Re: vkernel(7) usage and granularity of privileges

[Andreas Bartelt <dragonflybsd@xxxxxxxxxx>]

Hello Matt,

Matthew Dillon wrote:
. ..
>     You can then treat the TAP interface as a local IP space (or even bridge
>     it if you want).  If you treat it as a local IP space you can then use
>     something like PF to NAT it to the outside world and control the
>     bandwidth usage.

thanks a lot for the hint. After playing with both variants, I think 
I'll stick with the 'local IP space' setup which is connected via NAT to 
the outside world.

However, I've noticed a minor problem in combination with PF: since the 
tap interface gets created AFTER vknetd is run, enabling PF in 
/etc/rc.conf doesn't work in case filtering is also done on the tap 
interface (unknown interfaces give a parsing errror...). I suppose think 
it would be a good idea to add an option for vknetd to rc/rc.conf, in 
order to ensure that the tap interface is already created when PF starts 
(this further requires the kernel module for the tap interface to be 
enabled in /boot/loader.conf -- perhaps a comment in the rc.conf man 
page would help...). Basically the same problem applies to the bridging 
setup. What do you think?

regards,
Andreas