On Wed, 2023-02-15 at 18:43 +0100, Johannes Berg wrote:
> On Thu, 2023-02-09 at 06:09 +0800, Ryder Lee wrote:
> > This is utilized to pass LDPC configurations from user space (i.e.
> > hostap) to driver.
>
> again, hostapd
>
> why do we need this?
>
User can configure hostapd.conf "# LDPC coding capability: [LDPC] =
supported" to decide if an AP mode has LDPC caps. They are actaully
long-standing configurations of hostapd, but have never been propegated
to kernel. One of our use cases is for certification setup.
> > + * @ht_ldpc: in AP mode, indicates interface owns HT LDPC
> > capability.
> > + * @vht_ldpc: in AP mode, indicates interface owns VHT LDPC
> > capability.
> > + * @he_ldpc: in AP mode, indicates interface owns HE LDPC
> > capability.
>
> what do you mean by "owns" the capability? Has? Uses it?
Copy-and-pasted from @mu_mimo_owner, but indeed, "has" is better. Will
fix.
> > diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c
> > index 6bf1cdf254f6..a080010da8fb 100644
> > --- a/net/mac80211/cfg.c
> > +++ b/net/mac80211/cfg.c
> > @@ -1252,7 +1252,16 @@ static int ieee80211_start_ap(struct wiphy
> > *wiphy, struct net_device *dev,
> > prev_beacon_int = link_conf->beacon_int;
> > link_conf->beacon_int = params->beacon_interval;
> >
> > + if (params->ht_cap) {
> > + link_conf->ht_ldpc =
> > + params->ht_cap->cap_info &
> > + cpu_to_le16(IEEE80211_HT_CAP_LDPC_CODIN
> > G);
> > + }
>
> No need for braces.
>
OK.
Ryder
