[PATCH 1/2] De-couple MBO and WNM

[Krishna Chaitanya]

On Sun, Dec 10, 2023 at 12:21 AM Krishna Chaitanya
<chaitanya.mgit at gmail.com> wrote:
>
> On Wed, Dec 6, 2023 at 4:16 PM Jouni Malinen <j at w1.fi> wrote:
> >
> > On Tue, Nov 28, 2023 at 12:49:52AM +0530, Chaitanya Tata wrote:
> > > MBO is mandatory for WFA certification but WNM is not, so, in order to
> > > have a lean build for WFA certification only, de-couple WNM from MBO so
> > > that it can be compiled out.
> > >
> > > WNM is still auto-selected automatically when MBO is enabled, this
> > > behaviour is unchanged.
> >
> > I'm not sure what this is trying to achieve.. As far as the claim about
> > WNM not being needed for MBO is concerned, the Wi-Fi Agile Multiband
> > specification mandate STA to support WNM-Notification Request frame.
> The idea is to only include the support for that Notification request frame
> and disable the rest of WNM to save memory (wnm_sta.o is ~24K text).
> >
> > > diff --git a/wpa_supplicant/mbo.c b/wpa_supplicant/mbo.c
> > > +#ifdef CONFIG_WNM
> > >  static void wpas_mbo_send_wnm_notification(struct wpa_supplicant *wpa_s,
> > >                                          const u8 *data, size_t len)
> >
> > And as far as adding CONFIG_WNM in mbo.c is concerned, this would have
> > no impact since this file is included only if CONFIG_MBO=y is defined
> > and if CONFIG_MBO is defined, then so will CONFIG_WNM.
> Yes, I don't want to change this behaviour, but if someone wants to save
> memory then WNM can be disabled and MBO would still work.
>
> Of course, this patch is wrong, I will push a v2 that properly fixes the issue.
Looks like, we need to separate out BTM code to a separate file btm.c
and include
it for both MBO and WNM (NEED_BTM=y). This way we can get rid of rest of the
WNM. WDYT?