On 5/26/20 6:01 AM, Johannes Berg wrote:
> Hi Thomas,
>
> This looks mostly good (and I've applied the other 4 patches with some
> fixups), but ...
>
>> @@ -911,6 +931,15 @@ bool cfg80211_chandef_usable(struct wiphy *wiphy,
>> control_freq = chandef->chan->center_freq;
>>
>> switch (chandef->width) {
>> + case NL80211_CHAN_WIDTH_1:
>> + width = 1;
>> + break;
>> + case NL80211_CHAN_WIDTH_2:
>> + width = 2;
>> + break;
>> + case NL80211_CHAN_WIDTH_4:
>> + width = 4;
>> + break;
>> case NL80211_CHAN_WIDTH_5:
>
> aren't you missing some bandwidths here? You'd fall through to a
> WARN_ON(), which doesn't seem good?
Yep. Looks like hwsim tests needs a test for larger bandwidths.
--
thomas
