QEMU Accelerator Achieves Near-Native Performance

cool stuff by Prophetic_Truth · 2005-02-19 05:38 · Score: 3, Interesting

but how fast is "near native"? Some would consider WINE near native in certain aspects, in short, WHERES THE BEEF?

--
time is a perception of a being's consciousness
time is your 6th sense, the wierd ones are 7+

Re:cool stuff by Anonymous Coward · 2005-02-19 05:47 · Score: 2, Informative

WINE *does* run at native speed becuase WINE Is Not an Emulator (see where the name comes from now?).

Instead of emulating, WINE runs the programs using Windows libraries, either the original DLLs or rewritten versions of them.
Re:cool stuff by tepples · 2005-02-19 05:52 · Score: 3, Informative

RTFA. "Near native" in this case means about half as fast as native execution on the bare hardware. This should be enough to run the vertical market apps that a lot of business IT departments female-dog about when clinging to their Windows environment.
Re:cool stuff by bcmm · 2005-02-19 06:16 · Score: 4, Informative

Wine Is Not an Emulator.

Wine is a re-implementation of some of the Win32 API for Linux. Unless combined with an actual hardware emulator, it only works on x86. It just runs x86 binaries on an x86 processor like Windows does, but providing some commonly used Windows API calls.

In short, running programs using Wine is "native", and, programs that work at all are often actually faster with Wine than with Windows for the same reasons that anything runs faster on Linux.

--
# cat /dev/mem | strings | grep -i llama
Damn, my RAM is full of llamas.
Re:cool stuff by Guspaz · 2005-02-19 06:25 · Score: 2, Interesting

The QEMU accelerator can also only run x86 binaries on x86 processors. It only works on x86.

This is because it's not emulating, it's virtualizing, and tries to get as much code as possible to run native on the host CPU.
Re:cool stuff by iamacat · 2005-02-19 06:48 · Score: 1

I suspect the reason qemu accelerator runs at native speed is that it's not an emulator either. It probably just runs user-mode code directly as a Linux process and then catches a signal when a privileged instruction is attempted and emulates OS code.
Re:cool stuff by MadChicken · 2005-02-19 07:16 · Score: 1

oops, I saw another comment that was saying "Qemu", then I saw this one and thought I would avoid being redundant. My mistake.

Your turn.

--
SYS 64738 NO CARRIER
Re:cool stuff by snorklewacker · 2005-02-19 07:31 · Score: 1

> It probably just runs user-mode code directly as a Linux process and then catches a signal when a privileged instruction is attempted and emulates OS code.

That's how all supervisor-mode virtualizers work, by only trapping certain instructions. The only emulator out of the whole bunch is bochs.

CoLinux is my favorite solution for Linux on Windows. It's basically a port of User Mode Linux to Windows. Fast as native, just no native devices (network is done through tunnelling or a pcap hack). KDE still runs fine over X to a cygwin X server.

--
I am no longer wasting my time with slashdot
Re:cool stuff by Money+for+Nothin' · 2005-02-19 07:48 · Score: 2, Insightful

In short, running programs using Wine is "native", and, programs that work at all are often actually faster with Wine than with Windows for the same reasons that anything runs faster on Linux.

Faster than on Windows? That has not been my experience at all.

Maybe 80-90% as fast (just a seat-of-the-pants guesstimate)...

--

Is Capitalism Good for the Poor?
Re:cool stuff by steeviant · 2005-02-19 08:10 · Score: 1

While it's true that QEMU runs on Mac OS X and various other OSes and Architectures than x86, the article points to a page where the author talks about QEMU Accelerator. QEMU Accelerator is described as being comparable to Virtual PC in the way it works, with most of the code being run natively without translation/emulation, so unless you're running Mac OS X on an x86 CPU, there is no way that the QEMU Accelerator module could possibly work.

So, in conclusion...

It is true, and it always has been that QEMU Accellerator can only run x86 binaries on x86 processors, and indeed, it does only work on x86 processors.

Other than that, I agree with everything you said.
Re:cool stuff by MrResistor · 2005-02-19 08:36 · Score: 1

Depends on how you do it. If you're lazy (like me) and are using Cedega on KDE, then yeah, there's going to be a speed hit.

But, some people like to create a special runlevel, with X but no WM/DE, for gaming. That's much faster.

--
Under capitalism man exploits man. Under communism it's the other way around.
Re:cool stuff by snorklewacker · 2005-02-19 08:57 · Score: 1

> The only emulator out of the whole bunch is bochs.

Replying to myself ... obviously VirtualPC is an emulator as is PearPC. Looks like qemu started as a bochs clone, became more of a virtualizer on x86, but remained an emulator on other architectures. I never knew it even ran on other architectures.

--
I am no longer wasting my time with slashdot
Re:cool stuff by reub2000 · 2005-02-19 09:00 · Score: 4, Informative

Qemu is usually an emulator, unless you use the kqemu module which makes it a virutualizer like VMware.
Re:cool stuff by Guspaz · 2005-02-19 09:18 · Score: 1

Errm, while you are right that I indeed said qemu accelerator, and x86 only product, perhaps your response is a bit harsh...
Re:cool stuff by dinivin · 2005-02-19 09:58 · Score: 1

From American Heritage Dictionary:

1. To strive to equal or excel, especially through imitation: an older pupil whose accomplishments an style I emulated.
2. To compete with successfully; approach or attain equality with.
3. Computer Science. To imitate the function of (another system), as by modifications to hardware or software that allow the imitating system to accept the same data, execute the same programs, and achieve the same results as the imitated system.

Sorry, but that third definition clearly applies. WINE most certainly is an emulator.

Dinivin
Re:cool stuff by Guspaz · 2005-02-19 10:56 · Score: 1

For pete's sake, RTFA! The QEMU accelerator (KQEMU) is a virtualization solution that ONLY runs on x86 and ONLY can run x86 apps.

I suggest you read this:

http://fabrice.bellard.free.fr/qemu/qemu-accel.h tm l

Perhaps you shouldn't talk if you don't know what you're talking about?
Re:cool stuff by jrockway · 2005-02-19 16:18 · Score: 1

> bash-2.05b# cat /dev/mem | strings | grep -i llama
> Shit, my RAM is full of llamas...

Perhaps because the argument 'llama' to grep is... stored in memory?

--
My other car is first.
Re:cool stuff by Anonymous Coward · 2005-02-19 16:41 · Score: 1, Informative

WINE Is Not a Emulator. Period. And I find it rather strange that people here are STILL so misinformed about it.

WINE is an incomplete implementation of the Win32 API for X/Unix based systems. It should run almost at native speed, but there is a very small performance hit because of the translation of the WIN32 API calls to POSIX ones. While it is barely noticeable on old machines (like mine :-() it should run nearly at native speed on newer machines.

Geez!
Re:cool stuff by encia · 2005-02-19 18:18 · Score: 1

Refer to http://www4.tomshardware.com/howto/20020531/window s_gaming-05.html for the opposite view i.e. WineX vs Windows 2K.
Re:cool stuff by bcmm · 2005-02-19 23:59 · Score: 1

Yes, but in my case there were some less explicable ones too. Like "Llamaboy". Or "CHEESE STEAK JIMMYS" when grepping for cheese...

--
# cat /dev/mem | strings | grep -i llama
Damn, my RAM is full of llamas.
Re:cool stuff by puetzk · 2005-02-20 05:40 · Score: 1

I've definitely had the experience of WINE being faster than 'native' WinXP performance. But I wasn't doing gaming or other driver-intensive things, I was running some FPGA layout and simulation software. Running my testsuite in the simulator didn't work in windows; I let it go for over an hour and then it crashed. In wine it took ~5minutes :-)

Now, in all fairness, I'll attribute that to at least 2 factors:
- in Linux, my machine has 2 processors, and both were getting used. XP home only sees one, and I've never cared enough to pay additional microsoft tax to solve that.
- The simulation run needed significantly more memory than I have; WinXP basically thrashed itself to death and then eventually the process crashed on an allocation failure. I have linux setup with swap on the second hard drive (and thus separate from the data files).

So windows was playing with double handicap, because I don't usually do real work in windows :-)

However, I do think the real deciding factor was the kernel's VM layer; linux had an advantage in having the swap on a separate spindle, but it also was doing a better job at managing the working set, and wasn't degenerating into random thrashing.

--
The Matrix is going down for reboot now! Stopping reality: OK. The system is halted.
Re:cool stuff by puetzk · 2005-02-20 05:45 · Score: 1

Do you run mythtv? I finally figured out that that's where some of the stranger ones were coming from - they're out of show descriptions on the weirder cable tv channels :-)

--
The Matrix is going down for reboot now! Stopping reality: OK. The system is halted.
Re:cool stuff by bcmm · 2005-02-21 03:12 · Score: 1

No, but mldonkey was running...
Some of the more bizzare usernames of other p2p users maybe...

--
# cat /dev/mem | strings | grep -i llama
Damn, my RAM is full of llamas.
Re:cool stuff by tabrnaker · 2005-02-21 09:30 · Score: 1

I'm not sure what lowing is, but i assume then you mean that we should all be insensitive clods then?

Processor emulator by Saeed+al-Sahaf · 2005-02-19 05:40 · Score: 3, Funny

open source processor emulator?

So I can emulate my P4 on my P4? Swift!

--
"Who are in control, they are not in control of anything - they don't even control themselves!" - Glen Beck

Re:Processor emulator by puiahappy · 2005-02-19 05:43 · Score: 1

Yes or even better i can emulate a P4 on my AMD AthlonXp ? ;)

--
Think like a hacker, act like a hacker, but never become a hacker !
Re:Processor emulator by Lisandro · 2005-02-19 05:52 · Score: 4, Funny

You can. Just put the heatsink in the oven for 15 minutes at 400C before use.
Re:Processor emulator by grotgrot · 2005-02-19 09:33 · Score: 1

So I can emulate my P4 on my P4? Swift!
More accurately you can emulate the full instruction set of the P4 including all priviledge levels using only unpriviledged P4 instructions. Due to the design and implementation of the x86 that is actually a very big achievement. It can also emulate various bits of hardware which is less of a challenge, but still hard work.

Re:CherryOS by Saeed+al-Sahaf · 2005-02-19 05:42 · Score: 1

Didn't CherryOS turn out to be a fraud?

--
"Who are in control, they are not in control of anything - they don't even control themselves!" - Glen Beck

Re:if it really works... by greypilgrim · 2005-02-19 05:43 · Score: 2, Insightful

CherryOS has been released. It's called PearPC http://pearpc.sourceforge.net/

Doesn't work on PPC or SPARC by mrbrown1602 · 2005-02-19 05:43 · Score: 4, Informative

It should be noted that the Accelerator only works with x86 computers. It will not work on SPARC or PPC.

Re:Doesn't work on PPC or SPARC by hoppy · 2005-02-19 05:51 · Score: 1

Yes but x86 is dirty cheap. So it make sens to emulate PPC on x86 to run Mac OSX.

To be efficient it would have to support Altivec emulation. Clone are back!
Re:Doesn't work on PPC or SPARC by damiam · 2005-02-19 05:57 · Score: 1

More precisely, the accelerator only works when emulating x86 using an x86. So it's not gonna help you with running OSX. You're likely to have much better luck with PearPC.

--
It's hard to be religious when certain people are never incinerated by bolts of lightning.
Re:Doesn't work on PPC or SPARC by defy+god · 2005-02-19 06:10 · Score: 1

apparently, they have a Mac OS X version in the works:

# User interface polishing and porting for Linux, Windows and Mac OS X.
# Full MAC OS X support as guest OS.
you can download a test copy for windows and mac os x here: http://www.freeoszoo.org/download.php

--
hackers of the world unite!
Re:Doesn't work on PPC or SPARC by Bruce+Perens · 2005-02-19 06:43 · Score: 2, Interesting

That tells us how it works, doesn't it? It must set up traps so that things that need to be emulated are caught, and then directly execute the code that would otherwise be emulated. This requires some systems programming smarts and a lot of knowledge of processor internals, but even if the module is never Open Sourced, it's possible for some third party to produce an Open Source equivalent.
Bruce

--
Bruce Perens.
Re:Doesn't work on PPC or SPARC by mrbrown1602 · 2005-02-19 07:00 · Score: 1

Note, I said the Accelerator. QEMU works just fine on OS X... but the accelerator only works on x86.
Re:Doesn't work on PPC or SPARC by fm6 · 2005-02-19 07:13 · Score: 1

Wrong.
Re:Doesn't work on PPC or SPARC by Queer+Boy · 2005-02-19 07:21 · Score: 2, Insightful

Oh, so it's Virtual PC for PCs. Doesn't Microsoft already make that?

--
Not since Marie-Antoinette played milkmaid has looking simple and honest been so fake and complicated.
Re:Doesn't work on PPC or SPARC by timeOday · 2005-02-19 08:08 · Score: 3, Interesting

That tells us how it works, doesn't it? It must set up traps so that things that need to be emulated are caught, and then directly execute the code that would otherwise be emulated.
If so, it isn't an emulator and "near native on some code" performance isn't a big deal - about the same as VMWare.
Hopefully it really is an emulator, and the accelerator simply hasn't been written for other platforms yet.
If the claim were true, and could run even graphical apps decently fast, Intel should throw a couple million bucks at this guy for the fastest possible X86 on IA64 emulation - they need it.
Re:Doesn't work on PPC or SPARC by yanestra · 2005-02-19 08:09 · Score: 1, Informative

It should be noted that the Accelerator only works with x86 computers. It will not work on SPARC or PPC.
Sorry, but that's wrong. QEMU can emulate SPARC on x86 like it can emulate x86 on SPARC. Only certain speed-ups won't work everywhere.
Re:Doesn't work on PPC or SPARC by WillerZ · 2005-02-19 08:20 · Score: 2, Insightful

He was talking about the accelerator, not QEMU in its entirety.

--
I guess today is a passable day to die.
Re:Doesn't work on PPC or SPARC by timeOday · 2005-02-19 08:32 · Score: 1

Well, I just checked qemu's website and you're right - it does just run x86 application code directly unstead of emulating it.
Re:Doesn't work on PPC or SPARC by fm6 · 2005-02-19 08:44 · Score: 1

I stand corrected. Though I have to point out that Bruce's first point (how it works) is redundant. And his second point is kind of questionable. The difficulty of making the IA-32 do this sort of thing is precisely why VMWare has no real competitors.
Re:Doesn't work on PPC or SPARC by Shadowlore · 2005-02-19 13:43 · Score: 1

Then why is there an installer for OSX?

--
My Suburban burns less gasoline than your Prius.
Re:Doesn't work on PPC or SPARC by sketerpot · 2005-02-19 15:13 · Score: 1

It doesn't run on Linux, but QEMU does.
Re:Doesn't work on PPC or SPARC by reverius · 2005-02-19 15:36 · Score: 2, Insightful

the "accelerator", not qemu!

the "accelerator" refers to the accelerator module, kqemu, which can be found here: http://fabrice.bellard.free.fr/qemu/qemu-accel.htm l

it is a -module- for qemu.
Re:Doesn't work on PPC or SPARC by incabulos · 2005-02-19 23:27 · Score: 1

But the base project can emulate sparc/sparc64 and ppc, thats no small feat, even if they dont run terribly fast. For the older cpus like the 32 bit SPARCs, an emulated sparc environment might be almost as fast as a hardware sparc system, given the power of the current crop of x86 cpus.

Supported target and source processors

I have to admit, the thought of running MacOSX/PPC and Solaris10/Sparc64 VMs simultaneously on a Linux/x86 box makes me go weak at the knees. I wonder if Alpha support for Target CPU is in the works..
Re:Doesn't work on PPC or SPARC by Bert64 · 2005-02-20 03:34 · Score: 1

Alpha would be very hard to emulate, atleast to get it right.. I imagine the PALCode emulation would be difficult and slow... but using an alpha to emulate other procs is quite easy since the cpu was actually designed to do that..

--
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Re:Doesn't work on PPC or SPARC by GoRK · 2005-02-20 05:19 · Score: 1

Indeed.

IIRC at one point FX!32 was benchmarking better on Alpha under emulation than on the fastest X86 chip (This was when you could get something like 200MHz alphas while the fastest x86 chip was a 486DX2/66 or 80).. Even at a pretty achievable 50-60% speed after emulation overhead they were pulling ahead.
Re:Doesn't work on PPC or SPARC by Bert64 · 2005-02-20 07:25 · Score: 1

FX!32 would emulate roughly a p3 running at the same clockrate as the alpha on which it's running, so the alpha still had the advantage until x86 started overtaking it for clockrate, which must have been 1999 or 2000 afaik.. Alpha was the first chip to reach 1ghz, but pretty much stopped there..

--
http://spamdecoy.net - free throwaway anonymous email - avoid spam!

This is great! by Husgaard · 2005-02-19 05:44 · Score: 5, Funny

I am really looking forward to emulating an Opteron at near native speed on my good old 386sx processor...

Re:This is great! by gstoddart · 2005-02-19 07:12 · Score: 1

I am really looking forward to emulating an Opteron at near native speed on my good old 386sx processor...

Keep looking. =)

--
Lost at C:>. Found at C.
Re:This is great! by jonadab · 2005-02-19 09:09 · Score: 2, Interesting

> I am really looking forward to emulating an Opteron at near native speed
> on my good old 386sx processor...

Actually, this is very similar to one of the features that the original
RAIF-POOL implementation boasted. It worked by using your internet connection
to co-opt available cycles that would otherwise go unused on other computers
on the internet. However, the software stuck in beta and was never officially
released, due to some minor process control glitches that were never fully
worked out, and then the Pentium III was released, and the prices on Celeron
processors dropped, and nobody seemed to need any extra processing power any
more, so the project was just dropped. There are still a couple of copies of
the beta floating around on the net, I think, but they're hard to find.

--
Cut that out, or I will ship you to Norilsk in a box.

Money? by cronius · 2005-02-19 05:44 · Score: 4, Insightful

Open Sourcing the QEMU Accelerator Technology ?
As a supporter of open source, the author accepts to open source the QEMU Accelerator Technology provided a company invests enough money to support the project and to recompense the author from the potential loss of revenue. Interested companies can look at the roadmap and make suggestions to the author.

If it actually achieves near-native performance right now, how much better can it get? And since it's already gratis, would anyone want to pay for one that achives actual native performance?

I don't think there's much money up for grabs here, to be honest. But that depends on how good it really is right now.

--
Life is Reality

Re:Money? by damiam · 2005-02-19 06:01 · Score: 1

Look at VMWare and then look at QEMU. QEMU's nice, but it's not quite there yet; Windows XP isn't supported and 2000 is still a bit glitchy. There's still a lot of room for QEMU to improve support, eliminate bugs, and add ease-of-use features.

--
It's hard to be religious when certain people are never incinerated by bolts of lightning.
Re:Money? by vinsci · 2005-02-19 06:38 · Score: 2, Informative

Windows XP isn't supported and 2000 is still a bit glitchy.
Actually, plenty of poeple are running Windows XP on top of QEmu. Why wouldn't they?

--

Trusted Computing FAQ | Free Dawit Isaak!
Re:Money? by damiam · 2005-02-19 06:49 · Score: 1

Reading the official QEMU page, they say that some people have reported success with it. That's a far cry from actually supporting it. I'm not saying QEMU's not a great program, just that it has a bit of work to do before it becomes as reliable as VMWare.

--
It's hard to be religious when certain people are never incinerated by bolts of lightning.
Re:Money? by nadadogg · 2005-02-19 07:22 · Score: 1

And since it's already gratis, would anyone want to pay for one that achives actual native performance?

The same reason anyone pays for open-source stuff. Support and ongoing development for specialized needs.

--
i use linux and windows oh god how can i have an opinion
Re:Money? by vinsci · 2005-02-19 07:23 · Score: 1

Well. I started running XP on QEmu in September, last year. No QEmu related problems. I'd say that page is outdated, as is some other QEMu pages (speed comparison from 2003, for example).

--

Trusted Computing FAQ | Free Dawit Isaak!
Re:Money? by timeOday · 2005-02-19 08:14 · Score: 1

As a current VMWare user, I would be very interested if you could compare VMWare to QEmu.
VMWare is a very good product, but here are my complaints:
1) being closed source, it lags behind the kernel releases sometimes.
2) it seems to run in kernel mode all the time, so isn't a very polite multitasker.
3) the cost prevents me from using it at home.
That said, I'm sometimes surprised at things VMWare can do, like drive windows-only USB devices from a linux host. And the set of virtual devices it provides, such as ethernet interfaces and sound support, is impressive too.
Re:Money? by kronchev · 2005-02-19 08:16 · Score: 1

If it ever achieves 99% full speed (I know thats asking a lot) I will pay for a copy, even if it is free.

If it can ever emulate OS X at 80%, I will pay for a copy. Not because I like OS X, but because that would be an incredible feat of programming.

I do want to see this guys code though, it's probably a work of art by now.
Re:Money? by mollymoo · 2005-02-19 12:35 · Score: 1

Windows XP isn't supported
Alright, you're an idiot. It must be Windows PX that I'm running under QEMU right now then.

These Nike trainers I'm wearing don't oficially support anal insertion, but that won't stop me kicking you up the ass for being such a jerk.

--
Chernobyl 'not a wildlife haven' - BBC News
Re:Money? by adam_j_bradley · 2005-02-19 13:03 · Score: 1

I agree. I will only contribute when it exceeds native speeds. A factor of 2 sounds nice.

--
Come and help me pay off my mortgage - small donations preferable! http://www.paymymortgage.com.au
Re:Money? by SupremeTaco · 2005-02-19 16:03 · Score: 1

Yes, you're right. Windows 2000 and XP have lots of room for improvement in support, and eliminating bugs.

Oh wait . . . .

--
You have a constitutionally protected right to be wrong, and I the right to ignore you.

Re:if it really works... by boeserjavamann · 2005-02-19 05:44 · Score: 1

:-) who knows?

Re:CherryOS by greypilgrim · 2005-02-19 05:45 · Score: 2, Informative

Yup. Just a frontend for PearPC http://pearpc.sourceforge.net/ .

Only Reasonable by kiljoy001 · 2005-02-19 05:45 · Score: 2, Insightful

It's only reasonable, considering the time and effort that this guy put into the program... wasn't ibm pushing out money to linux ? Stuff like this should be funded.

Re:Only Reasonable by mmu_man · 2005-02-19 21:56 · Score: 1

Also take into account that Fabrice is also the mighty author of FFMpeg, which codecs are used by all the media players in use in Linux, and on other platforms (VLC, mplayer, Xine, ...)
So we do owe him more than a beer I guess :)

Don't want Real Windows no way no how. by wiggling · 2005-02-19 05:45 · Score: 1, Offtopic

... run Windows (or another OS) on a Linux machine at near native speeds without buying a commercial emulator.

Only by running Windows® That could be convenient for the near term, but my goal is to not send another penny to Redmond.

Re:Don't want Real Windows no way no how. by aetherspoon · 2005-02-19 08:32 · Score: 1

You haven't built enough computers then. It is really easy to build your own machine much much cheaper than what you can get from Dell - as long as you don't buy from brick retail stores.

--
--- Ãther SPOON!

Re:if it really works... by Anonymous Coward · 2005-02-19 05:45 · Score: 1, Insightful

First off, QEMU is x86.

Second, CherryOS is a ripoff of PearPC. The reason for the CherryOS delay is most likely that it's violating the licensing of PearPC and the developer is worried about litigation.

Oh yeah! by Saint+Aardvark · 2005-02-19 05:46 · Score: 5, Interesting

I've been trying out VMware's latest beta at work (journal entry on that in an hour or so) to run W2K on Linux, in order to try some patch automation software. It's great, but I'd love it if there was a Free program that'd do this.

I dearly hope the accelerator gets GPL'd. Between sysadmin work (reverting to a snapshot ROX) and, just maybe, being able to move W2K people to Linux (there's only a handful of applications we need [damn you, Texas Instruments! Where's your Linux version of Code Composer?], and remote admin is just soooooooo much better with a Unix), I'd be very happy if a) this thing works as well as it's supposed to, and b) if there was some sort of tip jar I could kick in a few bucks to (like with Blender, I believe), and get it released when there was enough money.

Incidentally, I tried installing W2K on qemu w/o the accelerator. When I left work on Friday, it was finishing up the second stage of installation; it was slow as molasses, but seemed to be working. This seems to contradict the note re: disk full during install problem noted on the support page. It's always possible I just haven't hit it yet, but does anyone else have any experience with W2K and qemu?

--
Carousel is a lie!

Re:Oh yeah! by stevey · 2005-02-19 06:52 · Score: 3, Informative

I installed Win2k on a Debian installation without the accelerator - and wrote about it here Running Microsoft Windows inside Debian : Qemu.

I found the process of installation took a couple of hours, which didn't feel really excessive. Once installed it's pretty good to use, for small things.

Right now I'm using qemu to practise the upgrade of a busy server from Woody to Sarge (whenever its ready!). All I can say is that the program rocks, and many thanks to its authors.
Re:Oh yeah! by mml · 2005-02-19 07:41 · Score: 1

> there's only a handful of applications we need
> [damn you, Texas Instruments! Where's your Linux
> version of Code Composer?

The code composer CD from December 2004 includes
statically linked x86 linux binaries.

I use it every day and it works just fine. Admittedly, it doesn't include the GUI tools (simulator, debugger), but the compiler is there, and that's all I care about.

Matt
Re:Oh yeah! by Spy+Hunter · 2005-02-19 07:56 · Score: 2, Interesting

It has always been Fabrice's intention to have the accelerator GPL'd; he uses the word "when" to describe the event rather than "if". He is just looking for a sponsorship first; hopefully some company will soon provide him with one (Lindows^H^H^H^Hspire perhaps?) If nobody does sponsor him within a year or so, I imagine that somebody else will write a different accelerator unless he releases his own; there is quite a bit of demand.

--
main(c,r){for(r=32;r;) printf(++c>31?c=!r--,"\n":c<r?" ":~c&r?" `":" #");}
Re:Oh yeah! by yope · 2005-02-19 08:40 · Score: 1

damn you, Texas Instruments! Where's your Linux version of Code Composer?

CodeComposer compiler binaries run just fine with wine. The IDE sucks anyway IMHO, so you can live without windows here. Just use you favorite linux-IDE, gnu-make and some wine-wrapper scripts around the windows command-line executables.

Just to stay on-topic, I have had the same mentioned problem installing Win2k on QEMU, the virtual HD really gets filled up with strange log files of errors that seem to occur during the hardware detection stage. You can either make a really big virtual HD image, or freeze the process half way and manage to erase those files somehow, then continue the installation.
I think that's also the reason it is so slow during install. Once it is installed, speed is actually quite usable.
Re:Oh yeah! by Saint+Aardvark · 2005-02-19 08:54 · Score: 1

The code composer CD from December 2004 includes statically linked x86 linux binaries.
Really! Man, didn't know that. We've been using the binaries with Wine, so that's good; it's more the GUI that I'm after for Linux. (Yeah yeah yeah Emacs, but some of the engineers really want the GUI.)
Thanks for the tip about the binaries, though. That's bound to be easier than trying to get the Solaris binaries (which I assumed without checking included the GUI) to work on FreeBSD or Linux...

--
Carousel is a lie!
Re:Oh yeah! by Saint+Aardvark · 2005-02-19 08:57 · Score: 1

As I mentioned above, we're using the compilers w/Wine w/o problems, just as you say. However, the IDE is quite liked by some of the engineers (not being one, I have no opinion on the matter), and that's what I'd like to get going on Linux.
Thanks for the tip on QEMU; if I go back on Monday and it's crapped out, I'll give that a try.

--
Carousel is a lie!
Re:Oh yeah! by beuges · 2005-02-19 10:09 · Score: 1

I really dont mean to sound like a troll, but I have a question:

what difference does it make if the accelerator is GPL'd or not? Its free, and it appears to work. I'm a programmer by professions, and i like to think i'm at least better than average if not good, but I doubt there are that many people who would benefit from having the source to something as low-level as an emulator available (apart from the obvious additional contributors, but i'm talking in the general open-source "i have the source so i can hack it if i want to" sense). Why can't you use and support it in its closed-source state?
Re:Oh yeah! by DasIst · 2005-02-19 12:39 · Score: 1

what difference does it make if the accelerator is GPL'd or not?

A couple years down the line, if he became bored with the project, it might not work with modern distros. Without the source available to recompile, it'd suddenly become unusable. It wouldn't be the first time I've seen that happen with closed source applications in Linux.
Re:Oh yeah! by Saint+Aardvark · 2005-02-19 16:30 · Score: 1

In addition to the two other excellent replies: you're right -- I probably wouldn't be able to make full use (or any use, beyond compiling) of the source code. I'm not a programmer, let alone a good enough programmer to understand or improve what's going on.
But if it's open-source, then lots of people who are that good can look at it and talk about how well it works, what its problems might be, how they can be improved. It's not guaranteed, of course, but it's a lot more likely than if it's closed-source. There will be (the potential for) more informed opinions about the software, and that will benefit me.
It's not that I can't, or won't, support it in its closed-source state. It's just that it would be so much more useful to me, though in indirect ways, if it were open-source.

--
Carousel is a lie!
Re:Oh yeah! by Bert64 · 2005-02-20 04:49 · Score: 1

Well, it works as a kernel module on current 2.4.x and 2.6.x linux kernels, and thats it..
If there are many changes to 2.7.x etc, it will stop working, similarly it won't work on freebsd or netbsd, or whatever other os people may want to run it on..

--
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Re:Oh yeah! by Bert64 · 2005-02-20 07:21 · Score: 1

I installed win2k and encountered the disk full during install problem, it seems to create thousands of logfiles.. if you delete them, it will continue just fine.. I never found it all that slow tho, without the accelerator.. I will try it with the accelerator when i get to work tomorrow.

--
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Re:Oh yeah! by Saint+Aardvark · 2005-02-22 07:50 · Score: 1

How did you manage to delete them? I tried getting at the files with a Knoppix disk (version 3.6) and only succeeded in borking the image badly enough that Windows refused to even reinstall on it.

--
Carousel is a lie!
Re:Oh yeah! by Saint+Aardvark · 2005-02-22 07:54 · Score: 1

So how did you manage to delete those logfiles? And how big a disk image did you have to make? Right now I'm filling up a 2GB image w/the files you talk about. I tried deleting them by mounting the image w/a Knoppix 3.6 CD, but only succeeded in borking the image so badly that Windows wouldn't even reinstall on it.

--
Carousel is a lie!
Re:Oh yeah! by Bert64 · 2005-02-22 22:06 · Score: 1

Well, knoppix doesn't play well with propriatory filesystems like NTFS.. i actually followed the guide on the qemu site, and hit reset when it borked.. then it resumed the install and worked, and once it had booted up for the first time i deleted all the files.

--
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Re:Oh yeah! by yope · 2005-02-23 01:45 · Score: 1

Weeellll... I don't remeber the details anymore, since it is quite some time back, and I tried so many things back then....

I think, that in the end I just installed it on a 5 or 6Gb image, and deleted the files afterwards, but there must be a way to do it in the middle of the process....

Of course, if you delete files on a frozen system, you need to be aware that you are deleting files on a running system without it knowing, so this will be quite tricky. Parts of the disk-image might still be in RAM, and are about to be written back later.
You probably have to start off installing on a FAT partition because it is the easiest to tweak or fix by hand, and use a disk-editor to just unlink the FAT chains that are in use by those files, and hope the FAT is not entirely cached in RAM at that moment. Anyway that's a guess.... you'll have to try.

Re:CherryOS by NetNifty · 2005-02-19 05:46 · Score: 1

Last I heard it was suspected of actually being PearPC under a different name (compiled executables compared and found that many variable names matched), but the author(s) of Cherryos denied it - no idea if it was proven one way or another for certain though.

--
Linux Wireless Hardware in the UK

qemu DOES work by supersuckers · 2005-02-19 05:49 · Score: 5, Informative

A lot of people seem skeptical so far. I've been using qemu to run windows98 under linux for close to a year now. This was before this "accelerator" It was definitely usable. I needed windows for an application from my job, and this let me use it without rebooting. I installed it using the following instructions: http://forums.gentoo.org/viewtopic-t-186001-highli ght-qemu+howto.html

Re:qemu DOES work by advocate_one · 2005-02-19 06:11 · Score: 1

nice one Cyril... excellent link... now my old ms-windows 98SE license is no longer fit for just using as a beermat... and it means the company I work for will have fewer excuses for not changing over to Linux... the power users will still be able to run their windows only apps

Mind you, the thing I've now got to attempt is to get w2k pro and OSX running simultaneously on a Linux box...

--
Donald 'Duck' Dunn: We had a band powerful enough to turn goat piss into gasoline.
Re:qemu DOES work by neumayr · 2005-02-19 13:45 · Score: 1

Well, running an 1998 OS on a modern PC of course doesn't prove this emulators 'Near-Native Performance'.

--
Truth arises more readily from error than from confusion. -Francis Bacon

Re:CherryOS by duncan · 2005-02-19 05:49 · Score: 1

Check thier web site. They had a press release on October 12 last year and they have announced that it will be availible Q1 2005.

What more do you want? It's all right there on thier home page at www.cherryos.com.

Is there anything else I can do for you?

Dual licence ? by Alain+Williams · 2005-02-19 05:51 · Score: 4, Insightful

Why not release under a dual licence as mysql have done ... they seem to rake in money while giving the product away. Yes: most people will have it for free, but some will want to pay (to include in a proprietary bundle/...) - the extra market/awareness that open sourcing it will bring will mean that he will get a slice of a larger cake.

Re:Dual licence ? by fabu10u$ · 2005-02-19 05:59 · Score: 3, Insightful

Especially because with enough performance this is a good candidate for embedding in a commercial product, i.e. if your code is really tied to a specific architecture you could "port" by wrapping QEMU around it.

--
They say the mind is the first thing to ... uh, what's that saying again?
Re:Dual licence ? by gnuman99 · 2005-02-19 07:05 · Score: 1

So why not use the emulator (100% free), and if you need more speed, just buy the accelerator module, if you with? http://fabrice.bellard.free.fr/qemu/license.html

Re:if only they could get this to run faster by fabu10u$ · 2005-02-19 05:55 · Score: 5, Funny

From your link:

Hear God Anywhere! - Get BiblePlayer Deluxe for $29 +s/h

If you need an iPod to hear God, maybe you should start looking for a different faith...

--
They say the mind is the first thing to ... uh, what's that saying again?

Qemu - information by Richard_J_N · 2005-02-19 05:58 · Score: 5, Informative

Here's a bit more about Qemu. I've now used it quite a bit, and am impressed by it!

Qemu can essentially run any guest OS (Linux,Windows,MacOS,BSD,menuetOS...) under any other. This normally incurs approximately a 4x slowdown, which is pretty good. I've personally installed XP on Linux this way. It's also a great way to test out the latest knoppix.

The KQEmu (accelerator) is a kernel module which allows near-native speeds, if both the guest and host architecture are x86.

Qemu is Free (speech and beer); the accelerator is free (beer), but not, at least for now, open source.

In the specific case of needing to run windows apps on Linux, we have now several options:

API emulation: Wine
PC emulation: Qemu (free)/VMWare (expensive)
VNC: (and just move the display)

Re:Qemu - information by Kevin+DeGraaf · 2005-02-19 06:16 · Score: 1

In the specific case of needing to run windows apps on Linux, we have now several options:

API emulation: Wine
PC emulation: Qemu (free)/VMWare (expensive)
VNC: (and just move the display)

Also, rdesktop, a client for RFB (remote framebuffer) servers such as Windows Terminal Services.

--
We have more to fear from the bungling of the incompetent than from the machinations of the wicked.
Re:Qemu - information by dmaxwell · 2005-02-19 06:39 · Score: 1

If you're remastering Knoppix, it is handy to see if your new ISO will do what you will expect without having to burn and reboot.
Re:Qemu - information by newr00tic · 2005-02-19 07:15 · Score: 1

####

Does this mean you can run MAC-OSX in QEMU even on x86/intel processors? ..sounds freaking great!!

.

--
A horse can't be sick, you know, even if he wants to.
Re:Qemu - information by Richard_J_N · 2005-02-19 07:28 · Score: 1

I principle, yes. In practice, not quite yet... See: http://fabrice.bellard.free.fr/qemu/ossupport.html
Re:Qemu - information by Cthefuture · 2005-02-19 07:42 · Score: 1

PearPC is what you want.

Development was going really awesome on that thing then all of a sudden it stopped. Possibly due to the death of Stefan Weyergraf.

--
The ratio of people to cake is too big
Re:Qemu - information by timeOday · 2005-02-19 08:21 · Score: 1

Can you use it to drive windows only USB devices from a Linux host? (i.e. set up a windows guest and share the printer, then mout it under Linux with Samba.) This is my favorite capability of VMWare, unfortunately it is too expensive for home use.
Re:Qemu - information by Richard_J_N · 2005-02-19 11:08 · Score: 1

Sadly, not yet. But it would be a killer feature if it were added. I should think it's probably not *that* hard to do...
Re:Qemu - information by Bert64 · 2005-02-20 04:56 · Score: 1

VNC is horrendously slow, i dont know why anyone would want to use it.. windows supports rdesktop, which is MUCH faster and X11 supports remote displays natively

--
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Re:Qemu - information by Richard_J_N · 2005-02-20 05:30 · Score: 1

VNC is actually quite fast - if you have the correcto configuration (look carefully at the options for encoding, and in Windows, turn off any animations). It's also cross platform. I find tightvnc very useful to have remote access to Win98 (old versions of windows don't support rdesktop). x11vnc is wonderful on Linux, since you can remotely view your own destkop, rather than starting a new xsession. Of course, now there's NX (nomachine) to add in.

Theoretically by no_choice · 2005-02-19 05:59 · Score: 2, Insightful

This means you could theoretically run Windows on a Linux machine at near native speeds

Theoretically, does it work or not?

iPC? by Doc+Ruby · 2005-02-19 06:02 · Score: 4, Interesting

Can you cut & paste between the Windows::apps and Linux::apps? Can each OS instance get its own IP# for IPC?

This app could offer a nice technique for "embedded" Windows: Run Linux, and QEMU::Windows with the Windows "screen" hidden or suppressed. Run vncserver on the Windows instance, and a vncviewer on the Linux desktop. Run a watchdog app that pings Windows and its apps, restarting them when they freeze. Put the host in the closet, and never hear surf music again.

--

--
make install -not war

Re:iPC? by Lehk228 · 2005-02-19 10:00 · Score: 2, Interesting

another way to have a cross system clipboard would be using the network adaptor, both the real and the guest OS would need a service running and they would communicate clipboard data back and forth, but on ly to specific IP's for security reasons.

--
Snowden and Manning are heroes.

Re:I'm getting pretty skeptical of these things by caseih · 2005-02-19 06:05 · Score: 5, Informative

The qemu claims are accurate. I am currently using this module with qemu and find that I do get 60-70% native speed (just as he advertises) and it is only going to get better. mind you this is x86 on x86. I/O performance (just like on vmware) is still a pig.

Look. Just go download it and try it. Don't post stupid pointless comments about how skeptical you. Don't know how that rated insightful.

Re:if it really works... by jericho4.0 · 2005-02-19 06:11 · Score: 1

No, CherryOS was a GPL violation of PearPC. The idiot simply repackaged PearPC and called it CherryOS.

--
"A language that doesn't affect the way you think about programming, is not worth knowing" - Alan Perlis

Re:CherryOS by jericho4.0 · 2005-02-19 06:14 · Score: 1

Yes, it was proven. The matching variable names proved it, in fact. The "author"s denials were; "but we wrote similar programs, of course the variable names are the same". It doesn't work like that.

--
"A language that doesn't affect the way you think about programming, is not worth knowing" - Alan Perlis

Boy, it's a screamer! by Just+Some+Guy · 2005-02-19 06:14 · Score: 4, Funny

I like the speedup table:

Emulator
Typical application slow down ratio compared to native

QEMU
5 to 10

QEMU + QEMU accelerator
1 to 2

I read those "ratios" as, well, ratios: "Wow, QEMU has a 5:10 slowdown, while the accelerator only has a 1:2 slowdown! I should write one with a .1:.2 slowdown! Oops, done."

--
Dewey, what part of this looks like authorities should be involved?

Re:Boy, it's a screamer! by Karma+Farmer · 2005-02-19 06:51 · Score: 1

The latter makes no sense as you were saying, which is why you need to read it correctly.

Actually, 5 to 10 is a ratio. Between 5 and 10 is not a ratio.

I think what you mean is that you need to read it incorrectly.
Re:Boy, it's a screamer! by jci · 2005-02-19 06:55 · Score: 1

I think its supposed to be interpreted as "between a 5 to 10x slowdown" which turns into 1:5 to 1:10 (which makes much more sense),

not "a ratio of 5:10"

but oh well, still teh funnay...
Re:Boy, it's a screamer! by Yotsuya · 2005-02-19 07:16 · Score: 1

Of course, that's 5:1 to 10:1 depending. And with the accelerator it's in between 2:1 and 1:1.

--
Claude Angers
Re:Boy, it's a screamer! by Karma+Farmer · 2005-02-19 07:42 · Score: 1

Well, the author cleary says "ratio," but as you point out, he didn't mean ratio. Basically, the author uses the word incorrectly. If you read what he writes, it's gibberish. If you read what you think he was tring to say, it makes sense.

Now, I'll argue that "reading what's actually written" is reading correctly. You'll argue that "reading what you think the author was trying to say" is reading correctly.

To-mae-toe/Toh-mah-ta, etc.

Opening the source by Sweetshark · 2005-02-19 06:15 · Score: 1

If it actually achieves near-native performance right now, how much better can it get?
It is at 1:4 to 1:5 speeds without the accelerator. With the accelerator the ratio is claimed to be between 1:1 and 1:2.
I don't think there's much money up for grabs here, to be honest. But that depends on how good it really is right now.
Depends. If he sells the accelerator to a company and releases the source a few month later under the GPL it might be interesting. A bit like the wine/winex stuff started. If the company manages this in a honest and clever way (unlike transgaming) it might be interesting for the linux community and the company.
Basically, I think Fabrice Bellard just shouted: "I am willing to double-licence this cool stuff.". qemu was already a good alterative to VMware without the accelerator. With a 3D video card driver for a windows guest OS (closed source by the company) it will be a killerapp.

Re:Opening the source by TheRaven64 · 2005-02-19 07:24 · Score: 1

I think what he's saying is `I want to work on this full time, but I can't afford to. If someone wants to pay me to, then that would be good. Oh, and I want to keep it open source'. I would not be surprised if IBM decided to offer him a job (or a stipend), since a good quality x86 emulator for PowerPC would be of huge benefit to them if they are intending to push PowerPC into the mainstream, since it would enable people to run legacy apps at a reasonable speed.

--
I am TheRaven on Soylent News

Call me silly by mr_zorg · 2005-02-19 06:17 · Score: 1

I understand that qemu is also an entire system emulator, but this post calls it a processor emulator. Does the accelerator also help with the general system emulation? Why would you need a processor emulator accelerator that only runs on x86 to emulate an x86 to run Windows? You're already ON an x86... Or is that all this accelerator really is? Some kind of native passthrough? Assuming I had the source and could compile it myself, would it benefit me in emulating a x86 on my PPC?

Re:Call me silly by abborren · 2005-02-19 06:53 · Score: 1

Except emulating hardware it will also emulate processor features such as paging, memory protection , interrupts etc. You don't want to/cannot use those features on the host processor so you have to emulate them.

--
><////>
Re:Call me silly by 1tsm3 · 2005-02-19 07:40 · Score: 1

Silly!

--
-ItsME
Re:Call me silly by mr_zorg · 2005-02-19 13:50 · Score: 1

I wonder how portable this code is? Would it fare as well on a PPC? The darwine project is already working on integrating QEMU for a completely non-MS emulation layer on OS X, it would be awesome if this code actually made it perform well enough to run those handful of Windows apps that you just can't get or substitute on Mac.

Re:If you need an iPod to hear God, maybe you shou by ShieldW0lf · 2005-02-19 06:17 · Score: 1

If you need an iPod to hear God, maybe you should start looking for a different faith...

All you need to do is hack the thing to play the tunes backwards!

--
-1 Uncomfortable Truth

Re:How do we know it;'s legal? by ctid · 2005-02-19 06:21 · Score: 1

I hope this will be treated with caution until it can be ascertained to be fully legitimate...

Pardon? This is Fabrice Bellard ! Show a little respect!

--
Reality is defined by the maddest person in the room

What is "Near Native"? by polyp2000 · 2005-02-19 06:25 · Score: 1

Near native to what ? To the clock rate of which processor? the physical one or the virtual one? I mean - Im sure you can emulate a Z80 at "Near Native" speeds without batting an eyelid but what if the Z80 in question was running at 20ghz ?
For example - having used VMWare in the past I have found that the performance of the virtual machine to be roughly half the speed of the host machine (performance might be better these days). Say my machine had a 2ghz cpu - I could say that I was getting "Near Native" speed of a 1ghz chip.

--
Electronic Music Made Using Linux http://soundcloud.com/polyp

Re:What is "Near Native"? by vinsci · 2005-02-19 07:15 · Score: 2, Informative

It's fast enough that booting Knoppix off the CD-ROM with the QEmu acceleator module loaded, much of the boot process is IO-bound rather than CPU-bound.
You can play Frozen-Bubble in QEmu and still have idle CPU on the host... It's fast enough! :-)
Now this also means that all of a sudden, the cost of migrating large organisations to GNU/Linux, while running that last application in a Windows install in an emulator got a _lot_ cheaper. No more VMWare needed. München, are you listening?

--

Trusted Computing FAQ | Free Dawit Isaak!
Re:What is "Near Native"? by ZeekWatson · 2005-02-19 11:42 · Score: 2, Informative

Near native to what ?
Native to the host processor running QEMU.
What else could it be? Perhaps you were thinking that you could emulate a beowulf cluster on a single P4?
Note that this accelerator only works on x86 emulating x86 so if you're thinking about z80 and G5 then you need to RTFA ...

fabrice's other projects by harlemjoe · 2005-02-19 06:29 · Score: 5, Interesting

afaik the only way to run wine on linux-ppc (WINE is not an emulator, and so is x86 dependent) is by using qemu...

also check out darwine... integrated qemu + wine under OS X so you can http://darwine.opendarwin.org/ click on windows apps and run them seamlessly in OS X

fabrice bellard is a processor emulating god imho

--
shooting is not too good for my enemies

Re:fabrice's other projects by bullitB · 2005-02-19 07:49 · Score: 1

He's also been behind the ffmpeg project for a long time. Most Linux users probably use a fair bit of his code when watching videos.
Re:fabrice's other projects by Spy+Hunter · 2005-02-19 08:25 · Score: 5, Informative

He is not merely a processor emulating god; he is a coding god in general. Look at his project page. QEmu is not even his most significant project! He the main force behind the FFMpeg project, which is the premier open-source library for all things video-related (including open-source encoders and/or decoders for nearly every video/audio codec known to man). You can thank him for much of the progress that MPlayer/Xine have made, especially on non-x86 systems.
In addition, he has implemented a complete C99 compiler, and a software modem (unfortunately incomplete), which is the hard part of making open-source WinModem drivers. Also, an emacs clone which also happens to have full Unicode support (including bidirectional editing), *and* a built-in HTML/CSS2 renderer with WYSIWYG editing.
And if that wasn't enough, he has won awards in the IOCCC twice. If that doesn't prove he's a true coding god, I don't know what would. He has done all of this in his free time, for no pay. I think it is safe to say that he *really* deserves a sponsorship.

--
main(c,r){for(r=32;r;) printf(++c>31?c=!r--,"\n":c<r?" ":~c&r?" `":" #");}
Re:fabrice's other projects by Wesley+Felter · 2005-02-19 08:55 · Score: 2, Informative

Don't forget his crazy hack of putting a C compiler in a bootloader and compiling the Linux kernel at boot time.
Re:fabrice's other projects by The+Wicked+Priest · 2005-02-19 11:40 · Score: 1

The first time I heard of him was back in '89 or so, with LZEXE, the executable compressor for DOS that spawned a host of imitators. (Microsoft's EXEPACK predated it, but that was a lot less useful; also less safe.)

--
Share and Enjoy: 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
Re:fabrice's other projects by Replicant7 · 2005-02-20 05:51 · Score: 1

Nowadays the main force behind FFmpeg is Michael Niedermayer, a coding god in his own right. Fabrice hasn't touched FFmpeg in ages.

Qemu Speed by nurb432 · 2005-02-19 06:36 · Score: 1

The normal QEMU is more then useable, so i would tend to believe his claims.

Its also limited to emulating a i386 on a i386, so again its quite plasuable. Ever run VMware? I bet they are using the same sort of tricks.

Too bad I cant test them, as its only for Linux, and im running BSD... Perhaps down the road...

--
---- Booth was a patriot ----

Install/Run Speeds by nurb432 · 2005-02-19 06:40 · Score: 1

Ive found that once it does finally get installed, its not bad at all.

But yes, *installation* is a pain.. I gave up on my first attempt to install W2K, thought it had died...

--
---- Booth was a patriot ----

is it legal by yupa · 2005-02-19 06:41 · Score: 2, Interesting

If you read http://people.redhat.com/arjanv/COPYING.modules, you will see that if you write a binary module from scratch it is considered as a derived work from linux kernel, so it should respect the GPL license and be open source.

I am a bit disapointed to see a guy like fabrice bellard which have contributed to lot's of famous open source projects (ffmpeg for exemple) to choose a such decission.

Re:is it legal by yupa · 2005-02-19 08:06 · Score: 1

Note that fabrice don't seem familiar whith license issue : see http://sourceforge.net/mailarchive/message.php?msg _id=6932290 where is tried to relicense lgpl work in under bsd license...
Re:is it legal by Jason+Hood · 2005-02-19 10:35 · Score: 1

Its his code, he wrote it and doesnt need to give anyone his source. Now if he sold the module that is a different story.

What he is looking for is someone to give him financial backing to continue work on it. Its kinda like a shareware program at the moment except you dont have to pay anything. If a company steps up and puts him on the payroll, he will release the source and get to dedicate more time to it, but for money.

I completely agree with his direction, he has a right to make money off his work and an obligation to the open source community. He _can_ have it both ways.

--
Are you intolerant of intolerant people?
Re:is it legal by truedfx · 2005-02-19 10:50 · Score: 1

That (arguably unlike the kernel module) is completely his own code (according to the copyright notice, anyway). He distributed it to others under the LGPL, but as the copyright owner he himself isn't bound by that.

Comment removed by account_deleted · 2005-02-19 06:42 · Score: 2, Informative

Comment removed based on user account deletion

Not exactly by ari_j · 2005-02-19 06:43 · Score: 1

It's even better than that. You can emulate my P4 on your P4.

ffmpeg by daserver · 2005-02-19 06:45 · Score: 1

This is Fabrice Bellard of ffmpeg fame. Now I know why he hasn't touched that ffmpeg in a while :)

Derived work? by David+McBride · 2005-02-19 06:55 · Score: 2, Interesting

From the description, I'd be concerned that releasing the Accelerator code under a non-free (as in speech) license would be incompatible with the linux kernel's GPL license as it could be argued that it is a derived work.

See also http://kerneltrap.org/node/1735.

In practice, it may be enough of a gray area that it won't be a problem -- although it may scare off any company wishing to invest in it.

Personally, I'm just getting sick and tired with the maintainability and reliability issues that binary modules usually incur..

Re:Derived work? by Erpo · 2005-02-19 13:03 · Score: 1

Personally, I'm just getting sick and tired with the maintainability and reliability issues that binary modules usually incur.

I'm not. I love the way binary modules work. You just double click the exe file you downloaded from the hardware manufacturer's web site and follow the prompts. Then you reboot and your hardware is up and running without any compiling or messing about with manuals and documentation. Then, several years down the road and a few service packs later, you can take the same exe and get your OS up and running again. Of course, downloading the latest driver updates is always a good idea and it's not much harder than using the old installer, but it's good to have the guarantee that the old driver will work if you need it.

In all seriousness, though, I have to agree with you in that the Linux binary modules situation is nearly intolerable. I really believe that having the source to the kernel and all the drivers I use would be better than not having them. But the fact that they really are, for the most part, available in source form has allowed a particularly frustrating idea to seep into the heads of the kernel developers, that being that a stable binary interface for drivers should not be maintained.

I agree with you. It's bad right now. But the module developers are not the only ones causing the problem.
Re:Derived work? by crulx · 2005-02-19 16:52 · Score: 1

Linus has said that he will never stabalize the kernel interface for binary drivers because that would encourage companies to release more binary drivers instead of GPL drivers. So if you feel "sick of it" get another opperating system because Linux won't ever have that feature as long as Linus maintains the kernel.
Re:Derived work? by Erpo · 2005-02-20 10:45 · Score: 1

Linus has said that he will never stabalize the kernel interface for binary drivers because that would encourage companies to release more binary drivers instead of GPL drivers.

This is a perfectly valid stance for him to take if his goal is to encourage companies to release driver source under the GPL instead of releasing binaries. It is probably working out very well for him.

However, binary drivers make for more broad hardware support, and a stable binary (and source!) interface solves all sorts of problems for end users.

So if you feel "sick of it" get another opperating system because Linux won't ever have that feature as long as Linus maintains the kernel.

Getting another operating system only solves part of the problem. I've been using windows and dos since I don't know how long, and dual booting linux since 1996 I think. To be perfectly honest, I really like linux. I like the fact that I can dig down into any component of the system if I want to, or that I can change anything I want or distribute it to other people freely. I like the wide variety of very capable software that's available for the platform at no cost.

When I'm indulging my geeky side, linux can be a lot of fun for those reasons, but when I'm trying to accomplish a task instead of play around it can get very frustrating very quickly. I think the most frustrating thing is that all the mundane system administration tasks I'm used to performing in windows (e.g. installing a third party video driver, setting up a printer, installing software) are so much more difficult in linux. It sucks. I deal with it. Switching away from linux is not something I _have_ to do a whole lot to get my work done.

The reason why this sort of thing bothers me so much is that there are a lot of people out there who can't or won't deal with the additional trouble of runing linux instead of windows on a home machine, which is troubling since it means more machines running a microsoft OS. The key ingredient in pulling off something scary like Palladium is market share. If I could wave a magic wand and make microsoft not evil, I wouldn't care anywhere near as much about how usable linux happens to be.

You forgot win4lin by 6800 · 2005-02-19 06:57 · Score: 1

And win4lin home edition 30$, worked great for me and now I have to try this Qemu with accelerator! Of course win4lin only runs win95/98/me under linux.

http://www.win4lin.com/

Re:You forgot win4lin by 6800 · 2005-02-19 08:53 · Score: 1

I wasn't running a custom kernel so I just used the 2.6 precompiled kernel they provided. It is a 16MB download so I guess they compiled in right many options. The win98SE I am running on a 750mhz P3 boots in 15 seconds.
I had to refer to there support knowledge base for several items, all were there, most would have been unneeded had I rtfm first :-).

Re:EMULATOR?? by sanermind · 2005-02-19 06:59 · Score: 5, Informative

The main bulk of QEMU (which is all open source BTW) is entirely an emulator. It uses just-in-time dynamic binary translation to convert from the guest architecture to the host architecture, using an internal intermediate representation. At this level it is altogether an emulator... you can run x86 on powerPC or sparc on x86, or x86 on x86, etc. But because this the overhead is not insignificant. There seems to be a 5 - 10 times slowdown when I run winXP under linux in an emulated environment.

What the new KQemu accelerator does, is replace some of the emulation [specifically, of user-space code not in ring 0] with direct VMware-style virtualization, where the code is being run natively and trapped by the monitor. Important to note with KQemu is that kernel code is still being wholly emulated. Virtualizing only user-space is so much easier... I believe that Plex86 gave up on trying to virtualize ring0 code on x86 a while back, because the x86 isn't well designed for this. Running winXP with the accelerator causes it to run pretty close to native, though. Where before under just emulation, trying to play a video with windows media player took over a minute for a frame to show up and otherwise froze, with KQemu I can actually play video relatively smoothly. This is a wonderous thing! Basically, you can get almost all of the functionality of VMware, but FOR FREE.

--

---
the pen is mightier than the sword, the sword is mightier than the court, the court is mightier than the pen.

Re:if it really works... by gabebear · 2005-02-19 06:59 · Score: 2, Insightful

Actually QEMU runs on x86, x86_64, PowerPC, Alpha, Sparc32, ARM, S390, Sparc64, ia64, and m68k to different degrees. The X86 emulation is the only one completely working, but he PowerPC emulation in QEMU is good enough to boot PowerPC Linux, hopefully soon you will be able to boot OSX on a X86 via QEMU, and hopefully the speed will be at least 1/10th.

Think of what you just said by iamacat · 2005-02-19 07:05 · Score: 1

How can be something you wrote from scratch be a derived work of something else? This kind of like SCO wanting control of the code written by IBM. Even if that's allowed by law, it's morally wrong to demand compliance. Yes, I know Linus is the author of this document, but I think he is doing his own "GNU/Linux" thing here and should change his mind.

Re:Think of what you just said by cronius · 2005-02-19 08:51 · Score: 1

It is a very interesting read, I never thought that could be a legal issue.

Linus' point is that if you write a kernel module you own the copyright to it and can do whatever you want with it. However, it is impossible to compile that code without it containing some kernel code (kernel header files etc). So once it's compiled it contains at least some (if very little) GPL code and thus must be GPL.

If that's true/correct legally is another matter, I've never heard about it. But if he's right, the term "binary-only modules" can't exist legally (with regards to the Linux kernel), and every binary-only module that do exist today are violating the GPL.

--
Life is Reality
Re:Think of what you just said by Shadowlore · 2005-02-19 15:53 · Score: 1

No, Linus said that ports *of existing drivers* to use the kernel code are acceptable, but new from cloth kernel modules are not. Thus there can be binary only modules. For example, nvidia took their initial code from their windows drivers and ported it to Linux. Thus those drivers are legal according to Linus.

But if nvidia had no drivers at all and made a binary only kernel module, Linus' statement is that those are violations.

--
My Suburban burns less gasoline than your Prius.
Re:Think of what you just said by cronius · 2005-02-20 00:37 · Score: 1

You are right. I stand corrected.

--
Life is Reality

Other goodies in QEMU by sanermind · 2005-02-19 07:06 · Score: 4, Informative

Not otherwise mentioned here: full sound support via SB16 emulation, built in NAT network support [the guest OS can use DHCP to get an address], and easy access to your linux filesystem on hosted windows environments via automatically configured SMB shares! It runs smbd as a slave process and comunicates with it over pipes, so it dosen't even require any root permisions.

qemu -hda /ahuge/unt/qemuXP.img -boot c -m 384 -user-net -enable-audio -smb /home/michael -pci &

Simple as that, and you're running XP with audio, network, and local filesystem access.

Of course, having to insmod a closed-source kernel module is unnerving, admitedly. But all of the above still applies and is usable without it, it's just not nearly as fast.

--

---
the pen is mightier than the sword, the sword is mightier than the court, the court is mightier than the pen.

Re:Other goodies in QEMU by Spy+Hunter · 2005-02-19 14:12 · Score: 1

Can you run Windows from the partition where it is already installed on your hard disk? That's all I'd ever want to use it for.

--
main(c,r){for(r=32;r;) printf(++c>31?c=!r--,"\n":c<r?" ":~c&r?" `":" #");}

The ignorant laugh by fm6 · 2005-02-19 07:06 · Score: 4, Insightful

Ha ha. The point is not just that you emulate the processor, but that you emulate the complete environment. In theory, you can do this without processor emulation using the Virtual Machine Monitor feature in IA-32 chips. Unfortunately, nobody except VMWare has figured out how to do this well enough to support guest OSs reliably.

I sometimes want to play old DOS games on my P3 machine. A lot of these games work in Real Mode, period. And they often can't do sound except on specific sound cards. I can:

Get a second machine for game playing. I don't spend that much time playing games. At least I hope not...
Reboot to DOS every time I want to run a game. Inconvenient, and the sound doesn't always work.
Fire up VMWare. Except I can't afford a copy right now...
Run DOSBox, which emulates not only a 286-class processor, but other legacy hardware such as sound cards and Hercules Graphics.

DOSBox is a really impressive bit of software, but it demands a lot of cycles to get the job done. (Typically, 75% of the CPU time on my P3 when I do VGA games full screen. Playing in a window is impossible unless I step down the color depth of my display.) So it's not good for much except real- and protected-mode games. An open-source emulator that doesn't have that kind of overhead would be very useful.

Re:How do we know it;'s legal? by cperciva · 2005-02-19 07:06 · Score: 2, Insightful

A closed-source module could contain stolen proprietary code

For that matter, an open-source module could also contain stolen proprietary code.

I hope this will be treated with caution until it can be ascertained to be fully legitimate...

I hope you never use any code which you didn't write yourself.

--
Tarsnap: Online backups for the truly paranoid

Status update on Linux/PPC test by zapp · 2005-02-19 07:10 · Score: 4, Informative

I decided to test this on my 1.25GHz G4 Powerbook running Yellow Dog Linux 4.0.

-It compiled in about 2 minutes flat
-The sample Freedos and Linux-Test images booted fine
- In the linux-test, /proc/cpuinfo reported an 18MHz Pentium Pro (ouch)
- I used qemu-img to create a 3gb disk image for testing Windows XP
- I booted off a Windows XP CD, and am in progress of installing on said disk image
( qemu -cdrom /dev/cdrom -boot d winxp.img )

--
no comment

Re:Status update on Linux/PPC test by The+Wicked+Priest · 2005-02-19 11:52 · Score: 1

This is actually off-topic... the article is about x86-on-x86, only. QEMU works on other architectures, yes, but not with "near-native" performance, and not with the closed-source module.

--
Share and Enjoy: 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

Escaping the Palladium Jail? by Jah-Wren+Ryel · 2005-02-19 07:12 · Score: 5, Interesting

Could this be used to sufficiently virtualize a 'palladiumized' system such that we could run a hypothetical DRM-up-the-ass version of Windows in it and then from the host OS side peek at all the secret data that the copyright cartel thinks is locked down?

--
When information is power, privacy is freedom.

Re:Escaping the Palladium Jail? by Weaselmancer · 2005-02-19 07:40 · Score: 1

You betcha! I'm sure that the hammer will come down as soon as "the powers that be" realize that no matter what DRM or other nasty stuff they come up with - it'll run in emulated machines.

For example, you could run QEmu and some sort of nasty DRM audio app like Napster. While it's running in the VM, drop back to windows and cat /dev/dsp into a file.

--
Weaselmancer
rediculous.
Re:Escaping the Palladium Jail? by Weaselmancer · 2005-02-19 07:42 · Score: 1

Argh. "Drop back to Linux". Too many layers of abstraction. ;^)

--
Weaselmancer
rediculous.
Re:Escaping the Palladium Jail? by vinsci · 2005-02-19 07:55 · Score: 1

If you can get the drivers that allows you to do that signed, sure. Don't hold your breath.
To avoid confusion: sure you could get that open source driver signed (for a hefty fee of course), but that driver wouldn't allow you to do any of the things you're thinking about. Changing the source to the driver after the signing won't help either, of course, as it then doesn't pass the signature test and thus isn't accepted into the so called trusted part the system.

--

Trusted Computing FAQ | Free Dawit Isaak!
Re:Escaping the Palladium Jail? by Anonym0us+Cow+Herd · 2005-02-19 08:43 · Score: 4, Interesting

NO. Not if the hardware has "trusted computing".

The trust begins at the hardware.

The motherboard must first examine the bootloader to see that it trusts it. If so, then trust is passed to the (un-compromised) boot loader.

The boot loader can look at the OS it is loading to see if it is trusted. If so, then trust can be given to the OS. An untrusted OS could be loaded, but it would have no trust. The untrusted OS can't ask the TPM chip "Hey, sign this for me using your secret private key."

If the OS is trusted, it can then determine if certian applications are trusted. QEMU would not be trusted. It could run, but would not be trusted.

When you run RIAA-Approved software, it is able to prove, via. a chain of trust (certificates) that it is trusted and running on a non-compromised system.

So why can't QEMU emulate the TPM chip? Because it does not know any secret keys. Each TPM has a secret key. And that key is a secret. Forever sealed in the chip. Even the manufacturer does not know the key. The key is a true secret. Nothing outside of the TPM chip itself has ever seen that private key. The TPM's public key was signed at the factory. You and everyone can know the public key of any TPM. And those public keys can be proven authentic because of their certificates (i.e. they are signed). You could generate an emulated TPM public/private key, but the public key is not signed by Dell. You know the public key of your box, but not the secret private key, which you can probably never know.

--
The price of freedom is eternal litigation.
Re:Escaping the Palladium Jail? by Anonym0us+Cow+Herd · 2005-02-19 08:47 · Score: 1

You betcha!

Boy thems stoopid trusted computin foaks shore is ignert. They must haves never thunk of emulation.

--
The price of freedom is eternal litigation.
Re:Escaping the Palladium Jail? by Jah-Wren+Ryel · 2005-02-19 09:45 · Score: 1

The motherboard must first examine the bootloader to see that it trusts it.

Exactly what does that mean? The BIOS on the motherboard? Why can't we emulate the "the motherboard," BIOS or whatever it is?

If we can convince the TPM chip that our emulator is "the motherboard" then we can bootstrap the rest of the system in fully trusted mode. As long as the TPM chip thinks it is talking to a fully trusted OS, we don't care diddly about the keys in the chip.

--
When information is power, privacy is freedom.
Re:Escaping the Palladium Jail? by JollyFinn · 2005-02-19 09:49 · Score: 1

I misread your post slightly. For a slight shock.

Argh. "Drop jhashdfh Linux". Too many lawyers of abstraction. ;^)

--
Emacs is good operating system, but it has one flaw: Its text editor could be better.
Re:Escaping the Palladium Jail? by sosume · 2005-02-19 10:14 · Score: 1

No way.

The OS is talking to emulated hardware, so you can generate whatever response is needed. You could for instance forward the call to the real TPM in the machine that runs the emulator, and then remove the protection by snooping /dev/audio. Windows won't be able to tell the difference.
Re:Escaping the Palladium Jail? by Anonym0us+Cow+Herd · 2005-02-19 11:24 · Score: 1

A little further down in this thread, I wrote the following...
Boy thems stoopid trusted computin foaks shore is ignert. They must haves never thunk of emulation.

You wrote...
The OS is talking to emulated hardware, so you can generate whatever response is needed.

The whole point is that YOU CAN NOT generate the needed response. The TPM's private key is a secret. See the rest of this thread. Your emulated TPM doesn't have a secret key whose counterpart public key was signed. The TPM is not going to give you, or anyone, its secret key. The TPM will sign something, but only if you can provide a certificate that you are trusted. You can only get this from the OS. The certificate can be inspected, just like happens on the web. I can see that the OS signed your certificate saying you are trusted, but can I trust the OS. Well, that certificate has embedded within it, a certificate from the boot loader that certifies that the OS is trusted. That certificate has an embedded certificate from the BIOS/motherboard saying that the boot loader is trusted. And that certificate has one saying that the hardware trusts the BIOS. The hardware trusts the BIOS because it was signed by someone that the TPM can trust (like Pheonix, etc.). I know that a trusted (not emulated) TPM signed it, because the TPM has a certificate that, lets suppose, Dell has signed saying that the TPM's public key is genuine. (Nobody has ever known, nor ever will know the TPM's secret key.)

So unless you are already in the chain of trust, YOU CAN NOT generate the needed response. Unless you have a certificate with embedded chain of certificates that you can present to the TPM, it isn't going to sign anything for you. Therefore, your plug-in cannot "prove" to the RIAA that it is okay to send me that media stream. And horrors(!), someone not in the room with you right now, not crouched around the monitor right now, might lay eyes on the video or hear the audio.

--
The price of freedom is eternal litigation.
Re:Escaping the Palladium Jail? by Anonym0us+Cow+Herd · 2005-02-19 11:31 · Score: 1

If we can convince the TPM chip that our emulator is "the motherboard"

Therein, I think, lies the problem. How do you convince it that you are the BIOS/motherboard. What if the TPM is only in a hardware mode to accept this convincing when the processor first boots from the ROM? What if the ROM's checksum must be signed by someone that the TPM itself trusts? (i.e. the TPM will only trust a ROM, whose checksum is signed by, let's suppose, Dell or Pheonix.)

Nevermind that the TPM is not in the right hardware mode to verify trust of the ROM right now.

Consider: don't you think that the "trusted" computing forces and their minions have thought of this? "Oh, duh! We were too stoopid to think of emulation. Dang!"

(I'm not saying that it might not be impossible to defeat, but it is not just going to be by using an emulator.)

--
The price of freedom is eternal litigation.
Re:Escaping the Palladium Jail? by Jah-Wren+Ryel · 2005-02-19 13:08 · Score: 1

What if the ROM's checksum must be signed by someone that the TPM itself trusts? (i.e. the TPM will only trust a ROM, whose checksum is signed by, let's suppose, Dell or Pheonix.)

That's not a problem because the emulator would be feeding the TPM chip the real ROM. Just like we would be feeding it the real bootloader and the real OS. Remember it is the CPU that executes the code, not the TPM. The CPU that we have absolute control over in an emulation environment.

Nevermind that the TPM is not in the right hardware mode to verify trust of the ROM right now.

That's a reasonable defense. But how strong is it really? What if we can glitch the TPM into "rebooting" without disturbing the rest of the system? Perhaps through software or perhaps through a careful physical hack, like shorting the chip's reset pin.

Consider: don't you think that the "trusted" computing forces and their minions have thought of this? "Oh, duh! We were too stoopid to think of emulation. Dang!"

Either you've said that twice now or there was someone else with the same opinion. Either way, my answer to that is BULLSHIT! If I wanted to believe in the designers' infallibility, I would never have asked the question in the first place. It's the suits in the copyright cartel who buy that shit and the geeks on the net who prove them wrong.

--
When information is power, privacy is freedom.
Re:Escaping the Palladium Jail? by Rich0 · 2005-02-20 01:44 · Score: 1

That's not a problem because the emulator would be feeding the TPM chip the real ROM.

The TPM chip is examing the BIOS before the system even boots. The BIOS examines the OS before it runs, etc. By the time the emulator even starts running, the TPM chip has already stored the state of everything on the system in its protected memory.

The emulator could feed it a new BIOS, but the chip would not accept it, since it knows that the boot sequence is already done. The emulator can emulate the CPU, but the CPU doesn't have any control over the TMP chip. The only reason you can do magical things with an emulated CPU is because it is EMULATED. The TMP chip is not, and you can't emulate it without knowledge of the private key...
Re:Escaping the Palladium Jail? by KiloByte · 2005-02-20 03:04 · Score: 1

I'll give it about 1 month after release for some hardware geek to fish a secret key out of one of those chips.

Indeed, all it takes is a single signed private key to break the whole network of "trust". This means, if a leak occurs at any of the plethora of hardware manufacturers -- or, a hardware geek manages to fish out the key from his piece of junk, you can use feed this key to QEMU or any similar emulator.

Such a key would work with any piece of software released prior to the leak, and most likely, with all software released later until the point when support for all hardware signed with the original signing key (one that you have never even seen) is dropped. As long as it's you who control the software and "hardware", there is no real way to revoke a compromised key -- the cartel would have to issue a new signing key and use it on any new equipment.

--
The creatures outside looked from Alt-Right to Antifa; but already it was impossible to say which was which.

Re:I'm getting pretty skeptical of these things by TheRaven64 · 2005-02-19 07:14 · Score: 5, Informative

First of all, this is by the guy who already wrote FFMPEG (MPEG transcoder), the Tiny C Compiler, and QEMU - an emulator that is capable of emulating several different CPUs running on several other host CPUs at impressive speeds (compare it to Bochs sometime). Quite an impressive resumé, and someone whose claims I would be inclined to take seriously.

Secondly, QEMU Accelerator is not an emulator, it is a virtualisation layer. It executes most instructions on the host CPU, using the emulator as a sand-box and system emulator, rather than a CPU emulator. Emulating x86 on x86 at a reasonable speed is really not that hard (non-trivial, but certainly not unbelievable).

--
I am TheRaven on Soylent News

What For? by Superduck-Canuck · 2005-02-19 07:22 · Score: 1

Why would you even need an emulator? You can have a Workstation (Say A Windows PC) running your Windows Apps. Load Cygwin on it, use it as an XTerminal. Get a headless Linux Box, network the two together. Get a Sparc, network it. Have all your displays on your Windows system. Use NFS and share files. You can run multiple OSes by having a mutltihost networked system.

Re:CherryOS by PatrickThomson · 2005-02-19 07:24 · Score: 1

Yeah, iirc there was a variable called SPIRO_MULTIMAX_3000

Not really common, I would say.

--
I am one of many. My idea is not unique, nor do I expect my voice alone to sway you. I speak in a chorus of opinion.

Re:to wine or not to wine by Weaselmancer · 2005-02-19 07:31 · Score: 3, Informative

But what, exactly, is best?

Well, in a lot of ways that's like asking what's the best tool in your toolbox? It depends on what you're trying to do. Hammer a nail, use a hammer. Tighten a bolt, use a wrench. A wrench makes a lousy hammer and a hammer makes a lousy wrench.

Wine is great for running x86 Windows apps natively on Linux. It has no Microsoft code in it whatsoever - that's it's advantage. No MS tax to run MS apps. Downside is that it isn't 100% accurate - some apps don't work. By its very nature it'll always be a work in progress.

Qemu (and VMware and their like) emulate a complete computer. They're slower because they have to emulate hardware. And to run Windows apps, you'll need a copy of a Microsoft OS, so you've got a licensing issue there. As a plus, you can do development on a virtual machine and cause any kind of damage, and (if you've backed up before you tried it), fixing things is as simple as tar -xvzf yourbackup.tgz.

Explore both options a bit and you'll find that each one solves a particular class of problems. More tools in the toolbox is always good.

--
Weaselmancer
rediculous.

Re:People have to ask... by 0racle · 2005-02-19 07:31 · Score: 2, Insightful

This isn't insightful, they have no obligation to you to sell systems without Windows. Go to a car dealership and ask them where are their models without engines are. Then when they look at you strange say, "sorry, in the market for a cheap new car, but want my own engine and not paying you extra for it when I am not going to use it." Makes you sound crazy now doesn't it. You sound just as crazy when you say that about a little computer, especially considering you already know they don't a configuration like that and you know exactly where you can get one.

They don't care if you don't buy from them, you just make yourself look like an idiot.

--
"I use a Mac because I'm just better than you are."

little disappointed... but it's still good news by CaptainPinko · 2005-02-19 07:37 · Score: 1

from the headline I got really excited but then I read that it only speeds things up on Linux. Well this is great for people running Windows on Linux it does nothing for those running Linux on Windows. I use a TabletPC that won't even accept regular Windows XP install discs... just the Acer supplied ones. I've tried running DSL Linux under QEMU but the performance isn't there at all (laptop is a 800Mhz, consider what XP uses and then divide the remainder by 5-10 and you get the idea).

Hopefully there is enough deman to have a similar layer developed for Windows.

Then, if only I could figure out how to communicate between the Windows host environment and the emulated Linux one, I'd be all set.

--
Your CPU is not doing anything else, at least do something.

Re:little disappointed... but it's still good news by Quiberon · 2005-02-19 08:41 · Score: 1

Does SAMBA work ?
Re:little disappointed... but it's still good news by johndoe42 · 2005-02-19 09:26 · Score: 1

What kind of Acer tablet do you have? My C300 runs my MSDN Windows install just fine, and it also runs SuSE 9.1.

You cold also give CoLinux a try.
Re:little disappointed... but it's still good news by CaptainPinko · 2005-02-19 09:50 · Score: 1

I can't even seem to ping the Windows from the DamnSmallLinux :(

--
Your CPU is not doing anything else, at least do something.

Re:People have to ask... by ip_fired · 2005-02-19 07:53 · Score: 1

A better car analogy would be to ask the dealership for the cars without stereo systems. Because I have my own, good stereo system and I don't need a crappy factory installed one. That is more like that OS than the engine (CPU).

--
Don't count your messages before they ACK.

DRM implications by cryptoluddite · 2005-02-19 07:57 · Score: 1

Having a free, open-source virtual machine means you can easily copy songs from napster or any other source regardless of DRM. It should be trivial to patch the sound blaster emulation to start dumping to disk when the sound starts and stop once the device is closed or the sound pauses for 1 sec or so. This would make it trivial to copy napster songs regardless of what they do to lock down Windows itself, and it requires almost no work on the part of the user (they just play the song in a qemu-xp and it's saved as qemu###.wav... it would be a bit more difficult to automatically get the song name). But no manual starting/stoping of the recording is needed.

Ultimately this will be used to justify locking the system down at the hardware level with a verification key of some sort (aka paladium or whatever). If the BIOS has a private key and the OS includes the public key then it can encrypt something only the real bios can decrypt, so a virtual pc will need to either have the private key or be able to fake the OS. But ultimately the OS depends on the processor and system to be "honest" so the only way to prevent a QEMU from tricking the OS is to prevent it from running at all (the actual OS must prevent arbitrary programs from running).

I suppose locking up the computer so only offical, blessed software can run so that megacorps can deny us fair-use rights is called progress. But until then I'll continue to donate to wikipedia (so we know) and add more GPL code to the world (so we can do something about it).

Re:DRM implications by Wesley+Felter · 2005-02-19 08:52 · Score: 1

Emulating a Soundblaster won't bypass Secure Audio Path AFAIK. You'd have to patch the Windows kernel to bypass it (which shouldn't be too hard for some people, if only they'd use their mad skillz for good).

(BTW, I wonder if a crypto-luddite is opposed to cryptography or secretly opposed to all technology...)
Re:DRM implications by spongman · 2005-02-19 09:08 · Score: 1

this is true if the application (ie napster) isn't using the full capabilitie of DRM (which it currently doesn't, see below).
however, if it were, the decoded PCM stream that the sound card gets still would contain significant amounts of noise. it's up to the sound card to remove this noise in hardware. unless you're a trusted hardware vendor, microsoft won't tell you how to do this.
napster doesn't use this secure audio path mainly because nobody has sound cards that support it.
Re:DRM implications by GoRK · 2005-02-20 05:32 · Score: 1

The proposed Secure Audio Path stuff requires Microsoft to sign and certify a driver that they feel is sufficiently providing a secure audio path. Although this is feasable with many older creative labs soundcards such as the SB Live! (where you could in theory put decryption code on the DSP and send only crypted data for playback), it's likely not possible for the SB16 device that most virtualizers choose to implement. In other words, there will probably never be a secure audio path driver for SB16 because the hardware is not capable of any kinds of security.

The workaround of course is to emulate more modern hardware or exploit the DRM in some other area of the windows kernel. Then there's always the possibility that MS will simply sign a SB16 driver to prevent customer complaints but thereby leaving a huge gaping hole in their system.

It reminds me of this whole recent airport security business.. TSA will happily tear through my baggage with their little wand and mass spectrometer (ask them; they don't even know what their $30K machine that does the testing is called) to look for explosives while bypassing explosives screening on every single golf bag that comes through. The reason? Golf bags come up positive for explosives on their initial tests very often because of fertilizers. When they do, they have to use a more expensive test to more specifically identify the concentrations and type of the residual -- This takes too much money and time, apparently for TSA to stomach, so the policy at some airports has been simply to bypass the check. Idiots.

Having only one computer by J_Omega · 2005-02-19 08:04 · Score: 1

For people trying to migrate away from Windows, the solution isn't usually going to be to buy more hardware.

Re:People have to ask... by 0racle · 2005-02-19 08:06 · Score: 1

I chose to use the engine because the car would be useless without it , just like without a OS, a desktop is essentially simply a paperweight.

--
"I use a Mac because I'm just better than you are."

Damned impressive...(how to test with Knoppix) by jbwiv · 2005-02-19 08:10 · Score: 3, Informative

Wow...I'm simply staggered by this. I've been waiting around for Netraverse to release a version that supported Windows XP...now I no longer have to wait!

All you haters need to quit your bitchin'....Mr. Fabrice Bellard, THANK YOU FOR THIS PROJECT!

If you'd like to give it a try with Knoppix and you have an ISO lying around, type:

qemu -cdrom /path/to/knoppix.iso

Boots up perfectly. WOW!

Re:Damned impressive...(how to test with Knoppix) by jbwiv · 2005-02-19 13:21 · Score: 1

Ummm...no they haven't. Their site still lists Workstation 5.0 as the latest, and it only supports 95/98/ME...

he's probably best off just releasing it by idlake · 2005-02-19 08:12 · Score: 3, Insightful

The author writes:

As a supporter of open source, the author accepts to open source the QEMU Accelerator Technology provided a company invests enough money to support the project and to recompense the author from the potential loss of revenue. Interested companies can look at the roadmap and make suggestions to the author.

What matters isn't really the potential loss of revenue, but the expected loss of revenue, i.e., large amount of money multiplied by the low probability of actually succeeding.

I'm sure the accelerator is skillfully written, but I think chances of turning QEMU+Accelerator into something commercially successful are next to nil. Why? Foremost because the market already has VirtualPC and VMware in it, created and maintained by big companies with deep pockets, lots of lawyers, and large patent portfolios. Oh, and then there are coLinux, Xen, and UML that he would be competing with as well.

To compete, he'd have to get startup funding, a management team, developers, support, and, worst of all, sales people: a lot of work, a lot of money, and next to impossible if you don't either get a lot of buzz or know the right people. Then he'd have to develop a product; a product isn't just a working piece of code, it's documentation, training, tutorials, travel, presentations. People don't expect that from FOSS, but they do expect it as soon as they pay a couple of bucks for something. Even if he does all that, the most likely outcome is still that the startup fails. But even if it succeeded, he would probably only end up with a small slice of a moderately successful company (competition and all that), and only after spending several years of his life doing things he probably doesn't enjoy.

I think people underestimate how hard it is to commercialize something, even something that is really good and novel (his software may be really good, but it isn't novel).

I think his best bet would be release it under a dual license (GPL/commercial) right away, while people are paying attention, and build a consulting business and commercial licensing around it. He won't become an instant billionaire that way, but if he has a worthwhile product, it can be a steady source of good money doing mostly what he enjoys. If he sits on the software too long waiting for a sugar daddy, it will be less and less likely that he will be able to get anything out of it.

Oh, and in case you are wondering whether my argument is disingenuous and I just want to get a free virtual machine, it isn't: I already have VirtualPC and VMware, and I actually prefer the various user mode Linux solutions at this point.

Re:Willing to sell != willing to open source by cronius · 2005-02-19 08:17 · Score: 1

I think it's definitely desirable to open (or rather, Free) the source. There are other archs and other emulators that could benefit from a port of the technology, and there are Free Software ethics at stake which affect society, if nothing else.

Yes, I didn't think about other archs. Some posters misunderstood my original question, I was wondering if this mod was near perfect (as opposed to really perfect), and that once it reached that stage it would be complete, and couldn't add anything more to QEMU on its own (it's a single module doing a spesific task after all).

But porting it to other architectures is definitly worthwhile, and freeing it would ensure compatibility with future kernels. I just didn't think about that :)

--
Life is Reality

Re:People have to ask... by drfreak · 2005-02-19 08:22 · Score: 1

You are both morons. If someone is smart enough to know the difference between windows and linux, they are smart enough to build their own computer. The car reference just does not apply. Are people crazy who replace the engine in their car with a beefier one?

Re:People have to ask... by homer_ca · 2005-02-19 08:22 · Score: 1

The OS isn't analagous to the engine of a car, but aside from that, you don't have a Microsoft engine installed under the hood of every preassembled car on every dealership lot. I understand all the big name brand companies want to sell you a complete working system. The difference here is Microsoft has a monopoly on one component of a complete computer system, and many OEM agreements required computer makes to pay Microsoft for every system shipped even if it had a different OS under the hood.

Free by h4x0r-3l337 · 2005-02-19 08:24 · Score: 1

The catch is that although QEMU is released under various open source licenses, the Accelerator uses a free (as in beer) license

"Free as in beer" is good enough for all but the most demanding zealot.

Re:Free by Grey_14 · 2005-02-19 11:51 · Score: 1

Not really, The site state's it cannot be distributed seperatly, without specific approval from the author, meaning no distro's will be able to include it, or even package it.

Linuxes for Windows by Quiberon · 2005-02-19 08:27 · Score: 1

Pile of torrents at torrents for linux-under-windows isos. Look like SuSE for Windows . 2.4 kernels perform better than 2.6 ones, but all work !

Re:if only they could get this to run faster by CdBee · 2005-02-19 08:31 · Score: 1

don't - or every other .sig will be a link to freeiGods.com soon

--
I have been a user for about 10 years. This ends Feb 2014. The site's been ruined. I'm off. Dice, FU

Re:How do we know it;'s legal? by CdBee · 2005-02-19 08:37 · Score: 1

Well I apologise for the slight to M.Bellard's ego caused by my not knowing who he is. I often come across binary-only applications that promise to speed up my PC one way or another.. usually they're some sort of scam, or adware, spyware, or just plain and simple warez

It is only a matter of time - remember this - before Liux users will face as many hazardous closed-source apps as Windows users have to deal with.

Opening the source - even if under a restrictive licence - is a good way to encourage peer audit and safety

--
I have been a user for about 10 years. This ends Feb 2014. The site's been ruined. I'm off. Dice, FU

Re:How do we know it;'s legal? by snorklewacker · 2005-02-19 08:54 · Score: 1

OOh, Fabrice of tcc fame. tcc is really lots of fun, very useful when doing test compiles because it's just so fast. It's not perfect (seems to have problems compiling Python) but it's a nice little tool.

There was an effort to port it to Windows, but it went nowhere. It's not as important since cl.exe is quite fast, but it's great for testing C programs when you don't feel like waiting on pokey old gcc.

--
I am no longer wasting my time with slashdot

Not pedantism... by rbarreira · 2005-02-19 09:06 · Score: 2, Informative

There's a technical name for software like WINE. It's not an emulator, it's a compatibility layer. Look it up on wikipedia.

--

The AACS key is NOT 0xF606EEFD628B1CA427BEA93A9CA9773F

Re:Not pedantism... by rbarreira · 2005-02-19 23:28 · Score: 1

Look it up elsewhere then...

--

The AACS key is NOT 0xF606EEFD628B1CA427BEA93A9CA9773F

I can install OSX on qemu under Linux or Windows? by crivens · 2005-02-19 09:13 · Score: 1

So does this mean I can install OSX on qemu under Linux or Windows?

Re:Insideous GPL by SLi · 2005-02-19 09:33 · Score: 1

Funny. He proved my point of why the GPL is good and actually works.

I hope not. by user9918277462 · 2005-02-19 09:37 · Score: 2, Interesting

The author is essentially blackmailing the F/OSS community by refusing to release the source unless a benefactor comes forward to fund him. If he is successful it will only encourage others to try the same tactic.

He claims it's to cover "lost revenue" if he opens the source but the reality is that he is getting no revenue currently and is unlikely ever to. VMware is the undisputed king of the vitualization market (with MS Virtual PC catching up).

VMware Workstation works extremely well. QEmu would have an extraordinarily difficult time competing in the proprietary software space, and the author knows this. If he had asked for donations to help support development I would happily contribute and encourage others to, but this just stinks.

Re:I hope not. by Nailer · 2005-02-19 17:07 · Score: 1

The author is essentially blackmailing the F/OSS community by refusing to release the source unless a benefactor comes forward to fund him.

You're begging the question: how on earth does that equate to blackmail? Blackmail usually involves the perpetrator doing something negative. Fabrice has done something wonderful, which you can run without an non OSS apps. If you choose to however, you can use his other wonderful thing, which is prprietary until someone pays him for it. A pretty decent thing: having the ability to run Windows apps inside Linux is somethign a lot of commercial Linux distros would benefit from. Good on Fabrice for making some rent.

If he is successful it will only encourage others to try the same tactic.

Get paid for their work? Or Open Source cool apps if they're paid?

Awesome.

VMWare workstation is, BTW, incredibly overpriced.
Re:I hope not. by serbanp · 2005-02-19 21:46 · Score: 1

Considering the huge amount of very high quality code this guy donated to the community, I would be more considerate when opening the mouth.

Come on, a little decency, for God's sake!

Colinux by Aggrajag · 2005-02-19 09:44 · Score: 1

Colinux runs Linux at native speeds under WinXP.

Is this really going to matter? by Hack+Jandy · 2005-02-19 09:51 · Score: 1

Considering AMD and Intel's push for virtualization *in* the CPU, are projects like these really going to have a future? Why run Windows on top of Linux when in the near future I will be able to run them side by side.

Re:Is this really going to matter? by ctid · 2005-02-19 09:58 · Score: 1

Unless you have a time machine that you are not telling us about, it matters now, which is when I live.

--
Reality is defined by the maddest person in the room

Trying to hand a retailer money... by zogger · 2005-02-19 09:54 · Score: 1

is somehow strange, idiotic or wrong? Say what? It's why they have a store in the first place.

Actually, it works quite well once enough people do it.some examples here: I was one of those people who consistently refused to answer personal questions to buy something at Radio Shack, when it was their corporate policy. Did it for a couple of years and I know quite a few other people did it too,and this has been covered on slashdot before, eventually they stopped with the 50 questions. I have asked the local grocery store where I shop to carry this and that items they didn't stock,in the produce and in the meat section, now they stock those items and the department managers thanked me, said their sales went up. In fact I was directly responsible for the large grocery here always carrying fresh cooking herbs, first I asked if they would like to try some to sell, they said sure, I sold them a few bags of extra from our gardens. they sold well, now they carry them all the time and the guy is looking forward to my cheaper fresher stuff again this upcoming season. so he'll have two "brands" for people to choose from when before they had not much of any herbs there.

Back when we second amendment guys got a wild hair about rosie o donnel being kmart tv spokesmodel, I went to my local kmart manager and said I would boycott until she got removed, thousands of other people did the same thing, and we wrote letters to kmart and to their tv advertisers, etc, and eventually they fired her. Customer feedback works once it hits some magical critical mass level, always a variable, but it's there.

As to getting cars or trucks with or without this or that, you can do it, most dealers will arrange it if you got cash in hand. You won't get the same warranty,if any, but it's possible. I know for example just on the truck side of things you can order a truck with zero bed, just the frame, and put your own bed on it. It's not common, but it's not all that uncommon either, and to get to computers, please, installing an OS is just a tad easier than putting a new engine in a car, easier, cheaper, computers are designed to accept disks and keyboard and mouse commands, it's just not that strange of a request, and there are some chains that do this now, just not all of them, but eventually I think you'll see it be pretty much universal.

There's no harm in asking your retailer to provide you with something you want to buy, this is called customer feedback and can work if *enough* people do it. They are in business to sell you stuff and make money, that's their bottom line. Sure, a dozen people doing it won't result in much, but say if some place like office depot got 10,000 requests over a one week period or a month, and the requests were to actual main store managers? That info would filter upstream and at least get seriously considered. If people want no OS or x-Linux OS preinstalled, then I say they should ask for it at the retail level. It's no different than asking for anything else.

Re:Insideous GPL by nurb432 · 2005-02-19 09:54 · Score: 1

I suppose that is true if you dont mind losing rights to your code..

As a *user* you win since they are forced to give it to you for free.

As a *devloper* you lose, as you are forced to give your work away..

--
---- Booth was a patriot ----

I just tried the CVS+kqemu---It ROCKS by waferhead · 2005-02-19 10:20 · Score: 1

Saw this and (strangely) the site wasn't /.ed...
It built and works EXCELLENT.

To run a Knoppix CD:
qemu -cdrom /dev/cdrom

To boot off CD, and have a HD image (to install on, for example)
qemu -boot d -cdrom /dev/cdrom -hda (some image file)

Look in the fine docs for qemu-img docs.

VMware is do dead...

Re:I just tried the CVS+kqemu---It ROCKS by waferhead · 2005-02-19 10:37 · Score: 1

Added note: modprobe kqemu

(I added it to /etc/modprobe.preload)
I'm running Mandrake Cooker.

SoftWindows by Brian+Brian · 2005-02-19 10:41 · Score: 1

Near Native. Right. I have heard that story for years. For me it started with claims that on the PowerPC they could achieve 95% of native performance when emulating Windows. Maybe if you had some hardware betwwen memory and the CPU that translated on the fly then it might work. Look at Transmeta. Even in micro-code they can't compete on performance.

Re:Insideous GPL by The+Wicked+Priest · 2005-02-19 11:45 · Score: 1

You're not forced to do anything. You chose to accept the GPL by working with GPL'd code in the first place. If you don't want to GPL your code, then stay clear of existing GPL'd code -- or get a separate, commercial license from the original developers, just as you would for proprietary code. It's that simple.

--
Share and Enjoy: 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

IBM and Cell processor by fbonnet · 2005-02-19 12:10 · Score: 2, Interesting

I understand that, QEMU being an unfunded project, the author hasn't enough resources to port this technology to other host platforms.

However if the Accelerator Technology is portable to platforms such as PowerPC or Cell, then it perfectly makes sense to keep it closed source if the goal is to open source it in exchange for financial or material support. Imagine what IBM could do with a near-native performance x86 emulator running on Cell... No more counter-arguments about compatibility, and Cell becomes instantly the fastest x86-compatible processor on the market. Given Cell's potential for parallel computation, this should beat the crap out of Opteron or whatever Intel could release in the near future.

Re:IBM and Cell processor by sketerpot · 2005-02-19 15:48 · Score: 1

It's not portable to anything other than x86, ever. Plain old QEMU is, though, and it gets pretty decent speed.
Re:IBM and Cell processor by puetzk · 2005-02-20 05:28 · Score: 1

First off, the techniques involved are such that you'd be emulating cell on cell, not x86 on cell - what it's doing is managing to run most of the instructions natively, without any emulation.

Porting this wouldn't make a whole lot of sense other than collecting existing work into one project; virtualization is already available on most other architectures, because it's not as hard there. If the instruction set was designed for it and the chip gives a good way to hook the instructions that need to be emulated, it can be fairly straightforward. See http://maconlinux.org/ (which is opensource, and does PPC-on-PPC), HP's VPAR, IBM's LPAR, etc.

x86 on x86 is a particularly difficult case though; plex86, Xen, etc all do it to some extent, but forbid certain operations that are too hard to emulate. If you have control of your guest kernel (which is the only place you have to worry about privileged instructions anyway), this works well. If you want to run unmodified copies of windows, it doesn't. VMware pioneered some techniques to trap these instructions, and thus provide a full x86 instruction set, Microsoft followed with Virtual PC.

Now qemu apparently can do this too, which is quite cool, but mostly because it's now possible in an architecture where the rest of the emulation system is GPL.

--
The Matrix is going down for reboot now! Stopping reality: OK. The system is halted.

A bit of good and bad by DasIst · 2005-02-19 12:15 · Score: 1

From trying out Knoppix, I'm pretty impressed by the speed. It's been long enough since I last tried qemu with knoppix that I can't say if the new module is helping out to a large extent, but it's still pretty cool. The downside is that I tried installing Windows98, and on rebooting it dies with a "windows protection error". On doing some searching, it seems this is a known problem with the 9x series using this module.

Re:NEWS! by mollymoo · 2005-02-19 12:23 · Score: 1

It is cheaper with greater performance to buy the most recent Intel or AMD system and dynamically translate Mac OSX on that x86 host, than it is to purchase that expensive PowerPC system to run Mac OSX.

Bullshit. Show me the tests. I want to see the x86 hardware which can emulate a Dual 2.5GHz G5 and beat it for speed, for less money.

--
Chernobyl 'not a wildlife haven' - BBC News

Re:How do we know it;'s legal? by kasperd · 2005-02-19 13:01 · Score: 1

I hope you never use any code which you didn't write yourself.

Even that is no guarantee the code is legal. Somebody may have patented your algorithm.

--

Do you care about the security of your wireless mouse?

Re:Free-Hardware loans. by h4x0r-3l337 · 2005-02-19 16:30 · Score: 1

Oh right, I forgot that my Radeon will stop working the moment ATI goes out of business...

Forget Samba... by gg3po · 2005-02-19 17:22 · Score: 1

...what I've found to be easiest method is using ssh/sftp from the guest to the host computer. You need to make sure to use the -user-net option and that sshd is running on the host (Cygwin can do this in Winblow$). It's worked for me every time.

--
---

Re:I can install OSX on qemu under Linux or Window by xororand · 2005-02-19 22:02 · Score: 1

You could use PearPC (http://pearpc.sourceforge.net/) for that. A friend of mine successfully finished his Mac OS X Installation just a few days ago and reported 'usable' speeds, whatever this means.

OT Free Ipods by nurb432 · 2005-02-20 03:09 · Score: 1

Sure they exist, i got one the other day for nothing.

The guy down the hall bought one and left it on his desk unattended. "free ipod".

--
---- Booth was a patriot ----

Re:NEWS! by resiak · 2005-02-20 04:14 · Score: 1

CherryOS!

(...for those of you who don't remember...)

How does it compare to CoLinux? by Ed+Avis · 2005-02-20 08:36 · Score: 1

Cooperative Linux already provides a way to run Linux under a kind of hardware emulation. If you wanted to run other operating systems then QEMU would be useful, but if you just want a way to run Linux on your Windows desktop then a good answer already exists.

--
-- Ed Avis ed@membled.com

Re:I can install OSX on qemu under Linux or Window by Bri3D · 2005-02-21 13:29 · Score: 1

To answer parent+grandparent: No and I agree. You cannot run OSX on qemu yet there's a cd emulation problem(see the qemu compatability page for more). I personally have a copy of 10.3 on PearPC. It's perfectly useable for web/email, though office is really slow. But I have a lowly P4 2.4, so others may have better experiences.

Slashdot Mirror

QEMU Accelerator Achieves Near-Native Performance

232 of 366 comments (clear)