> -----Original Message----- > From: Rameshkumar Sundaram (QUIC) <quic_ramess@xxxxxxxxxxx> > Sent: Monday, December 27, 2021 9:50 AM > To: Johannes Berg <johannes@xxxxxxxxxxxxxxxx> > Cc: linux-wireless@xxxxxxxxxxxxxxx; Rameshkumar Sundaram (QUIC) > <quic_ramess@xxxxxxxxxxx>; Lavanya Suresh <lavaks@xxxxxxxxxxxxxx> > Subject: RE: [PATCH v2] mac80211: disable BSS color collision detection in > case of no free colors > > > > -----Original Message----- > > From: Johannes Berg <johannes@xxxxxxxxxxxxxxxx> > > Sent: Monday, December 20, 2021 3:32 PM > > To: Rameshkumar Sundaram (QUIC) <quic_ramess@xxxxxxxxxxx> > > Cc: linux-wireless@xxxxxxxxxxxxxxx; Lavanya Suresh > > <lavaks@xxxxxxxxxxxxxx> > > Subject: Re: [PATCH v2] mac80211: disable BSS color collision > > detection in case of no free colors > > > > Hi, > > > > > > > include/net/cfg80211.h | 28 +++++++------ > > > net/mac80211/cfg.c | 16 ++++++-- > > > net/wireless/nl80211.c | 109 > > > ++++++++++++++++++++++++------------------------- > > > 3 files changed, 82 insertions(+), 71 deletions(-) > > > > > > > This is now a fairly big cfg80211 change, and not much mac80211. Can > > you please split it? > > > > But you didn't really address why we need to do this via element > > change detection, rather than letting hostapd do this via the > > NL80211_ATTR_HE_BSS_COLOR attribute even in change_beacon? > > > > Ah! I totally missed this attribute parsing option and went with ap_params > from last discussion. > > Yes, hostapd would send NL80211_ATTR_HE_BSS_COLOR for change_beacon > too, maybe we should relocate cfg80211_he_bss_color to beacon data and > do nl80211_parse_he_bss_color() in nl80211_parse_beacon() in that case to > have this data for both commands. > But channel_switch and color_change commands won't have this attribute set and We may disable color there. We will have to have a flag and set if NL80211_ATTR_HE_BSS_COLOR is present in the NL command and check that flag in assign_beacon() before using params->he_bss_color struct data to prevent that then. So, can we go with this IE checking itself or we can do above change, please suggest. > > johannes