How about this replacement of WINE.

bussuser · Post by **bussuser** » Thu Feb 28, 2008 11:26 am

http://www.reactos.org/forum/viewtopic.php?t=5240

It is based on WINE and ReactOS .

The project is sponsored by a chinese company.

Dan Kegel · Post by **Dan Kegel** » Thu Feb 28, 2008 11:38 am

On Thu, Feb 28, 2008 at 8:26 AM, bussuser <[email protected]> wrote:

http://www.reactos.org/forum/viewtopic.php?t=5240

It is based on WINE and ReactOS .

It's not a wine replacement; it uses wine.
Better description at
http://www.reactos.org/forum/viewtopic.php?t=5240

It looks like they added hooks to the linux kernel
to accept windows nt syscalls. Maybe they even allow using the
system's normal shared library loader instead of Wine's
special one. This is something
I've often wanted to do, but it was way lower priority
than getting wine working. I haven't looked at their
project at all, no idea if it was done well.
- Dan

DARKGuy . · Post by **DARKGuy .** » Thu Feb 28, 2008 11:47 am

However, it's a great, great project IMO. It was about time something like
that should've been made. Redirecting calls is simply acting as a proxy,
instead of emulating. However, it makes it only to be usable in Linux... but
hey, this is genius.

On Thu, Feb 28, 2008 at 12:07 PM, Dan Kegel <[email protected]> wrote:

On Thu, Feb 28, 2008 at 8:26 AM, bussuser <[email protected]>
wrote:
http://www.reactos.org/forum/viewtopic.php?t=5240

It is based on WINE and ReactOS .
It's not a wine replacement; it uses wine.
Better description at
http://www.reactos.org/forum/viewtopic.php?t=5240

It looks like they added hooks to the linux kernel
to accept windows nt syscalls. Maybe they even allow using the
system's normal shared library loader instead of Wine's
special one. This is something
I've often wanted to do, but it was way lower priority
than getting wine working. I haven't looked at their
project at all, no idea if it was done well.
- Dan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.winehq.org/pipermail/wine-us ... chment.htm

Dan Kegel · Post by **Dan Kegel** » Thu Feb 28, 2008 12:15 pm

On Thu, Feb 28, 2008 at 8:47 AM, DARKGuy . <[email protected]> wrote:

However, it's a great, great project IMO. It was about time something like
that should've been made. Redirecting calls is simply acting as a proxy,
instead of emulating.

Wine is not an emulator. No emulation is going on anyway.

We would, I think, like to move wineserver into the kernel
sometime. It's been discussed before. Linus is not
opposed to having native support for win32 system calls.
A fellow at Redhat wrote a kernel module for wine a number
of years ago, but it just wasn't time yet.

For a related BSD thingy, see
http://www.kernel-traffic.org/wine/wn20010313_87.html#1
That project was still going as of 2006, I think, but didn't
get terribly far.

- Dan

David Gerard · Post by **David Gerard** » Thu Feb 28, 2008 12:23 pm

On 28/02/2008, Dan Kegel <[email protected]> wrote:

We would, I think, like to move wineserver into the kernel
sometime. It's been discussed before. Linus is not
opposed to having native support for win32 system calls.
A fellow at Redhat wrote a kernel module for wine a number
of years ago, but it just wasn't time yet.

If Wine works, what would you gain by moving it out of userspace?

- d.

Dan Kegel · Post by **Dan Kegel** » Thu Feb 28, 2008 12:48 pm

On Thu, Feb 28, 2008 at 9:22 AM, David Gerard <[email protected]> wrote:

If Wine works, what would you gain by moving it out of userspace?

Performance, mostly. Maybe also more of a notion of
a whole system (right now, wine is kind of per-user).

It's pretty low priority.

David Baron · Post by **David Baron** » Thu Feb 28, 2008 12:53 pm

However, it's a great, great project IMO. It was about time something
like that should've been made. Redirecting calls is simply acting as a
proxy, instead of emulating.
Wine is not an emulator. No emulation is going on anyway.

We would, I think, like to move wineserver into the kernel
sometime. It's been discussed before. Linus is not
opposed to having native support for win32 system calls.
A fellow at Redhat wrote a kernel module for wine a number
of years ago, but it just wasn't time yet.

For a related BSD thingy, see
http://www.kernel-traffic.org/wine/wn20010313_87.html#1
That project was still going as of 2006, I think, but didn't
get terribly far.

OK. Has anyone actually tried this thing?

Patching kernels and building modules is no big thing but patching and
compiling wine, being tied to an older version or constantly upgrading and
redoing this might be a bit too much.

Steven Edwards · Post by **Steven Edwards** » Thu Feb 28, 2008 1:15 pm

On Thu, Feb 28, 2008 at 11:37 AM, Dan Kegel <[email protected]> wrote:

It's not a wine replacement; it uses wine.
Better description at
http://www.reactos.org/forum/viewtopic.php?t=5240

It looks like they added hooks to the linux kernel
to accept windows nt syscalls. Maybe they even allow using the
system's normal shared library loader instead of Wine's
special one. This is something
I've often wanted to do, but it was way lower priority
than getting wine working. I haven't looked at their
project at all, no idea if it was done well.

A long long time ago there was a project called Captive NTFS that was
mostly the same idea but in userspace, They tool the ReactOS kernel,
combined it with FUSE, and used that to load the native ntfs.sys. It
actually took a little more work because ReactOS Cache Manager was so
broken (another sign there is no way in hell the code was stolen from
Microsoft) that the author had to profile the native ntfs.sys and
rewrite the entire Cache Manger in glib. It would not have been much
of a stretch to add support for Win32 applications on top using Wine.
This might end up being a better solution for certain device drivers
like the copyprotection systems.

--
Steven Edwards

"There is one thing stronger than all the armies in the world, and
that is an idea whose time has come." - Victor Hugo

bussuser · Post by **bussuser** » Fri Feb 29, 2008 11:09 am

MS windows device drivers working on linux would be very very tempting. I hope they implemente the WDM device driver framework this year.

robert w hall · Post by **robert w hall** » Sat Mar 01, 2008 9:14 am

In article <[email protected]>, bussuser <wineforum-
[email protected]> writes

MS windows device drivers working on linux would be very very tempting. I hope
they implemente the WDM device driver framework this year.

But just implementing a 0x2e syscall entry point won't suffice surely,
for most drivers?
An exception might be usbscan.sys, (the usb bits of which can be done
quite happily by libusb in user mode), and for which (I believe) the
present stopper for is the need to patch or generalise CreateFile and
friends...
Bob

--
robert w hall

robert w hall · Post by **robert w hall** » Sat Mar 01, 2008 9:23 am

In article <[email protected]>
, Dan Kegel <[email protected]> writes

On Thu, Feb 28, 2008 at 9:22 AM, David Gerard <[email protected]> wrote:
If Wine works, what would you gain by moving it out of userspace?
Performance, mostly. Maybe also more of a notion of
a whole system (right now, wine is kind of per-user).

It's pretty low priority.

Anyone got this to work? What .config options do I _need_?

The original kernel .config file builds 1500 modules!, so I stripped it
right down, to get a manageable 2.6.23 kernel which appears to be OK in
linux; then patched and built wine. But the two don't appear to be
playing correctly together...
--
robert w hall

bussuser · Post by **bussuser** » Tue Mar 11, 2008 7:00 am

Dan Kegel wrote:On Thu, Feb 28, 2008 at 8:26 AM, bussuser <[email protected]> wrote:
http://www.reactos.org/forum/viewtopic.php?t=5240

It is based on WINE and ReactOS .
It's not a wine replacement; it uses wine.
Better description at
http://www.reactos.org/forum/viewtopic.php?t=5240

It looks like they added hooks to the linux kernel
to accept windows nt syscalls. Maybe they even allow using the
system's normal shared library loader instead of Wine's
special one. This is something
I've often wanted to do, but it was way lower priority
than getting wine working. I haven't looked at their
project at all, no idea if it was done well.
- Dan

The linux unified kernel uses wine and reactos' codes, but its technique is different from the technique of wine.

Just like KVM, it uses a lot of XEN's codes, but the technique of kvm is different from XEN.

here is the plan of the unified kernel---translate by google.com

http://translate.google.com/translate?u ... N&ie=UTF-8

-

Phobos · Post by **Phobos** » Thu Mar 13, 2008 9:15 am

bussuser wrote: The linux unified kernel uses wine and reactos' codes, but its technique is different from the technique of wine.

Just like KVM, it uses a lot of XEN's codes, but the technique of kvm is different from XEN.

I guess that you meant QEMU instead of XEN.... KVM is QEMU based and XEN also leverages functionality from QEMU... KVM doesn't use a lot of XEN code.... (AFAIK none, or very little, but I could be wrong)

bussuser · Post by **bussuser** » Sat Mar 15, 2008 7:48 am

Phobos wrote:
bussuser wrote: The linux unified kernel uses wine and reactos' codes, but its technique is different from the technique of wine.

Just like KVM, it uses a lot of XEN's codes, but the technique of kvm is different from XEN.
I guess that you meant QEMU instead of XEN.... KVM is QEMU based and XEN also leverages functionality from QEMU... KVM doesn't use a lot of XEN code.... (AFAIK none, or very little, but I could be wrong)

Chinese. translated by google.com

http://translate.google.com/translate?u ... N&ie=UTF-8

the writers are intel's engineer. they made the kvm running on the non-x86 computer first.

He ( in charge of Intel's multi-core software Sunil Saxena) referred to Xen and KVM are developed by the open source community, the majority of the code is also KVM directly from the XEN in the transplant, KVM many developers have come from XEN project team, although the KVM and XEN are two different Virtual Machine.

Phobos · Post by **Phobos** » Sat Mar 15, 2008 9:34 am

bussuser wrote:Chinese. translated by google.com

http://translate.google.com/translate?u ... N&ie=UTF-8

the writers are intel's engineer. they made the kvm running on the non-x86 computer first.

He ( in charge of Intel's multi-core software Sunil Saxena) referred to Xen and KVM are developed by the open source community, the majority of the code is also KVM directly from the XEN in the transplant, KVM many developers have come from XEN project team, although the KVM and XEN are two different Virtual Machine.

what I read there is that some developers of KVM came from the XEN project...

read some white papers and documentation and you will see that there are fundamental differences in how both XEN and KVM do the same thing... these differences don't allow them to directly use the same code on both project.

KVM is an extension of QEMU, working as an accelerator like KQEMU, but hardware based (hence, it needs the hardware extensions to work)

XEN is an paravirtualization hypervisor and has an option to use QEMU devices (QEMU-DM) and hardware extensions to use unmodified OSs, like windows...

read Daniel Berrange's presentation about using QEMU in XEN more like KVM does to get a better idea of how XEN compares to KVM: xen.org/files/xensummit_fall07/10_DanielBerrange.pdf

and read also for example from http://linuxvirtualization.com/articles ... versus-kvm :

The only slight beef (or question) I have with KVM is why has it diverged from QEMU? More specifically KVM relies on QEMU for it’s device emulation but it also requires hardware virtualization. Now that QEMU’s kernel accelerator is open source could KVM and QEMU not be merged back together to give us one high performing solution that would support both VT and non-VT capable hardware?

Edward Savage · Post by **Edward Savage** » Sat Mar 15, 2008 9:52 am

I'm sorry but what does this have to do with wine at all?

Phobos · Post by **Phobos** » Sat Mar 15, 2008 11:14 am

just clearing up some stuff that was not correct...

James McKenzie · Post by **James McKenzie** » Sat Mar 15, 2008 12:25 pm

Edward Savage wrote:

I'm sorry but what does this have to do with wine at all?

This is a discussion about the Linux Unified Kernel and where the
developers came from. At last check, it would be a good thing to
discuss possible competing projects and where they differ from Wine.
The Linux Unified Kernel attempts to put some, but not all, of the
Windows API functions in the Linux Kernel. Bloat if I ever saw it, and
that is my opinion on the matter.

James

bussuser · Post by **bussuser** » Sun Apr 06, 2008 8:54 am

Has the wine project team pondered whether to move WINE into the kernel?

hendrik · Post by **hendrik** » Sun Apr 06, 2008 8:59 am

On Sun, Apr 06, 2008 at 08:55:00AM -0500, bussuser wrote:

Has the wine project team pondered whether to move WINE into the kernel?

In my opinion, there's too much in the kernel already. But I'm not part
of the wine project or part of the kernel team.

-- hendrik

bussuser · Post by **bussuser** » Sun Apr 06, 2008 9:44 am

hendrik wrote:On Sun, Apr 06, 2008 at 08:55:00AM -0500, bussuser wrote:
Has the wine project team pondered whether to move WINE into the kernel?

In my opinion, there's too much in the kernel already. But I'm not part
of the wine project or part of the kernel team.

-- hendrik

In my opinion, the linux kernel should have two version - server version and desktop version, just like ms windows. so the technique like as unified kernel can enter into the desktop linux kernel. it can help linux to have a big desktop market share .

David Gerard · Post by **David Gerard** » Sun Apr 06, 2008 9:57 am

On 06/04/2008, [email protected] <[email protected]> wrote:

On Sun, Apr 06, 2008 at 08:55:00AM -0500, bussuser wrote:

Has the wine project team pondered whether to move WINE into the kernel?

In my opinion, there's too much in the kernel already. But I'm not part
of the wine project or part of the kernel team.

Although I can't find links right now, I understand the Linux kernel
developers have indicated they are not averse to a Wine kernel module
for services that are sensible to put into the kernel. I expect this
work will not be happening unless and until someone takes a strong
interest in it, though. And there'd still need to be userland versions
for operating systems that aren't LGPL-compatible (e.g. OpenSolaris at
present).

- d.

Dan Kegel · Post by **Dan Kegel** » Sun Apr 06, 2008 9:57 am

On Sun, Apr 6, 2008 at 6:55 AM, bussuser <[email protected]> wrote:

Has the wine project team pondered whether to move WINE into the kernel?

Yes, it's been discussed many times.
It turns out it's not worth the effort yet.
There are many lower-hanging fruit.
- Dan

Jean-Michel Brünn · Post by **Jean-Michel Brünn** » Sun Apr 06, 2008 10:57 am

On Sun, Apr 6, 2008 at 6:55 AM, bussuser <[email protected]> wrote:

Has the wine project team pondered whether to move WINE into the kernel?

I don't know wether thats interesting or not as no one asked about opinions from users. In my humble opinion wine shouldn't go into the kernel. A better way (just my opinion) would be just using stuff added already into the kernel, or adding features to 'improve wine' into the kernel.

Things like KVM. The Kernel is offering more and more virtualization things. Wine should try to use them instead of trying to add wine to the kernel.

*shrugs*

Cheers
Jean

L. Rahyen · Post by **L. Rahyen** » Sun Apr 06, 2008 11:14 am

On Sunday April 6 2008 14:44:15 bussuser wrote:

hendrik wrote:
On Sun, Apr 06, 2008 at 08:55:00AM -0500, bussuser wrote:
Has the wine project team pondered whether to move WINE into the
kernel?
In my opinion, there's too much in the kernel already. But I'm not part
of the wine project or part of the kernel team.

-- hendrik
In my opinion, the linux kernel should have two version - server version
and desktop version, just like ms windows.

No, I do not see any reason to have different kernels for desktop and server.
If someone *really* needs special kernel he/she can and should compile it
from source because such need indicate very specific (non-standard)
requirements.
BTW, I'm using similar kernels on my desktop and two servers.

so the technique like as unified kernel can enter into the desktop linux
kernel. it can help linux to have a big desktop market share.

Why do you think that servers don't need WINE?
Real world example: I have collocated server which runs Windows software. It
is doing important financial work so if it fails, I can and probably will
lose some money. Therefore stability is really important issue - so Windows
of any version isn't an option here; Windows always crashes sooner or later
(yeah, I checked that multiple times), I cannot use it.
Fortunately WINE solves this problem - if Windows software is 100% stable
with it (which is easy to determine after some weeks or months of testing)
then I can be sure that it will work really well even for important
applications.
Performance is also important so if there is some way to improve it
somewhat - that's would be good.
However, I didn't actually tried "unified kernel" so I cannot
actually tell how useful it is. But at least I can tell if it is useful for
desktop then it can be useful for server too (if it really improves
performance of Windows applications). Of course if it isn't 100% stable then
it is useless.
Unfortunately so far I didn't find any benchmarks results which indicate that
it really improves performance and I didn't find information about its
stability; personally I don't have enough free time to test such
experimental thing myself.
For now I have some doubts about its practical usefulness (lack of good
documentation, no real-world benchmark results which proves usefulness, no
sync with current kernel and WINE versions, and other problems).