colinux-users — coLinux users mailing list

Re: [coLinux-users] Steps to redirect ports in SLIRP mode ?

[Henry N. <Hen...@Ar...>]

Frédéric L. W. Meunier schrieb:
> On Jan 23, 2008 6:53 PM, Henry Nestler <Hen...@ar...> wrote:
>> Frédéric L. W. Meunier wrote:
>>> On Wed, 23 Jan 2008, Henry Nestler wrote:
>>>
>>> Everything you both listed is right, but http://localhost/
>>> doesn't work from Windows. It does only if I run a native
>>> Windows server.
>> I don't know your difference you means.
>> Windows != Windows Server ?
>> If not native Windows? What else is it?
>>
> 
> I meant Windows = XP Professional and Windows server = a Windows HTTP server.
> 
>>> In TCPView, I see:
>>>
>>> colinux-slirp-net-daemon.exe TCP PERVALIDUS:80 PERVALIDUS:0
>>> LISTENING
>> Please disable DNS Name Resolution on TCPview. It should be listen on
>> 0.0.0.0 (every networks)
>>
> 
> It returns 0.0.0.0.
> 
>> Try other port for remapping, for example
>>   tcp:2222:22/tcp:4000:80
>> Than on the windows side the port 4000 is your webserver and port 80
>> inside coLinux. Type the URL http://localhost:4000/ in your Windows browser.
> 
> Remapping, both work.Now I really don't have any idea why using the
> same ports doesn't.

Only one program can agent as server on one tcp port.
If you have a running program that is listen on port 80 (windows web
Server ?), than an other program can not open the same port for listen.
Or ir can open, but windows is faster and catch all the incommings. I
don't know.

Has your windows some more protections to not open ports lower than 1024
from a user program as server? Some firewalls or secure tools forbidden
and and blocks it (Zone Alarm, Norton, ...).

For example on Linux an user land program can only listen on ports
>1023. On ports 22 and 80 only a user "root" can listen.

-- 
Henry N.

Re: [coLinux-users] Steps to redirect ports in SLIRP mode ?

[<lis...@pe...>]

On Jan 23, 2008 6:53 PM, Henry Nestler <Hen...@ar...> wrote:
>
> Fr=E9d=E9ric L. W. Meunier wrote:
> > On Wed, 23 Jan 2008, Henry Nestler wrote:
> >
> > Everything you both listed is right, but http://localhost/
> > doesn't work from Windows. It does only if I run a native
> > Windows server.
>
> I don't know your difference you means.
> Windows !=3D Windows Server ?
> If not native Windows? What else is it?
>

I meant Windows =3D XP Professional and Windows server =3D a Windows HTTP s=
erver.

> > In TCPView, I see:
> >
> > colinux-slirp-net-daemon.exe TCP PERVALIDUS:80 PERVALIDUS:0
> > LISTENING
>
> Please disable DNS Name Resolution on TCPview. It should be listen on
> 0.0.0.0 (every networks)
>

It returns 0.0.0.0.

> Try other port for remapping, for example
>   tcp:2222:22/tcp:4000:80
> Than on the windows side the port 4000 is your webserver and port 80
> inside coLinux. Type the URL http://localhost:4000/ in your Windows brows=
er.

Remapping, both work.Now I really don't have any idea why using the
same ports doesn't.

> > I don't have any firewall running on Windows, but added
> > colinux-slirp-net-daemon.exe to the list of exceptions in
> > Windows Firewall.
>
> If you have no firewall on windows, why you can add exceptions to
> firewall? Try complete disable the windoes internal firewall. If you not
> would risk some, you can disconnect your networkcable for this testing
> time. Slirp is usable from windows to coLinux and via versa without any
> external network.

Windows Firewall is "off", but even with it disabled, the "Exceptions"
tab is enabled. Just for safety, I had added
colinux-slirp-net-daemon.exe to the list.

> Check: We talk about coLinux 0.7.1, 0.7.2 or 0.8.x? Not older, please.

0.8.0-20071105.

Re: [coLinux-users] Steps to redirect ports in SLIRP mode ?

[Henry N. <Hen...@Ar...>]

Frédéric L. W. Meunier wrote:
> On Wed, 23 Jan 2008, Henry Nestler wrote:
> 
>> Josep Ma [JAZ] wrote:
>>> As you say, you have to use a syntax like 
>>> "eth0=slirp,,tcp:22:22/tcp:80:80" to open ports from colinux in slirp mode.
>>>
>>> Let's verify a few steps:
>>>
>>> How have you configured the eth in linux?
>>> you know you have to use the IP 10.0.2.15/24 , the gateway 10.0.2.2, and 
>>> the DNS server 10.0.2.3, do you? (Configuring it with DHCP should also 
>>> work).
>> Be shure to have netmask 255.255.255.0
>> All other, or default generated netmask for ip 10.0.2.15 dos't work.
>>
>>> Do you get to internet from colinux?
>>> Ping doesn't work with slirp, but you can try telnet, ssh, wget....
>>>
>>> Now, if that works, how do you try to connect to colinux from windows? 
>>> Slirp connection makes colinux act as if it was the windows machine. 
>>> This means that you have to connect to localhost. Example: 
>>> http://localhost/ should connect to an http server in colinux, if using 
>>> the above eth0 configuration. (That is, of course, if there port isn't 
>>> already used in windows).
>>>
>>> If that still doesn't work, you would need to be more specific as to 
>>> which configuration are you using and how do you try to use it.
>> Yes. Plesase check the steps from one to other network step by step.
>> First inside coLinux check the running webserver, for example
>> $ cd /tmp
>> $ wget http://localhost/
>> Would get your index.htm without errors.
>> $ wget http://10.0.2.15/
>> Should do the same without error.
>>
>> Go to windows side and open your browser with url
>> http://localhost/
>> Should get your index.htm on windows.
>> If not, check your firewall. Disable firewall temporaly, shutdown and
>> start colinux again.
>>
>> If that not helps, check the running processes inside taskman.
>> colinux-slirp-net-daemon.exe must run and should have the parameters
>> "tcp:22:22/tcp:80:80". If taskman dos't give you the parameters, use the
>> ProcessExplorer from http://www.sysinternals.com (redirection to MS).
>>
>> After the process parameter are ok, than check the listen port on tcp.
>> You can do it also with ProcessExplorer or with tcpView.exe (from
>> http://www.sysinternals.com). YOu should see an listen 0.0.0.0 on port
>> 22 and 80.
> 
> Everything you both listed is right, but http://localhost/ 
> doesn't work from Windows. It does only if I run a native 
> Windows server.

I don't know your difference you means.
Windows != Windows Server ?
If not native Windows? What else is it?

> In TCPView, I see:
> 
> colinux-slirp-net-daemon.exe TCP PERVALIDUS:80 PERVALIDUS:0 
> LISTENING

Please disable DNS Name Resolution on TCPview. It should be listen on
0.0.0.0 (every networks)

Check: The first lines from Windows command NETSTAT.EXE /a /n (I don't
know the paremters right). The IP 0.0.0.0 on port 80 and 22 should be
"listen". Think it is, TCPview shows the same.

Check: Stop coLinux. The port 80 and 22 should no longer listen.

Try other port for remapping, for example
  tcp:2222:22/tcp:4000:80
Than on the windows side the port 4000 is your webserver and port 80
inside coLinux. Type the URL http://localhost:4000/ in your Windows browser.

> I don't have any firewall running on Windows, but added 
> colinux-slirp-net-daemon.exe to the list of exceptions in 
> Windows Firewall.

If you have no firewall on windows, why you can add exceptions to
firewall? Try complete disable the windoes internal firewall. If you not
would risk some, you can disconnect your networkcable for this testing
time. Slirp is usable from windows to coLinux and via versa without any
external network.

Check: We talk about coLinux 0.7.1, 0.7.2 or 0.8.x? Not older, please.

> On Linux there's iptables, but the same rules 
> work on Linux.
> 
> netstat on coLinux shows:
> 
> tcp        0      0 0.0.0.0:80            0.0.0.0:*               LISTEN

Ok, let coLinux config it untouched.

-- 
Henry N.

Re: [coLinux-users] Steps to redirect ports in SLIRP mode ?

[<2...@pe...>]

On Wed, 23 Jan 2008, Henry Nestler wrote:

> Josep Ma [JAZ] wrote:
>> As you say, you have to use a syntax like 
>> "eth0=slirp,,tcp:22:22/tcp:80:80" to open ports from colinux in slirp mode.
>> 
>> Let's verify a few steps:
>> 
>> How have you configured the eth in linux?
>> you know you have to use the IP 10.0.2.15/24 , the gateway 10.0.2.2, and 
>> the DNS server 10.0.2.3, do you? (Configuring it with DHCP should also 
>> work).
>
> Be shure to have netmask 255.255.255.0
> All other, or default generated netmask for ip 10.0.2.15 dos't work.
>
>> Do you get to internet from colinux?
>> Ping doesn't work with slirp, but you can try telnet, ssh, wget....
>> 
>> Now, if that works, how do you try to connect to colinux from windows? 
>> Slirp connection makes colinux act as if it was the windows machine. 
>> This means that you have to connect to localhost. Example: 
>> http://localhost/ should connect to an http server in colinux, if using 
>> the above eth0 configuration. (That is, of course, if there port isn't 
>> already used in windows).
>> 
>> If that still doesn't work, you would need to be more specific as to 
>> which configuration are you using and how do you try to use it.
>
> Yes. Plesase check the steps from one to other network step by step.
> First inside coLinux check the running webserver, for example
> $ cd /tmp
> $ wget http://localhost/
> Would get your index.htm without errors.
> $ wget http://10.0.2.15/
> Should do the same without error.
>
> Go to windows side and open your browser with url
> http://localhost/
> Should get your index.htm on windows.
> If not, check your firewall. Disable firewall temporaly, shutdown and
> start colinux again.
>
> If that not helps, check the running processes inside taskman.
> colinux-slirp-net-daemon.exe must run and should have the parameters
> "tcp:22:22/tcp:80:80". If taskman dos't give you the parameters, use the
> ProcessExplorer from http://www.sysinternals.com (redirection to MS).
>
> After the process parameter are ok, than check the listen port on tcp.
> You can do it also with ProcessExplorer or with tcpView.exe (from
> http://www.sysinternals.com). YOu should see an listen 0.0.0.0 on port
> 22 and 80.

Everything you both listed is right, but http://localhost/ 
doesn't work from Windows. It does only if I run a native 
Windows server.

In TCPView, I see:

colinux-slirp-net-daemon.exe TCP PERVALIDUS:80 PERVALIDUS:0 
LISTENING

and

firefox.exe TCP PERVALIDUS:3819 localhost:80 FIN_WAIT2
firefox.exe TCP PERVALIDUS:3819 localhost:80 ESTABLISHED
...

when I try to connect from Firefox, but it stays at 
"Loading...".

I don't have any firewall running on Windows, but added 
colinux-slirp-net-daemon.exe to the list of exceptions in 
Windows Firewall. On Linux there's iptables, but the same rules 
work on Linux.

netstat on coLinux shows:

tcp        0      0 0.0.0.0:80            0.0.0.0:*               LISTEN

Re: [coLinux-users] Steps to redirect ports in SLIRP mode ?

[Henry N. <Hen...@Ar...>]

Josep Ma [JAZ] wrote:
> As you say, you have to use a syntax like 
> "eth0=slirp,,tcp:22:22/tcp:80:80" to open ports from colinux in slirp mode.
> 
> Let's verify a few steps:
> 
> How have you configured the eth in linux?
> you know you have to use the IP 10.0.2.15/24 , the gateway 10.0.2.2, and 
> the DNS server 10.0.2.3, do you? (Configuring it with DHCP should also 
> work).

Be shure to have netmask 255.255.255.0
All other, or default generated netmask for ip 10.0.2.15 dos't work.

> Do you get to internet from colinux?
> Ping doesn't work with slirp, but you can try telnet, ssh, wget....
> 
> Now, if that works, how do you try to connect to colinux from windows? 
> Slirp connection makes colinux act as if it was the windows machine. 
> This means that you have to connect to localhost. Example: 
> http://localhost/ should connect to an http server in colinux, if using 
> the above eth0 configuration. (That is, of course, if there port isn't 
> already used in windows).
> 
> If that still doesn't work, you would need to be more specific as to 
> which configuration are you using and how do you try to use it.

Yes. Plesase check the steps from one to other network step by step.
First inside coLinux check the running webserver, for example
$ cd /tmp
$ wget http://localhost/
Would get your index.htm without errors.
$ wget http://10.0.2.15/
Should do the same without error.

Go to windows side and open your browser with url
http://localhost/
Should get your index.htm on windows.
If not, check your firewall. Disable firewall temporaly, shutdown and
start colinux again.

If that not helps, check the running processes inside taskman.
colinux-slirp-net-daemon.exe must run and should have the parameters
"tcp:22:22/tcp:80:80". If taskman dos't give you the parameters, use the
ProcessExplorer from http://www.sysinternals.com (redirection to MS).

After the process parameter are ok, than check the listen port on tcp.
You can do it also with ProcessExplorer or with tcpView.exe (from
http://www.sysinternals.com). YOu should see an listen 0.0.0.0 on port
22 and 80.

-- 
Henry N.

Re: [coLinux-users] Steps to redirect ports in SLIRP mode ?

[Josep Ma [JAZ] <jo...@gm...>]

As you say, you have to use a syntax like 
"eth0=slirp,,tcp:22:22/tcp:80:80" to open ports from colinux in slirp mode.

Let's verify a few steps:

How have you configured the eth in linux?
you know you have to use the IP 10.0.2.15/24 , the gateway 10.0.2.2, and 
the DNS server 10.0.2.3, do you? (Configuring it with DHCP should also 
work).

Do you get to internet from colinux?
Ping doesn't work with slirp, but you can try telnet, ssh, wget....

Now, if that works, how do you try to connect to colinux from windows? 
Slirp connection makes colinux act as if it was the windows machine. 
This means that you have to connect to localhost. Example: 
http://localhost/ should connect to an http server in colinux, if using 
the above eth0 configuration. (That is, of course, if there port isn't 
already used in windows).

If that still doesn't work, you would need to be more specific as to 
which configuration are you using and how do you try to use it.

[coLinux-users] Steps to redirect ports in SLIRP mode ?

[<2...@pe...>]

I'm really lost trying to get some services accessible from 
coLinux in SLIRP mode.

>From what I've read, you only need to add something like:

eth0=slirp,,tcp:22:22/tcp:80:80

in your configuration file and it's done.

The problem is that I can't access such services from Windows 
or a remote machine, only from coLinux.

If I run a server on Windows, I can access it from coLinux and 
remotely.

Any hints ? I'm using 0.8.0-20071105.

[coLinux-users] Cofs problem

[nano b. <nan...@ho...>]

Hello to everyone ,

I recently play with colinux and I've a problem.

I have a network disk on windows I would like to access under colinux and I=
 use cofs to do this. The problem is that when I'm creating a new file with=
 colinux on this disk I can't list it :

# touch homelnx/coucou
root@colinux:~# ls homelnx/   =20
root@colinux:~# ls -l homelnx/coucou
-rwxr-xr-x 1 root root 0 Jan 17 15:54 homelnx/coucou

Thanks for your help

Baptiste


_________________________________________________________________
Votre contact a choisi Hotmail, l'e-mail nouvelle g=E9n=E9ration. Cr=E9ez u=
n compte.=20
http://www.windowslive.fr/hotmail/default.asp=

Re: [coLinux-users] Loadable kernel modules

[Henry N. <Hen...@Ar...>]

jitubhai ambani wrote:
> Hi,
> I use colinux version 0.8.0 on linux as a host.
> I was able to insert a simplistic kernel module in the guest.
> 
> However, when I include linux/cooperative_internal.h it gives me
> errors.  following is a simple kernel module that dosen't work
> 
> #include <linux/kernel.h> /* Needed for KERN_ALERT */
> #include <linux/module.h> /* Needed by all modules */
> #include <linux/unistd.h>
> #include <linux/cooperative_internal.h>
> #include <linux/string.h>
> 
> 
> int init_module(void)
> {
>     unsigned long flags;
>     co_passage_page_assert_valid();
>     co_passage_page_acquire(&flags);
>     co_passage_page->operation = CO_OPERATION_PRINTK;
>     co_passage_page->params[0] = strlen("We did this!");
>     strcpy(co_passage_page->params[1], "We did this!");
>     co_switch_wrapper();
>     co_passage_page_release(flags);
>     printk("Hello world 1.\n");
>     return 0;
> }
> void cleanup_module(void)
> {
>     printk("Goodbye world 1.\n");
> }
> 
> 
> the errors we get are as follows :
> 
> make -C /lib/modules/2.6.15-co-0.8.0/build M=/root modules
> make[1]: Entering directory `/root/linux-2.6.15'
>   CC [M]  /root/hello.o
> /root/hello.c: In function `init_module':
> /root/hello.c:18: warning: passing arg 1 of `strcpy' makes pointer from
> integer without a cast
>   Building modules, stage 2.
>   MODPOST
> *** Warning: "co_passage_page_release" [/root/hello.ko] undefined!
> *** Warning: "co_switch_wrapper" [/root/hello.ko] undefined!
> *** Warning: "co_passage_page_acquire" [/root/hello.ko] undefined!
> *** Warning: "co_passage_page_held" [/root/hello.ko] undefined!
>   CC      /root/hello.mod.o
>   LD [M]  /root/hello.ko
> make[1]: Leaving directory `/root/linux-2.6.15'
> 
> 
> Why can't it find the functions if linux/cooperative_internal.h is
> included?... do i have to insert this module in the host kernel??
> (Another observation.. don't know how relevant it is in the context..
> but anyways, all symbols NOT defined extern in cooperative_internal.h
> work fine.. all declared extern don't)

All colinux internal functions and variables are not exported. They
would be very sensitive for using in any other way. You risk crashing
your host with smalles typofixies there!

For example your strcpy to wrong Pointer, that you have ;-)
-strcpy(co_passage_page->params[1], "We did this!");
+strcpy(&co_passage_page->params[1], "We did this!");

External modues should never use colinux_internal header. That's why is
named "internal".

To have access to variables or functions from an external mudule, you
needs to "EXPORT_SYMBOL" every variable or function *ones* in the file,
where the variable or function was declared, mostly of they are in
arch/i386/kernel/cooperative.c.

-- 
Henry N.

[coLinux-users] Loadable kernel modules

[jitubhai a. <jit...@re...>]

Hi,=0AI use colinux version 0.8.0 on linux as a host.=0AI was able to inser=
t a simplistic kernel module in the guest.=0A=0AHowever, when I include lin=
ux/cooperative_internal.h it gives me errors.  following is a simple kernel=
 module that dosen't work =0A=0A#include <linux/kernel.h> /* Needed for KER=
N_ALERT */=0A#include <linux/module.h> /* Needed by all modules */=0A#inclu=
de <linux/unistd.h>=0A#include <linux/cooperative_internal.h>=0A#include <l=
inux/string.h>=0A=0A=0Aint init_module(void)=0A{=0A    unsigned long flags;=
=0A    co_passage_page_assert_valid();=0A    co_passage_page_acquire(&flags=
);=0A    co_passage_page->operation =3D CO_OPERATION_PRINTK;=0A    co_passa=
ge_page->params[0] =3D strlen("We did this!");=0A    strcpy(co_passage_page=
->params[1], "We did this!");=0A    co_switch_wrapper();=0A    co_passage_p=
age_release(flags);=0A    printk("Hello world 1.\n");=0A    return 0;=0A}=
=0Avoid cleanup_module(void)=0A{=0A    printk("Goodbye world 1.\n");=0A}=0A=
 =0A=0Athe errors we get are as follows :=0A=0Amake -C /lib/modules/2.6.15-=
co-0.8.0/build M=3D/root modules=0Amake[1]: Entering directory `/root/linux=
-2.6.15'=0A  CC [M]  /root/hello.o=0A/root/hello.c: In function `init_modul=
e':=0A/root/hello.c:18: warning: passing arg 1 of `strcpy' makes pointer fr=
om integer without a cast=0A  Building modules, stage 2.=0A  MODPOST=0A*** =
Warning: "co_passage_page_release" [/root/hello.ko] undefined!=0A*** Warnin=
g: "co_switch_wrapper" [/root/hello.ko] undefined!=0A*** Warning: "co_passa=
ge_page_acquire" [/root/hello.ko] undefined!=0A*** Warning: "co_passage_pag=
e_held" [/root/hello.ko] undefined!=0A  CC      /root/hello.mod.o=0A  LD [M=
]  /root/hello.ko=0Amake[1]: Leaving directory `/root/linux-2.6.15'=0A=0A=
=0AWhy can't it find the functions if linux/cooperative_internal.h is inclu=
ded?... do i have to insert this module in the host kernel??=0A(Another obs=
ervation.. don't know how relevant it is in the context.. but anyways, all =
symbols NOT defined extern in cooperative_internal.h work fine.. all declar=
ed extern don't)

Re: [coLinux-users] Starting internet on colinux

[Henry N. <Hen...@Ar...>]

Vedang wrote:
> I am running colinux (version 0.8.0) with a debian image on linux
> kernel 2.6.15 (guest and host).

Be carefully with kernel 2.6.15 as guest. This have only backports. Some
functions are not full implemented, as they are in kernel 2.6.22

> Can somebody please tell how to start internet on colinux? I passed
> boot parameter eth0=slirp
> 
> ifconfig gives network address : 192.168.0.40 for eth0 which is statically set.

http://colinux.wikia.com/wiki/Network#Reconfigure_Network_Interface_On_the_Fly

With this commands you "Reconfigure Network Interface On the Fly":
# ifconfig eth0 10.0.2.15 netmask 255.255.255.0 broadcast 10.0.2.255
# route add default gw 10.0.2.2
# echo "nameserver 10.0.2.3" > /etc/resolv.conf

To set the config permanently follow the "Example: Typical Slirp
Configuration" in http://colinux.wikia.com/wiki/Network#Debian and write
into file /etc/network/interfaces:

auto eth0
iface eth0 inet static
  address 10.0.2.15
  network 10.0.2.0
  broadcast 10.0.2.255
  netmask 255.255.255.0
  gateway 10.0.2.2

-- 
Henry N.

[coLinux-users] Starting internet on colinux

[Vedang - 1. u|3. h. <ved...@gm...>]

Hi,
I am running colinux (version 0.8.0) with a debian image on linux
kernel 2.6.15 (guest and host).
Can somebody please tell how to start internet on colinux? I passed
boot parameter eth0=slirp

ifconfig gives network address : 192.168.0.40 for eth0 which is statically set.

-- 
Solaris is simple. It takes a genius to understand it's simplicity.
                                    - Anon
People think it must be fun to be a super genius, but they don't
realize how hard it is to put up with all the idiots in the world.
                                     - Calvin.
Cheers,
Vedang.

B.E. Computer Engineering,
P.I.C.T Pune.
http://blogs.sun.com/vedang
http://dose-of-wisdom.blogspot.com
http://tech-rantings.blogspot.com

Re: [coLinux-users] gcc/kernel modules on colinux

[Holger K. <hol...@gm...>]

chirayu shah schrieb:
> I am using colinux 0.8.0 with debian image on a compiled vanilla kernel
> (v 2.6.15), and I cannot seem to find gcc on  it. How do I install gcc
> on coLinux?

Howto install gcc depends on your distro. If you use debian "apt-get install gcc" will do.

[coLinux-users] gcc/kernel modules on colinux

[chirayu s. <chi...@gm...>]

Hi,

I am using colinux 0.8.0 with debian image on a compiled vanilla kernel (v
2.6.15), and I cannot seem to find gcc on  it. How do I install gcc on
coLinux? I wish to write my own kernel modules. Does coLinux support
insertion of external modules?

Any help would be greatly appreciated.
Thank you.

Re: [coLinux-users] Colinux crash on insmod colinux.ko

[jitubhai a. <jit...@re...>]

  =0Athnx a lot!!=0Afinally got colinux running!!... it is indeed impressiv=
e!!.. even multiple instances run as smoothly as anything!=0A=0A(lol.. and =
yeah... this time it IS reply all :D)=0A=0A=0A=0AOn Thu, 10 Jan 2008 Henry =
Nestler wrote :=0A>jitubhai ambani wrote:=0A> > I am using colinux 0.6.4 (I=
 know about the newer version, but cannot=0A> > switch to 0.7.1 for some re=
asons)=0A> >=0A> > I have suse-9.3.. i recompiled vanilla kernel 2.6.11 to =
work as host..=0A> > and using patched 2.6.11 to work as guest.. i've insta=
lled patched=0A> > fltk-1.4.4 and mxml-1.2 ... i used gcc 3.4.2=0A>=0A>PS: =
Host and guest kernel version have no dependency. Would only say here.=0A>=
=0A> > I made vmlinux (make vmlinux) in the patched kernel.. and follwed th=
e=0A> > stpes in docs/building (had to change stricmp() in config.c to=0A> =
> strcasecmp() )... but in the step insmod colinux.ko, my PC either hangs=
=0A> > or restarts..=0A> >=0A> > what are the reasons for this??.. and how =
can i resolve the issue??=0A>=0A>Grep in the SVN changelog with "Linux as h=
ost" to see all changes=0A>between 0.6.4 and 0.7.2 (eq 0.8.0).=0A>=0A>You c=
an not run the host Linux with CONFIG_REGPARM. This problems exist=0A>for a=
ll hostkernels >=3D 2.6.19 and coLinux version before 0.7.2. (see SVN=0A>re=
vision 727, 536)=0A>=0A>DEP/boexecute can be a problem on Intel CPU.=0A>=0A=
>On host with 4GB or more RAM needs to fix as it is in SVN rev 690.=0A>=0A>=
SVN rev 565: co_os_userspace_map=0A>SVN rev 561: Kernel 2.6.19 (and later) =
as linux host is broken currently=0A>=0A>Perhaps many more problems exist. =
I'm afraid that depends on your host=0A>config. I have tested vanilla kerne=
l 2.6.11 as host (with very=0A>minimalistic config and no acceloration for =
CPU, simple i486). So, I can=0A>say it was generaly usable.=0A>=0A>If you w=
ont to find the bug for crashing, you can insert some more=0A>printk("%s(%d=
)\n", __FILE__, __LINE__) and sleep(5 sec) in the=0A>colinux_module_init() =
inside src/colinux/os/linux/kernel/module/main.c=0A>and the followed subfun=
ctions. Than open a "tail  -f /var/log/messages",=0A>to see on what positio=
n it hangs while module loading.=0A>=0A>CoLinux on a Linux as host is more =
for developing coLinux, not for=0A>using. This port has not good tested and=
 is not full supported.=0A>You have two better ways for running a coLinux-I=
mage inside Linux:=0A>Simple use "chroot" into a loop mounted image. Or run=
 it as UML.=0A>=0A>--=0A>Henry N.=0A

Re: [coLinux-users] Colinux crash on insmod colinux.ko

[Henry N. <Hen...@Ar...>]

jitubhai ambani wrote:
>  
> thanks a lot!!
> I recompiled my host with CONFIG_REGPARM off and the module inserted all
> right...

:-)

> but.. how do i run the daemon (with linux as a host)
> what do i pass for the following parameters:
> cobd0=
> root=

Configure is the same way as under Windows:
  cobd0=root.fs
  root=/dev/cobd0

> if someone can send me a sample configuration file for linux as a host,
> it'll be really great...

[ For 'someone' should mail "answer all", not send to me only. ;-) ]


This is my debian.cfg under Linux as host:
>>> cat debian.cfg >>>
kernel=vmlinux
# mem=200
# initrd=initrd.gz
cobd0=Debian-3.0r0.ext3.100mb.img
cobd1=/tmp/swap256m
eth0=slirp
eth1=tuntap,host0
root=/dev/cobd0
<<< end <<<


This is my start script under Linux with network:
>>> cat run.sh >>>
export COLINUX_CONSOLE_FONT=:12
export LD_LIBRARY_PATH=$HOME/colinux/mingw32/i686-co-linux/lib
export PATH=.:$PATH

if [ ! -f /tmp/swap256m ]
then
   dd if=/dev/zero of=/tmp/swap256m bs=1k count=256 || exit 1
   mkswap /tmp/swap256m || exit 1
   chown hn.users /tmp/swap256m
fi

sync ; sleep 1
echo "Install colinux module now"
insmod colinux.ko

colinux-daemon @debian.cfg &
PID=$!

# Waits spawning net daemon, before config host side of net
sleep 3
ifconfig host0 192.168.200.1 netmask 255.255.255.0
echo "coLinux PID=$PID"
wait $PID
<<< end <<<

Please read also current doc/building, section "For Linux as host" in
http://colinux.wikia.com/wiki/BuildingColinux0.8.0  and
http://colinux.svn.sourceforge.net/svnroot/colinux/branches/devel/doc/building

-- 
Henry N.

Re: [coLinux-users] Colinux crash on insmod colinux.ko

[Henry N. <Hen...@Ar...>]

jitubhai ambani wrote:
> I am using colinux 0.6.4 (I know about the newer version, but cannot
> switch to 0.7.1 for some reasons)
> 
> I have suse-9.3.. i recompiled vanilla kernel 2.6.11 to work as host..
> and using patched 2.6.11 to work as guest.. i've installed patched
> fltk-1.4.4 and mxml-1.2 ... i used gcc 3.4.2

PS: Host and guest kernel version have no dependency. Would only say here.

> I made vmlinux (make vmlinux) in the patched kernel.. and follwed the
> stpes in docs/building (had to change stricmp() in config.c to
> strcasecmp() )... but in the step insmod colinux.ko, my PC either hangs
> or restarts..
> 
> what are the reasons for this??.. and how can i resolve the issue??

Grep in the SVN changelog with "Linux as host" to see all changes
between 0.6.4 and 0.7.2 (eq 0.8.0).

You can not run the host Linux with CONFIG_REGPARM. This problems exist
for all hostkernels >= 2.6.19 and coLinux version before 0.7.2. (see SVN
revision 727, 536)

DEP/boexecute can be a problem on Intel CPU.

On host with 4GB or more RAM needs to fix as it is in SVN rev 690.

SVN rev 565: co_os_userspace_map
SVN rev 561: Kernel 2.6.19 (and later) as linux host is broken currently

Perhaps many more problems exist. I'm afraid that depends on your host
config. I have tested vanilla kernel 2.6.11 as host (with very
minimalistic config and no acceloration for CPU, simple i486). So, I can
say it was generaly usable.

If you wont to find the bug for crashing, you can insert some more
printk("%s(%d)\n", __FILE__, __LINE__) and sleep(5 sec) in the
colinux_module_init() inside src/colinux/os/linux/kernel/module/main.c
and the followed subfunctions. Than open a "tail  -f /var/log/messages",
to see on what position it hangs while module loading.

CoLinux on a Linux as host is more for developing coLinux, not for
using. This port has not good tested and is not full supported.
You have two better ways for running a coLinux-Image inside Linux:
Simple use "chroot" into a loop mounted image. Or run it as UML.

-- 
Henry N.

[coLinux-users] Colinux crash on insmod colinux.ko

[jitubhai a. <jit...@re...>]

  =0Ahi!=0AI am using colinux 0.6.4 (I know about the newer version, but ca=
nnot switch to 0.7.1 for some reasons)=0A=0AI have suse-9.3.. i recompiled =
vanilla kernel 2.6.11 to work as host.. and using patched 2.6.11 to work as=
 guest.. i've installed patched fltk-1.4.4 and mxml-1.2 ... i used gcc 3.4.=
2=0A=0AI made vmlinux (make vmlinux) in the patched kernel.. and follwed th=
e stpes in docs/building (had to change stricmp() in config.c to strcasecmp=
() )... but in the step insmod colinux.ko, my PC either hangs or restarts..=
=0A=0Awhat are the reasons for this??.. and how can i resolve the issue??=
=0A=0Athanks in advance,=0Aregards,=0Ajitesh

Re: [coLinux-users] interface renumbering?

[<Use...@zo...>]

Hello

after upgrading winpcap  from 4.0.1 to 4.0.2 it works!


conet-bridged-daemon: Looking for interface "LAN-ADM851X"
conet-bridged-daemon: checking connection: LAN-TAP V9
conet-bridged-daemon: checking connection: LAN-TAP V8(colinux)
conet-bridged-daemon: checking connection: LAN-ADM851X
conet-bridged-daemon: listening on: ADMtek 851X Series AdapterNDIS 5.0 Miniport Driver...
conet-bridged-daemon: listening for: (ether dst 0a:c0:71:65:08:00) or (ether broadcast or multicast) or (ip broadcast or multicast)



winpcap 4.0.1

>Compiled on Dec 22 2007 05:10:09

>conet-bridged-daemon: Looking for interface "LAN-ADM851X"
>conet-bridged-daemon: checking connection: LAN-TAP V9
>conet-bridged-daemon: checking connection: LAN-TAP V8(colinux)
>conet-bridged-daemon: checking connection: LAN-RTL8139
>conet-bridged-daemon: no matching adapter
>conet-bridged-daemon: error initializing winPCap


>eth0=pcap-bridge,"LAN-ADM851X",0A:C0:71:65:08:00,

Rainer---<=====>                         Vertraulich
             //  
           //                              
         <=====>--------------ocholl, Kiel, Germany ------------

Re: [coLinux-users] interface renumbering?

[<Use...@zo...>]

Use...@zo...(Rainer Zocholl)  07.01.08 01:37

>Hello




>Compiled on Dec 22 2007 05:10:09

>conet-bridged-daemon: Looking for interface "LAN-ADM851X"
>conet-bridged-daemon: checking connection: LAN-TAP V9
>conet-bridged-daemon: checking connection: LAN-TAP V8(colinux)
>conet-bridged-daemon: checking connection: LAN-RTL8139
>conet-bridged-daemon: no matching adapter
>conet-bridged-daemon: error initializing winPCap

>eth0=pcap-bridge,"LAN-ADM851X",0A:C0:71:65:08:00,


Using "LAN-RTL8139" it works:
conet-bridged-daemon: Looking for interface "LAN-RTL8139"
conet-bridged-daemon: checking connection: LAN-TAP V9
conet-bridged-daemon: checking connection: LAN-TAP V8(colinux)
conet-bridged-daemon: checking connection: LAN-RTL8139
conet-bridged-daemon: listening on: Realtek RTL8139/810x Family Fast Ethernet NIC                                   ...
conet-bridged-daemon: listening for: (ether dst 0a:c0:71:65:08:00) or (ether
bro
adcast or multicast) or (ip broadcast or multicast)


Using wireshark i found:
Wireshark does not see "LAN-ADM851X" anymore too!
So it's problem of libpcap and not Colinux!

Where to complain what?

Is there a workarround?

Rainer---<=====>                         Vertraulich
             //  
           //                              
         <=====>--------------ocholl, Kiel, Germany ------------

[coLinux-users] interface renumbering?

[<Use...@zo...>]

Hello

from time to time my USB Ethernet interface "disappears"
I have to reinstall its driver.
After reinstall it got this time the name
"LAN-Verbindung 10"
which i, as usual, rename to the device name "LAN-ADM851X".
But now colinux colinux can't find it any more tho
it's there and working!

At the same time tried to run colinux as a service.
The Ethernet "disapeared" during that phase.
But i deinstalled the service mean while to get more usefull
messages.

May i assume the colinux is not expecting so many devices
or what's wrong or can be done to make it work again?

I rebooted etc.




Compiled on Dec 22 2007 05:10:09

conet-bridged-daemon: Looking for interface "LAN-ADM851X"
conet-bridged-daemon: checking connection: LAN-TAP V9
conet-bridged-daemon: checking connection: LAN-TAP V8(colinux)
conet-bridged-daemon: checking connection: LAN-RTL8139
conet-bridged-daemon: no matching adapter
conet-bridged-daemon: error initializing winPCap


kernel=vmlinux
initrd=initrd.gz
mem=128
cobd0=\colinux.disks\Debian-3.0r2.ext3-mit-backports.imi
root=/dev/cobd0 hda=:cobd0
cobd1=\colinux.disks\swap_file hdb=:cobd1
cofs0=\colinux.disks\exchange
cobd4=\Device\Cdrom0 hdc=:cobd4
ttys0=COM2,"BAUD=115200 PARITY=n DATA=8 STOP=1 dtr=on rts=on"
exec0=C:\Programme\Xming\Xming.exe,":0 -clipboard -multiwindow -ac"
eth0=pcap-bridge,"LAN-ADM851X",0A:C0:71:65:08:00,


C:\coLinux>netsh interface ip dump


# ----------------------------------
# Schnittstellen-IP-Konfiguration
# ----------------------------------
pushd interface ip


# Schnittstellen-IP-Konfiguration für ""LAN-TAP V9""

set address name="LAN-TAP V9" source=dhcp
set dns name="LAN-TAP V9" source=dhcp register=PRIMARY
set wins name="LAN-TAP V9" source=dhcp

# Schnittstellen-IP-Konfiguration für ""LAN-TAP V8(colinux)""

set address name="LAN-TAP V8(colinux)" source=dhcp
set dns name="LAN-TAP V8(colinux)" source=dhcp register=PRIMARY
set wins name="LAN-TAP V8(colinux)" source=dhcp

# Schnittstellen-IP-Konfiguration für ""LAN-RTL8139""

set address name="LAN-RTL8139" source=static addr=192.168.1.2
mask=255.255.255.0

add address name="LAN-RTL8139" addr=172.16.0.2 mask=255.255.255.0
set dns name="LAN-RTL8139" source=static addr=none register=PRIMARY
set wins name="LAN-RTL8139" source=static addr=none

# Schnittstellen-IP-Konfiguration für ""LAN-ADM851X""

set address name="LAN-ADM851X" source=dhcp
set dns name="LAN-ADM851X" source=dhcp register=PRIMARY
set wins name="LAN-ADM851X" source=dhcp


popd
# Ende der Schnittstellen-IP-Konfiguration


ipconfig:

Ethernetadapter LAN-ADM851X:
        Beschreibung. . . . . . . . . . . : ADM851X USB To Fast Ethernet
Adapter
...

Rainer---<=====>                         Vertraulich
             //  
           //                              
         <=====>--------------ocholl, Kiel, Germany ------------

Re: [coLinux-users] CD Boot failure due to missing squashfs

[H. N. <hen...@ar...>]

Hello Markus,

Markus Stockhausen wrote:
> I just tried to boot a linux distribution (CRUX) from an
> ISO file mounted on /dev/cobd0. Everything worked
> fine until the init script tried to mount an squashfs
> file(system) from the CD. This failed as squashfs
> support ist not included into the default coLinux kernel
> (as of version 0.7.1) - it is available as a module but
> this does not help.
>=20
> After an hour of compiling a new colinux kernel I finally
> was able to boot the CD. I think it would be helpful
> for other users and distros to compile squashfs
> support directly into the default kernel. So recompiling
> a user defined kernel would be unneccessary.

Please send me a diff between default coLinux kernel config and yours. Than=
 we can add it.

Second: Have you a short guide for booting this. Or was it very easy after =
your kernel build?

--=20
Henry

Viel oder wenig? Schnell oder langsam? Unbegrenzt surfen + telefonieren
ohne Zeit- und Volumenbegrenzung? DAS TOP ANGEBOT F=DCR ALLE NEUEINSTEIGER
Jetzt bei Arcor: g=FCnstig und schnell mit DSL - das All-Inclusive-Paket
f=FCr clevere Doppel-Sparer, nur  29,95 Euro  inkl. DSL- und ISDN-Grundgeb=
=FChr!
http://www.arcor.de/rd/emf-dsl-2

[coLinux-users] CD Boot failure due to missing squashfs

[Markus S. <mar...@co...>]

Hi,

I just tried to boot a linux distribution (CRUX) from an
ISO file mounted on /dev/cobd0. Everything worked
fine until the init script tried to mount an squashfs
file(system) from the CD. This failed as squashfs
support ist not included into the default coLinux kernel
(as of version 0.7.1) - it is available as a module but
this does not help.

After an hour of compiling a new colinux kernel I finally
was able to boot the CD. I think it would be helpful
for other users and distros to compile squashfs
support directly into the default kernel. So recompiling
a user defined kernel would be unneccessary.

Best regards.

M. Stockhausen
  

****************************************************************************
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail
irrtümlich erhalten haben, informieren Sie bitte sofort den Absender und
vernichten Sie diese Mail. Das unerlaubte Kopieren sowie die unbefugte
Weitergabe dieser Mail ist nicht gestattet.

Über das Internet versandte E-Mails können unter fremden Namen erstellt oder
manipuliert werden. Deshalb ist diese als E-Mail verschickte Nachricht keine
rechtsverbindliche Willenserklärung.

Collogia
Unternehmensberatung AG
Ubierring 11
D-50678 Köln

Vorstand:
Kadir Akin
Prof. Gerd M. Henselmann
Dr. Michael Höhnerbach

Vorsitzender des Aufsichtsrates:
Hans Kristian Langva

Registergericht: Amtsgericht Köln
Registernummer: HRB 52 497

This e-mail may contain confidential and/or privileged information. If you
are not the intended recipient (or have received this e-mail in error)
please notify the sender immediately and destroy this e-mail. Any
unauthorized copying, disclosure or distribution of the material in this
e-mail is strictly forbidden.

e-mails sent over the internet may have been written under a wrong name or
been manipulated. That is why this message sent as an e-mail is not a
legally binding declaration of intention.

Collogia
Unternehmensberatung AG
Ubierring 11
D-50678 Köln

executive board:
Kadir Akin
Prof. Gerd M. Henselmann
Dr. Michael Höhnerbach

President of the supervisory board:
Hans Kristian Langva

Registry office: district court Cologne
Register number: HRB 52 497

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

Re: [coLinux-users] Colinux Ubuntu Guest Extremely slow

[<Use...@zo...>]

Hen...@Ar...(Henry Nestler)  21.12.07 20:47


>> Maybe fiddling with "boot.ini" might help to find the core problem
>>
>> Some ideas:
>>
>> /burnmemory=1024
>> /maxmem=1024 (better use burnmemory)
>> /3GB (or not?)
>> /userva=2900 (or less?)
>> /PAE (?)
>> /pcilock  (?)

>Hm. An ugly idea. - Of curse, if you use /NOPAE and not all memory,
>then the coLinux driver would be faster. PAE handling is liddle more
>overhead. I think, it's not a good idea to downgrade the system memory
>to running faster :-(

it's just a test to find out if that would be a cause of the problem.
It's of course not a solution, but -maybe- the OP ist happier with 
"only" 1,9GB RAM but a good performing Guest. RAM is cheap, currently,
man power not.
 
Rainer

Re: [coLinux-users] Colinux Ubuntu Guest Extremely slow

[Henry N. <Hen...@Ar...>]

Rainer Zocholl wrote:
> pa...@wa...(Paul Waldo)  21.12.07 08:52
>> I am running the latest colinux (0.7.1), 
> 
> Please try the devlopement release 0.8.
> It's stable enough for a developer ,)

One interesting change exist for machines with DEP/noexecute and PAE
(Physical Adress Extension enabled): In 0.7.1 colinux does "flush tlb"
on every OS switch as workarrount to fix crashing Intel CPUs.
coLinux version 0.7.2 and 0.8.0 runs the "flush tlb" only ones on
startup. This can make CPU faster. The CPU cache is not cleared on every
OS switch.

Give the 0.7.1 from http://www.colinux.org/snapshots/ a try.
It's a release candidate. ;-)

>> and it seems to be running quite slow.  
>> The KDE UI is somewhat jerky because of the slowness, but
>> my main concern is raw processing power.  I am hoping to be able to do
>> Java software development at work without having to interact with
>> Windows.  I'm willing to deal with the fact that coLinux is only using
>> one of the two CPUs, but I can't justify using it on a daily basis if
>> it takes significantly longer to do my job.

CoLinux kernel runs only on the first CPU. That you can not change. The
daemons and your Xserver should (can) use the other CPU.

>> Here are the platform details:
> 
>> Host
>> *******
>> Windows XP
>> coLinux 0.7.1
>> Pentuim D, ~3 GHz
>> 3GB RAM
> 
>> Guest
>> *********
>> Kubuntu Gutsy i386
> 
>> I booted my Kubuntu Live CD and found that it reported the bogoMIPS at
>> 5990.33.  coLinux reports the bogoMIPS at 1520.43.  That means the
>> coLinux guest is running at ~25% of native speed.  I know that
>> bogoMIPS are not an true speed representation, but I think it is close
>> enough to give a general idea of what is going on.

bogoMIPS have no real timebase under coLinux. That means, you can not
compair it with native boot. The function inside bogoMIPS calcuation is
an OS switch under coLinux and a loop in native linux.

If you feel, coLinux is slow, check the network you used. Hope, you use
tap for the network and is right configured (no bad routes and dropping
packets).

After colinux was started, incrase the process priority for
colinux-daemon.exe and for the network daemons. Some users sayed, it
makes coLinux faster.

Incrase the memory for coLinux. "mem=1000" is the highest you can use.

> Maybe fiddling with "boot.ini" might help to find the core problem
> 
> Some ideas:
> 
> /burnmemory=1024
> /maxmem=1024 (better use burnmemory)
> /3GB (or not?)
> /userva=2900 (or less?)
> /PAE (?)
> /pcilock  (?)

Hm. An ugly idea. - Of curse, if you use /NOPAE and not all memory, then
the coLinux driver would be faster. PAE handling is liddle more
overhead. I think, it's not a good idea to downgrade the system memory
to running faster :-(

-- 
Henry N.