Open silverjam opened 10 years ago
Per comment from Thomas when we were reviewing the wpa_supplicant mesh support patches to send upstream:
> /* TODO: verify this peer matches MBSS before inserting! */ > /* TODO: process in SAE, which will allocate station if authenticated. */ > /* just immediately allocate peer for now, and insert into driver */ > diff --git a/wpa_supplicant/mesh_mpm.c b/wpa_supplicant/mesh_mpm.c > index 160d05a..c3febcb 100644 > --- a/wpa_supplicant/mesh_mpm.c > +++ b/wpa_supplicant/mesh_mpm.c > @@ -306,6 +306,7 @@ wpa_mesh_new_mesh_peer(struct wpa_supplicant *wpa_s, const u8 *addr, > struct mesh_conf *conf = wpa_s->ifmsh->mconf; > struct hostapd_data *data = wpa_s->ifmsh->bss[0]; > struct sta_info *sta; > + struct wpa_ssid *ssid = wpa_s->current_ssid; > int ret = 0; > > sta = ap_get_sta(data, addr); > @@ -348,6 +349,12 @@ wpa_mesh_new_mesh_peer(struct wpa_supplicant *wpa_s, const u8 *addr, > return; > } > > + if (ssid && ssid->no_auto_peer) { > + wpa_msg(wpa_s, MSG_INFO, "will not initiate new peer link with " > + MACSTR " because of no_auto_peer", MAC2STR(addr)); > + return; > + } This needs to be ratelimited, unless there is a way to supress NEW_PEER_CANDIDATE events?
Per comment from Thomas when we were reviewing the wpa_supplicant mesh support patches to send upstream: