RE: [coLinux-devel] private networks between colinux and windows

[Ben V. <bv...@ka...>]

See if your lockups aren't xorg-x11.

For an internal connection, you have two options:

TAP/TUN in windows.  The TAP looks like a NIC to Windows; you can configure
TCP/IP etc and set routes as you wish.  Colinux's TAP network implements the
TUN side in windows, and passes traffic to/from colinux.

WinPCAP bridging in windows.  This grabs traffic on a separate driver (i.e.
Microsoft Loopback), filters out traffic the colinux host should see (by MAC
address?) and passes traffic to/from colinux.

For an internal network, TAP should be better since it is designed for this
purpose (making a user process look like an additional NIC to windows).
Performance results on this list show TAP is faster.

The WinPCAP has the benefit of allowing traffic outside the Windows host b/c
it interacts with existing traffic on an existing NIC.  The TAP driver can't
bridge, but you can make it part of a higher layer bridge, like Windows XP
bridge support.  I run Windows 2000 so I use WinPCAP.

I'm not sure exactly what SLiRP does or how it works.  I'm guessing it's a
sockets proxy like SOCKS where the colinux can request an outgoing
connection which is tunnelled to it, without ever being able to listen?

Another theoretical option would be to merge both IP stacks.  Each socket on
the colinux side would have a one-to-one correspondence with one on the
Windows side.  All sockets calls would have to be passed through and handled
by Windows.  But then colinux could open listening ports inside Windows and
linux daemons and Windows services would be indistinguishable from the
outside (same IP address, same broken Windows TCP/IP connection behavior).
This is of effectively what cygwin does -- it provides a libc that calls the
Windows APIs.  But with cygwin the executable itself is in Windows PE
format, and is scheduled and managed by the MS VMM (virtual memory manager).
With colinux a hybrid is possible, with linux VMM and Windows I/O.  That's
actually exactly what the cofs file system does, right?

****************************

Ben Voigt
University of Pennsylvania
Electrical Engineering PhD Candidate

vo...@se... <mailto:vo...@se...>
BV...@ka... <mailto:BV...@ka...>

Support a Constitutional Amendment to protect the Pledge of Allegiance and
National Motto.
Click here for more information. <http://www.wepledge.com/>

****************************


> -----Original Message-----
> From: col...@li...
> [mailto:col...@li...]On Behalf Of Eric S.
> Johansson
> Sent: Saturday, March 19, 2005 2:21 PM
> To: Ian Bonnycastle
> Cc: Dave Hylands; col...@li...
> Subject: Re: [coLinux-devel] private networks between colinux and
> windows
>
>
> Ian Bonnycastle wrote:
> > Ok, I've been reading these threads for a while, and I really don't
> > see the difference between the Microsoft Loopback Adapter
> and the TAP
> > adapter. Both give you direct access between your host system and
> > coLinux, whether or not the external interface is plugged in or not.
> > Why use one over the other?
>
> good point. let me try to to use and/or clarify things
>
>           |==================|
> ------|xx| colinux guest    |
>           |===============-==|
>                           x
>                           x
>                           -
>                           |
>                           |
>                           -
>                           x
>                           x
>           |===============-==|
> ------|xx| windows host     |
>           |==================|
>
> xx represents an ethernet interface
>
> here I'm trying to represent schematically a guest and host system.
> Each system has two ethernet interfaces.  the ones on the
> left-hand side
> are the public interfaces.  These are the interfaces you use
> to get to
> the local area network or the Internet.
>
> the two interfaces represented by vertically aligned X's are the
> internal or private connection between the guest and the
> Windows host.
> the setting up of the public interfaces is well understood.  But the
> method of adding a second pair of interfaces is not clear to me.
>
> Now if adding a loopback interface will work and making it
> visible via
> some variant on  <network index="1" type="tap" name="mystery date"/>
> will work, and this is much goodness and should be documented.  (yes,
> yes I will write up what I find.)
>
> now I can see making this work on the linux side with multiple tap
> interfaces but I don't fully comprehend the loopback because I would
> think it would just loop back.  That is unless Microsoft has applied
> their usual clarity of naming to the loopback tool.
>
> so I guess I'm still trying to figure out how to do the
> equivalent of a
> virtual ethernet device on the Windows side of the house.  Maybe I
> should just try loopback and see just what it does.  Just as
> soon as I
> figure out what the hell is locking up my machine when I try
> to emerge
> updates for the gentoo image.
>
> ---eric
>
> --
> http://www.wired.com/wired/archive/13.03/view.html?pg=5
>
> The result of the duopoly that currently defines "competition" is that
> prices and service suck. We're the world's leader in Internet
> technology - except that we're not.
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from
> real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> coLinux-devel mailing list
> coL...@li...
> https://lists.sourceforge.net/lists/listinfo/colinux-devel
> --
> No virus found in this incoming message.
> Checked by AVG Anti-Virus.
> Version: 7.0.308 / Virus Database: 266.7.4 - Release Date: 3/18/2005
>