[PATCH] wpa_supplicant: send event to all interfaces

[Janusz Dziedzic]

On 25 April 2016 at 10:55, Jouni Malinen <j at w1.fi> wrote:
> On Mon, Apr 25, 2016 at 06:57:59AM +0200, Janusz Dziedzic wrote:
>> On 24 April 2016 at 10:42, Jouni Malinen <j at w1.fi> wrote:
>> > On Sat, Apr 02, 2016 at 07:20:45PM +0200, Janusz Dziedzic wrote:
>> >> This is test with hwsim:
>> >
>> >> diff --git a/tests/hwsim/start.sh b/tests/hwsim/start.sh
>> >> @@ -100,7 +100,7 @@ else
>> >> -test -f /proc/modules && sudo modprobe mac80211_hwsim radios=7
>> >> channels=$NUM_CH support_p2p_device=0
>> >> +test -f /proc/modules && sudo modprobe mac80211_hwsim radios=7
>> >> channels=$NUM_CH support_p2p_device=1
>
> Ah.. I was not looking at this closely enough. This start.sh change does
> not actually do anything for my main test setup, since I use a VM and a
> kernel without any modules..
>
>> In ibss_5ghz.log I see we only get this event for p2p-dev-wlan1/p2p-dev-wlan0.
>
> I can see this now with vm/vm-run.sh handling the support_p2p_device
> modification.
>
> Now that I can reproduce this issue, lets try to figure out what's the
> best way of addressing this. Your patch is doing quite a bit more than
> just modifying the specfici CTRL-EVENT-REGDOM-CHANGE event. Was there a
> reason for doing that or is that all just trying to make these test
> cases pass? Would the simpler patch that Ilan sent today work for the
> cases where you can see this issue with remote hosts? It is simply
> finding the highest level parent to get the IFNAME=wlan0 style prefix
> for the event regardless of whether the P2P Device interface is used.
>

I also remove for_each() from wpa_supplicant_update_channel_list()
while this function will be called now for each interface.
So, seems this for_each() was a workaround for a real issue (not
sending this message to all ifaces).

This is general question:
how should we handle events when ifidx==-1 and wdev_id is not set.

BR
Janusz



> --
> Jouni Malinen                                            PGP id EFC895FA