Hi Marcel.
 
> Can you please not try to squeeze such change in the same patch. Just
> have a separate one that does this change.
...
> The more I looked through the rest of this patch, the more I thought
> you need to split this patch into a cleanup and caif_rtnl usage patch.

Yes, I have broken this into three patches.
It's a bit hard to find the exactly what to put where, and
tear apart what is already squashed. But have a look and
see what you think.

> 
> So I realized now that you are using ifindex as input for
> caif_rtnl_delete_interface, now I am questioning why there is a result
> parameter in the callback for caif_rtnl_create_interface.

OK, agree - done.

> Why are we having this in the first place. CAIF network interfaces are
> point-to-point, right? In that case you should just leave this NULL.

Yes - agree. PtP link should not report GW.

> Please use '\0' instead of just 0. You might need to fix this at other
> places as well.

OK, done at least here..

> > +static void rtnl_callback(int result, gpointer user_data,
> > +           char *ifname, int ifindex)
> > +{
> > +   struct conn_info *conn = user_data;
> > +
> > +   strncpy(conn->interface, ifname, sizeof(conn->interface));
> > +   conn->ifindex = ifindex;
> > +
> > +   if (result == 0)
> > +           conn->created = TRUE;
> > +   else {
> > +           conn->created = FALSE;
> > +           DBG("Could not create CAIF Interface");
> > +   }
> > +}
> 
> So here is the think. This makes no sense. Either you get an error and
> ifname and index are not valid or you succeed.
> 
> So this should be something like
> 
>       if (ifname < 0) {
>               connman_error(...)
>               return;
>       }
OK, I've done something similar.


Regards,
Sjur




_______________________________________________
ofono mailing list
[email protected]
http://lists.ofono.org/listinfo/ofono

Reply via email to