VIF setup

VIF setup

[Kero-Chan]

Please look at this does this look OK?

guest# ifconfig -a
eth0      Link encap:Ethernet  HWaddr AA:00:00:0C:0E:D9  
          inet addr:172.16.100.1  Bcast:172.16.255.255  Mask:255.255.0.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)

domain0# ifconfig -a
eth0      Link encap:Ethernet  HWaddr 00:08:54:08:3E:C7  
          inet addr:w.x.y.z  Bcast:255.255.255.255  Mask:255.255.254.0
          UP BROADCAST NOTRAILERS RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:131747 errors:0 dropped:0 overruns:0 frame:0
          TX packets:12223 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:15796392 (15.0 Mb)  TX bytes:1114824 (1.0 Mb)
          Interrupt:9 Base address:0xd000 

vif10.0   Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF  
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:3345 errors:0 dropped:34 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:0 (0.0 b)  TX bytes:200784 (196.0 Kb)

xen-br0   Link encap:Ethernet  HWaddr 00:08:54:08:3E:C7  
          inet addr:w.x.y.z  Bcast:255.255.255.255  Mask:255.255.255.255
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:131569 errors:0 dropped:0 overruns:0 frame:0
          TX packets:12221 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:13934617 (13.2 Mb)  TX bytes:1100770 (1.0 Mb)

The vif interface in domain0 is not initialised correctly, right?


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Mark Williamson]

> Please look at this does this look OK?

It's OK.  Because you're using an L2 ethernet bridge, it's not necessary to 
set up any address details for the vif in domain0.

Does it work?

HTH,
Mark

> guest# ifconfig -a
> eth0      Link encap:Ethernet  HWaddr AA:00:00:0C:0E:D9
>           inet addr:172.16.100.1  Bcast:172.16.255.255  Mask:255.255.0.0
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:0 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:1000
>           RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)
>
> domain0# ifconfig -a
> eth0      Link encap:Ethernet  HWaddr 00:08:54:08:3E:C7
>           inet addr:w.x.y.z  Bcast:255.255.255.255  Mask:255.255.254.0
>           UP BROADCAST NOTRAILERS RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:131747 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:12223 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:1000
>           RX bytes:15796392 (15.0 Mb)  TX bytes:1114824 (1.0 Mb)
>           Interrupt:9 Base address:0xd000
>
> vif10.0   Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:0 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:3345 errors:0 dropped:34 overruns:0 carrier:0
>           collisions:0 txqueuelen:0
>           RX bytes:0 (0.0 b)  TX bytes:200784 (196.0 Kb)
>
> xen-br0   Link encap:Ethernet  HWaddr 00:08:54:08:3E:C7
>           inet addr:w.x.y.z  Bcast:255.255.255.255  Mask:255.255.255.255
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:131569 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:12221 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:0
>           RX bytes:13934617 (13.2 Mb)  TX bytes:1100770 (1.0 Mb)
>
> The vif interface in domain0 is not initialised correctly, right?
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by: Beat the post-holiday blues
> Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
> It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Derrik Pates]

Mark Williamson wrote:
> It's OK.  Because you're using an L2 ethernet bridge, it's not necessary to 
> set up any address details for the vif in domain0.

I doubt it does, since the IP address is still bound to both xen-br0 and 
eth0; this will break things. That's something that really needs to be 
fixed in the network config script, because networking won't work right 
until that's corrected.

-- 
Derrik Pates
demon@devrandom.net


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Adam Sulmicki]

> Mark Williamson wrote:
>> It's OK.  Because you're using an L2 ethernet bridge, it's not necessary to 
>> set up any address details for the vif in domain0.
>
> I doubt it does, since the IP address is still bound to both xen-br0 and 
> eth0; this will break things. That's something that really needs to be fixed 
> in the network config script, because networking won't work right until 
> that's corrected.

I would agree here. In my experience you must clear eth0 first. something 
like

 	/sbin/ifconfig eth0 0.0.0.0 up


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Ronald G. Minnich]

On Fri, 14 Jan 2005, Adam Sulmicki wrote:

>        /sbin/ifconfig eth0 0.0.0.0 up

Am I the only person who sees this as a tad odd? I can't escape thinking 
about a real bridge. Seems to me each eth or vif you add to a bridge 
should retain its IP and the bridge, basically being a device that does 
mac-layer routing, has no IP. 

I guess this is an artifact of how it is implemented?

ron


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Derrik Pates]

Ronald G. Minnich wrote:
> Am I the only person who sees this as a tad odd? I can't escape thinking 
> about a real bridge. Seems to me each eth or vif you add to a bridge 
> should retain its IP and the bridge, basically being a device that does 
> mac-layer routing, has no IP. 

No, because the IP(s) assigned to a device cause the device to only 
accept packets destinde for it; when the IP is 0.0.0.0, the interface 
goes into promiscuous mode, and so it accepts all packets, whether 
destined for it or not. Then, the bridge device enslaves all the 
interfaces, and transports the packets among the participant interfaces.

> I guess this is an artifact of how it is implemented?

You could say that.

-- 
Derrik Pates
demon@devrandom.net


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Jan Kundrát]

Derrik Pates wrote:
> Ronald G. Minnich wrote:
> 
>> Am I the only person who sees this as a tad odd? I can't escape 
>> thinking about a real bridge. Seems to me each eth or vif you add to a 
>> bridge should retain its IP and the bridge, basically being a device 
>> that does mac-layer routing, has no IP. 
> 
> 
> No, because the IP(s) assigned to a device cause the device to only 
> accept packets destinde for it; when the IP is 0.0.0.0, the interface 
> goes into promiscuous mode, and so it accepts all packets, whether 
> destined for it or not. Then, the bridge device enslaves all the 
> interfaces, and transports the packets among the participant interfaces.

Isn't promisciuous mode L2 based? If you do `ifconfig eth0 promisc`, 
it'll pass all packets, even those having foreign destination MAC 
address, to upper layers, AFAIK.

-jkt

-- 
cd /local/pub && more beer > /dev/mouth



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Felipe Alfaro Solana]

On 15 Jan 2005, at 04:40, Ronald G. Minnich wrote:

> On Fri, 14 Jan 2005, Adam Sulmicki wrote:
>
>>        /sbin/ifconfig eth0 0.0.0.0 up
>
> Am I the only person who sees this as a tad odd? I can't escape 
> thinking
> about a real bridge. Seems to me each eth or vif you add to a bridge
> should retain its IP and the bridge, basically being a device that does
> mac-layer routing, has no IP.
>
> I guess this is an artifact of how it is implemented?

In my xen-unstable configuration, neither xen-br0 nor vif have any IP 
addresses. BTW, xen-br0 is linked to eth1, which is also up but with no 
configured IP.



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Jody Belka]

On Fri, Jan 14, 2005 at 08:40:08PM -0700, Ronald G. Minnich wrote:
> On Fri, 14 Jan 2005, Adam Sulmicki wrote:
> >        /sbin/ifconfig eth0 0.0.0.0 up
> 
> Am I the only person who sees this as a tad odd? I can't escape thinking 
> about a real bridge. Seems to me each eth or vif you add to a bridge 
> should retain its IP and the bridge, basically being a device that does 
> mac-layer routing, has no IP. 

Actually, if you think about it, it makes complete sense. I really can't
imagine a switch where each port has it's own IP address. The switch itself
having an IP address on the other hand is perfectly reasonable, as in the
case of a managed switch.

> I guess this is an artifact of how it is implemented?

I think it'd be wierd if it worked any other way, taking the above into account.


J

-- 
Jody Belka
knew (at) pimb (dot) org


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Martin Maney]

On Sat, Jan 15, 2005 at 05:20:56PM +0100, Jody Belka wrote:
> Actually, if you think about it, it makes complete sense. I really can't
> imagine a switch where each port has it's own IP address. The switch itself

Yes, exactly.  Switching is at a level below IP - indeed, switches work
fine with non-IP packets.

> having an IP address on the other hand is perfectly reasonable, as in the
> case of a managed switch.

Which is, from a protocol point of view, not an address that's actually
"the switch" - it's the address of this entirely separate device that
happens to be in the same box of hardware with the switch and has some
backside connection to the switch that's entirely separate from the
network.  That the switch and the management unit may share some of the
same hardware, however intimately, is just an implementation detail. :-)

-- 
The reason [limited term of copyright is] important is this:
Publishers are in the business of expanding capital.
The writers who supply them are in the business of
expanding civilization itself.  -- John Bloom



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: VIF setup

[Mike Wray]

Jody Belka wrote:
> On Fri, Jan 14, 2005 at 08:40:08PM -0700, Ronald G. Minnich wrote:
> 
>>On Fri, 14 Jan 2005, Adam Sulmicki wrote:
>>
>>>       /sbin/ifconfig eth0 0.0.0.0 up
>>
>>Am I the only person who sees this as a tad odd? I can't escape thinking 
>>about a real bridge. Seems to me each eth or vif you add to a bridge 
>>should retain its IP and the bridge, basically being a device that does 
>>mac-layer routing, has no IP. 
> 
> 
> Actually, if you think about it, it makes complete sense. I really can't
> imagine a switch where each port has it's own IP address. The switch itself
> having an IP address on the other hand is perfectly reasonable, as in the
> case of a managed switch.
> 
> 
>>I guess this is an artifact of how it is implemented?
> 
> 
> I think it'd be wierd if it worked any other way, taking the above into account.
> 
> 
> J
> 

Check the linux bridge howto at

http://bridge.sourceforge.net/howto.html

This site also has pointers to other info.
There's really nothing special about the xen-br0 bridge, it's
just a normal linux bridge.

Mike


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

VIF setup

[Jacob Gorm Hansen]

hi,

I just noticed that my vif setup code from buscreate no longer works
after upgrading to 3.0.3. Is there anywhere I can find documentation on
the current state of the Xenstore protocol for VIF setup? My current
code is below:

    sprintf(s,"%s/handle",frontend);
    sprintf(s2,"%d",uuid);
    xs_w(s, s2);

    sprintf(s,"%s/mac",frontend);
    xs_w(s, "aa:00:00:11:a6:02");

    sprintf(s, "%s/backend-id", frontend);
    xs_w(s, "0");

    sprintf(s,"%s/backend",frontend);
    xs_w(s, backend );

    sprintf(s, "%s/state", frontend);
    xs_w(s, "1");

    xs_end();
    xs_begin();

    sprintf(s,"%s/handle",backend);
    sprintf(s2,"%d",uuid);
    xs_w(s, s2);

    sprintf(s,"%s/mac",backend);
    xs_w(s, "aa:00:00:11:a6:02");

    sprintf(s,"%s/script",backend);
    xs_w(s, "");

    sprintf(s,"%s/frontend",backend);
    xs_w(s, frontend );

    sprintf(s,"%s/frontend-id",backend);
    sprintf(s2,"%d",domid);
    xs_w(s, s2 );

    sprintf(s, "%s/state", backend);
    xs_w(s, "1");

    xs_end();

Thanks,
Jacob

Re: VIF setup

[Ewan Mellor]

On Mon, Nov 20, 2006 at 11:35:43PM +0100, Jacob Gorm Hansen wrote:

> hi,
> 
> I just noticed that my vif setup code from buscreate no longer works
> after upgrading to 3.0.3. Is there anywhere I can find documentation on
> the current state of the Xenstore protocol for VIF setup? My current
> code is below:

Nothing changed, AFAIR.  What's the nature of your failure?

Ewan.

Re: VIF setup

[Jacob Gorm Hansen]

On Mon, 2006-11-20 at 22:43 +0000, Ewan Mellor wrote:
> On Mon, Nov 20, 2006 at 11:35:43PM +0100, Jacob Gorm Hansen wrote:
> 
> > hi,
> > 
> > I just noticed that my vif setup code from buscreate no longer works
> > after upgrading to 3.0.3. Is there anywhere I can find documentation on
> > the current state of the Xenstore protocol for VIF setup? My current
> > code is below:
> 
> Nothing changed, AFAIR.  What's the nature of your failure?
> 

The domU logs the following in dmesg:
XENBUS: Timeout connecting to device: device/vif/11600 (state 0)

Thanks,
Jacob

Re: VIF setup

[Keir Fraser]

On 20/11/06 11:19 pm, "Jacob Gorm Hansen" <jacobg@diku.dk> wrote:

>> Nothing changed, AFAIR.  What's the nature of your failure?
>> 
> 
> The domU logs the following in dmesg:
> XENBUS: Timeout connecting to device: device/vif/11600 (state 0)

I'd stick xenstored into logging mode and create a guest using the python
tools. Then see what nodes are now written by xend (might need to extend the
logging to make it clear who writes what into the store). I can't think of
anything essential that's been added though.

 -- Keir

Re: VIF setup

[Jacob Gorm Hansen]

On Tue, 2006-11-21 at 00:19 +0100, Jacob Gorm Hansen wrote:

> The domU logs the following in dmesg:
> XENBUS: Timeout connecting to device: device/vif/11600 (state 0)

Turns out Xenstored was hosed, a reboot seems to have cured it.

Thanks,
Jacob