Hi
On 03/28/2014 03:59 PM, Tomasz Bursztyka wrote:
Hi,
Due to the fact wpa_supplicant might not (it's a bug) send the device
address, the peer identifier will be constructed on its name in this
case.
I have a suggestion that maybe we can get the device address from the
obj_path temporarily.
You know, name is not unique to distinguish the different remote peers.
I'm afraid that we may meet many issues with name as a identifier.
No, that would be really ugly to "parse" the object path.
I did a patch for wpa_supplicant which exposes peer's mac address and
it's upstream now.
It's wpa_supplicant's DBus API which was broken.
Falling back to the name is just for the time being so it can still
work, not optimally for sure but issues are limited anyway.
It will still take some time to get p2p fully supported to connman,
once done, there will probably
be a new version of wpa_supplicant with my fixes.
Fine. Thanks
I have an another question, this patch only use a remote device address
as a identifier,
it may doesn't work well when mutli-p2p interface. maybe this question
is too
early this time, but the case about mutli-p2p interface is a reasonable.
so what I mean is that using a combination of local device address and
remote device address
as a identifier maybe have a good expansibility:dfd, just like the wifi
service identifier.
Cheers
Guoqiang
_______________________________________________
connman mailing list
[email protected]
https://lists.connman.net/mailman/listinfo/connman
