Hi Florian,
Florian Fainelli <f.faine...@gmail.com> writes:
> On 10/12/2017 03:51 PM, Vivien Didelot wrote:
>> The dsa_port structure has a "netdev" member, which can be used for
>> either the master device, or the slave device, depending on its type.
>>
>> It is true that today, CPU port are not exposed to userspace, thus the
>> port's netdev member can be used to point to its master interface.
>>
>> But it is still slightly confusing, so split it into more explicit
>> "master" and "slave" members.
>
> I do see some value in doing that, although I also see value in having
> structure members be named after what they are, rather than their use
> (oh well, it's all debatable anyway), see below for a suggestion on how
> to reconcile the two:
>
>> struct dsa_port {
>> + /* Master device, physically connected if this is a CPU port */
>> + struct net_device *master;
>> +
>> + /* Slave device, if this port is exposed to userspace */
>> + struct net_device *slave;
>> +
>
> How about using:
>
> union {
> struct net_device *master;
> struct net_device *slave;
> } netdev;
>
> Such that this serves both purposes of clearly communicating what the
> structure member is, and it can be either one of the two, but not both
> at the same time?
I love that! It makes clear that master is not available for a non-CPU
port. Using this union is correct for the moment because DSA and CPU
ports don't have a slave device attached to them. If this becomes true
one day (unlikely), we'll remove the union.
Thanks,
Vivien
