[PATCH v2] driver_nl80211: report invalid signal and noise when info is unavailable

[Brian Norris]

On Tue, Oct 27, 2020 at 4:00 PM Matthew Wang <matthewmwang at chromium.org> wrote:
>
> Sorry for the late reply, this got lost in my inbox. In e.g.
> wpa_supplicant_need_to_roam_within_ess, we consider RSSI >= 0 to be
> "unspecified units (not in dBm)", and it's best to keep these all

To add to this: "unspecified" units come from something like
NL80211_BSS_SIGNAL_MBM vs. NL80211_BSS_SIGNAL_UNSPEC, which apply to
scan results. I don't think there's an equivalent UNSPEC possibility
for the CQM / EVENT_SIGNAL_CHANGE events where wpa_bss_update_level()
is used -- at least in this case, the signal values are coming from
NL80211_STA_INFO_SIGNAL, which has documented units (dBm).

But the below seems more compelling:

> consistent IMO. If we should actually be treating >= 0 RSSI values as
> valid dBm values, then there should be a separate patch making all the
> relevant changes at once. For the purpose of this patch, that seems
> like an unnecessary (and potentially non-trivial) change in
> functionality that isn't mentioned in the commit message anyway.

+1, it doesn't seem like a necessary change to accomplish the goal of
this patch.

Brian