On Thu, Nov 10, 2011 at 2:47 PM, Markus Armbruster <armbru@xxxxxxxxxx> wrote: > Pekka Enberg <penberg@xxxxxxxxxx> writes: > >> Hi Anthony, >> >> On Thu, Nov 10, 2011 at 3:43 PM, Anthony Liguori <anthony@xxxxxxxxxxxxx> wrote: >>> It's not just the qcow2 implementation or even the block layer. This pull >>> requests adds a userspace TCP/IP stack to the kernel and yet netdev isn't on >>> the CC and there are no Ack's from anyone from the networking stack. I'm >>> fairly sure if they knew what was happening here they would object. >> >> It's something we consider extremely important because it allows easy >> non-root networking. But you're right, we definitely ought to ping the >> networking folks before the next merge window. > > The problem is real. The solution "duplicate in user space" sucks. If > you engaging with the kernel networking folks leads to one that doesn't > suck, we should bathe you in free beer. Look at disks, the problem is addressed by the udisks daemon on dbus. Anything can try talking to it. If you have permissions or can get the user to authenticate then you can manipulate LVM volumes, mount file systems, etc. We could do something similar for tap networking. The Ubuntu folks seem to want VDE instead to solve the same problem. Create a VDE switch with a single tap device on startup. Then let all VMs talk to the VDE without privileges. I don't think going through VDE is nice or performant, would be better to have add virtual network functionality over dbus just like udisks did for disks. Stefan -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
