> How would the P2P Client know which BSSID the GO is using in this
> instance? The BSSID, i.e., the P2P Interface Address, may change between
> each invocation of a persistent group.
The use case in ChromeOS and Android is out-of-band discovery, where
the BSSID is communicated via e.g. Bluetooth.
> The params->bssid here is the bssid value in the special network profile
> that is used to store the persistent group information. That special
> case uses the bssid value to store the P2P Device Address of the GO; not
> the P2P Interface Address (which would be the BSSID). How is this
> supposed to work if the GO uses P2P Interface Addresses that differ from
> the P2P Device Address?
Ah I see. In that case, how about directly passing in a BSSID via
D-Bus? I've sent a new patch, PTAL!
On Tue, Feb 21, 2023 at 3:47 AM Jouni Malinen <j@xxxxx> wrote:
>
> On Tue, Dec 20, 2022 at 07:15:46PM -0800, Matthew Wang wrote:
> > Allow clients to force the BSSID of an auto GO. If the auto GO has been
> > discovered on another interface, optimize scan frequency by performing
> > a single channel scan first. Android and ChromeOS use this to streamline
> > auto GO discovery.
>
> How would the P2P Client know which BSSID the GO is using in this
> instance? The BSSID, i.e., the P2P Interface Address, may change between
> each invocation of a persistent group.
>
> > diff --git a/wpa_supplicant/bss.c b/wpa_supplicant/bss.c
> > @@ -240,7 +240,7 @@ void wpa_bss_remove(struct wpa_supplicant *wpa_s, struct wpa_bss *bss,
> > - * @bssid: BSSID
> > + * @bssid: BSSID, or %NULL to match any BSSID
>
> > @@ -252,7 +252,8 @@ struct wpa_bss * wpa_bss_get(struct wpa_supplicant *wpa_s, const u8 *bssid,
> > if (!wpa_supplicant_filter_bssid_match(wpa_s, bssid))
> > return NULL;
>
> That could result in NULL pointer dereferencing the bssid parameter,
> i.e., this call needs to be made conditional on bssid being not-NULL.
>
> > diff --git a/wpa_supplicant/dbus/dbus_new_handlers_p2p.c b/wpa_supplicant/dbus/dbus_new_handlers_p2p.c
> > +++ b/wpa_supplicant/dbus/dbus_new_handlers_p2p.c
> > @@ -356,6 +356,7 @@ DBusMessage * wpas_dbus_handler_p2p_group_add(DBusMessage *message,
> > + int force_go_bssid = 0;
>
> That should really be bool/false here.
>
> > @@ -382,6 +383,9 @@ DBusMessage * wpas_dbus_handler_p2p_group_add(DBusMessage *message,
> > + } else if (os_strcmp(entry.key, "force_go_bssid") &&
>
> os_strcmp() returns 0 on match, i.e., this needs " == 0" to be added to
> work correctly.
>
> > + entry.type == DBUS_TYPE_BOOLEAN) {
> > + force_go_bssid = entry.bool_value;
>
> > if (wpas_p2p_group_add_persistent(wpa_s, ssid, 0, freq, 0, 0, 0,
> > 0, 0, 0, 0, NULL, 0, 0,
> > - false, retry_limit)) {
> > + false, retry_limit, true)) {
>
> true? Was that supposed to be force_go_bssid instead?
>
> > diff --git a/wpa_supplicant/p2p_supplicant.c b/wpa_supplicant/p2p_supplicant.c
>
> > static int wpas_start_p2p_client(struct wpa_supplicant *wpa_s,
> > struct wpa_ssid *params, int addr_allocated,
> > - int freq, int force_scan, int retry_limit)
> > + int freq, int force_scan, int retry_limit,
> > + bool force_go_bssid)
>
> > + if (force_go_bssid && params->bssid_set) {
> > + ssid->bssid_set = 1;
> > + os_memcpy(ssid->bssid, params->bssid, ETH_ALEN);
> > + }
>
> The params->bssid here is the bssid value in the special network profile
> that is used to store the persistent group information. That special
> case uses the bssid value to store the P2P Device Address of the GO; not
> the P2P Interface Address (which would be the BSSID). How is this
> supposed to work if the GO uses P2P Interface Addresses that differ from
> the P2P Device Address?
>
> --
> Jouni Malinen PGP id EFC895FA
_______________________________________________
Hostap mailing list
Hostap@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/hostap
