[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] Fwd: netback (Network backend) in domU



Please don't top post, makes it harder to follow the conversation.

On Thu, 2012-06-21 at 19:17 +0100, Shakeel Butt wrote:
> I have messed up my copy/paste on details of netback-domU and
> other-domU. So, for these two the clear details are below:
> 
> Detail of netback-domU:
> -----------------------
[...]
> Detail of other-domU:
> -----------------------
[...]

So if I understand correctly netback-domU's route to the outside world
if via a virtual interface connected to dom0 -- is that right?

If so then I'm afraid I think this doesn't work -- you can bridge
several netback devices together or you can bridge netfront devices
together but in the current implementation of netfront/netback I believe
it is not possible to bridge a mixture of the two.

The reason is that the netfront/back protocol currently does not make
use of "transitive grants" which are necessary for the netback-domU
domain to forward pages from a netfront to the netback running in dom0.
Trying to forward in this way without using a "transitive grant" is what
results in your original "grant_table.c:387:d0 Could not pin grant frame
9d44e" message. I'm afraid I don't know of anyone who is currently
working on implementing this feature.

The more normal configuration (which is what I thought you were using at
first) would be to pass a physical NIC to netback-domU using PCI
passthrough and then optionally create a VIF on dom0 in order to give it
access. Perhaps that would be suitable for your usecase?

Ian.




_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.