Jouni Malinen j
Sat Mar 14 02:56:53 PDT 2015

On Thu, Mar 12, 2015 at 08:14:43PM +0100, J?rg Krause wrote:
> I answer myself to my mail. I did some debugging and studied the
> sources. For hostapd AP_EVENT_DISABLED is propagated in
> hostapd_disable_iface() after receiving a 'DISABLE' command.

Yes, but that event was not indicated for other paths (e.g., the main
path of removing a full radio interface which is the case with
wpa_supplicant AP mode).

> In case of wpa_supplicant after receiving a 'DISABLE_NETWORK' for an AP
> network at a certain point wpa_supplicant_mark_disassoc() is called
> which in return calls wpa_supplicant_ap_deinit(). And I think somewhere
> here the event messages AP-DISABLED and CTRL-EVENT-DISCONNECTED should
> be propagated.
> What do you think? Where should the wpa_msg for AP-DISABLED and

wpa_supplicant_ap_deinit() sounds reasonable for the wpa_supplicant
specific CTRL-EVENT-DISCONNECTED event and hostapd_bss_deinit() for
AP-DISABLED. I added these to hostap.git to make the events more

Jouni Malinen                                            PGP id EFC895FA

More information about the Hostap mailing list