On Thu, Jan 03, 2013 at 04:45:14PM -0800, Ethan Jackson wrote:
> > This will cause the Ethernet address of vports to change randomly
> > whenever ovs-vswitchd is restarted. I don't know whether that is
> > important.
>
> I don't think this is a problem. The ethernet addresses are really only
> u
> This will cause the Ethernet address of vports to change randomly
> whenever ovs-vswitchd is restarted. I don't know whether that is
> important.
>
I don't think this is a problem. The ethernet addresses are really only
used in protocols (e.g. CFM) which will be pretty disrupted on restart
any
> I didn't notice this before but using sizeof mac here (and in the get
> function) is actually going to result in copying a pointer's worth of
> data so I think you want to explicitly use ETH_ADDR_LEN.
>
Good catch, changed.
Ethan
___
dev mailing list
On Wed, Dec 26, 2012 at 05:16:46PM -0800, Ethan Jackson wrote:
> Letting netdev-vport manage ethernet addresses itself instead of
> relying on the datapath has several advantages. It simplifies the
> code, is significantly more efficient, and will work when there is
> no longer a one to one mappin
On Wed, Dec 26, 2012 at 8:16 PM, Ethan Jackson wrote:
> diff --git a/lib/netdev-vport.c b/lib/netdev-vport.c
> index acf7d18..a95e09e 100644
> --- a/lib/netdev-vport.c
> +++ b/lib/netdev-vport.c
> netdev_vport_set_etheraddr(struct netdev *netdev,
> const uint8_t mac[ET
Letting netdev-vport manage ethernet addresses itself instead of
relying on the datapath has several advantages. It simplifies the
code, is significantly more efficient, and will work when there is
no longer a one to one mapping from netdev-vports to datapath
vports.
Signed-off-by: Ethan Jackson
