none of them gives the optimal solution, which is unmodified guest OSs on processors lacking specific hardware support.

This is apparently hard or impossible on many i386 processors due to the difficulty of intercepting certain instructions. But since Linux updates the processor microcode at boot time, would it be possible to modify the processor microcode to change the way the offending instructions operate ?
(just invalidating them would proably be enough)

3D acceleration please

[Skeith]

I've slowed worked Linux into everything else I do on a computer, now just let me never have to switch out of it and I'll be set.

lolcat!

[Anonymous Coward]

i'm in ur kernel
wasting ur ram

Re:Why? (What is "IDE")

[Anonymous Coward]

It's not a bus or port or other connection type, it designates a drive type.

IDE = Integrated Drive Electronics = smart drives (as opposed to dumb drives needing controller chips on the motherboard end of the cable) = everything since the early Winchesters and whatnot. So it's been a fairly meaningless term for some time now...

Then when SATA arrived some bright bulb equated (Parallel) "ATA" with "IDE" and we still suffer from the confusion :-(

Anybody running OS/2 Warp 4 under linux?

[knorthern knight]

I keep an old PII clunker kicking around to run Galactic Civilizations V2.5, an OS/2-only game. I'd really like to get rid of it, but keep OS/2 for the game. With QEMU and Virtualbox, I've occasionally managed to "install OS/2" but the VM crashes when trying to do much more than merely bring up the OS/2 desktop. I'd be interested in any working solutions. TIA.

How freakin' big is the Linux kernel now?

[NateTech]

Anyone attempting to compile a full Linux kernel with every conceivable feature that doesn't clash with another turned on, non-moduluar, will be able to measure the build time in months...

Unless they're running a virtualized cluster of machines! :-)

Re:legality

[G Morgan]

In any definition it is more legal than the way most people run Windows to begin with.

VMs are handy.

[WoTG]

I've got a couple VMWare VMs on my laptop. I'm running Windows XP (boo hiss, I know...).

The first is Linux system, Ubuntu, for running tools that aren't available for Windows and web site testing. Technically the OSS DBs, PHP, and Apache have Windows builds, but the pathing and other differences are bothersome, especially if the ultimate goal is to run the site on a Linux server.

The second is a bare bones Windows partition where I test suspect websites or software. Clients ask me to quickly evaluate a program that they want to buy. I'm not about to use my computer as a guinea pig (any more) now that VMWare is free and the hardware requirements are low enough (by todays standards).