Re: [RFC] IWD as wifi backend

[Andrew Zaborowski <andrew zaborowski intel com>]

Hi Thomas,

On 30 November 2017 at 17:20, Thomas Haller <thaller redhat com> wrote:
> Maybe it's simpler to have just two indpendent types NMDeviceWifi (for
> supplicant) and NMDeviceIwd.
>
> They both can implement the D-Bus interface
> org.freedesktop.NetworkManager.Device.Wireless, there is no strong
> requirement that they share a common parent class.
>
> I was thinking that "nm-wifi-factory.c"'s create_device() would either
> call nm_device_wifi_new() or nm_device_iwd_new(), depending on some
>     nm_config_data_get_value (nm_config_get_data_orig (NM_CONFIG_GET),
>                               "main",
>                               "wifi-backend")
>
> Common code should be shared, but it's not clear that you need a common
> parent GObject type for that.

Sounds good, but eventually I think a common parent class would work
well because the functions that manage the AP list can all be common,
also most of the activation stages (act_stage1_prepare etc.) as the
wpa_supplicant-specific part is small, most of it is querying the
secrets and validation.

> > > Say " (experimental)"?
> > > Since the plugin doesn't include any new public API, every
> > > misbehavior
> > > is just a bug that we can fix later. We don't commit to new API
> > > here.
> > > So, "experimental" is just cosmetic to set user expectations
> > > straight.
> >
> > Ok, do you prefer that this be enabled by default so that it gets
> > build-tested and the iwd backend be guarded by the
> > NetworkManager.conf
> > setting (also with "experimental" comment)?
>
> I would disable it by default (at least initially).

Ok.

Best regards