[PATCHv2 0/3] Improving the IBSS/RSN mechanism

Antonio Quartulli ordex
Wed Aug 7 01:16:38 PDT 2013


On Sun, Jul 21, 2013 at 04:04:59PM +0300, Jouni Malinen wrote:
> On Tue, Jul 02, 2013 at 12:04:57PM +0200, Antonio Quartulli wrote:
> > with this patchset based on the work done by Nicolas Cavallari, I want to bring
> > some improvements to the IBSS/RSN support in wpa_supplicant.
> > 
> > So far there were some issues due to peer reboots or faulty key exchange. The
> > solution was to detect any peer re-initialisation by explicitly using the Open
> > Authentication Systems. This feature was implemented in kernel space, but, of
> > course, it was triggering several race conditions due to the fact that the
> > kernel was re-initialising the peer state (DEL/NEW_STATION events in a row)
> > while wpa_supplicant was performing a key exchange.
> > 
> > To avoid the issue the idea has been to move the entire logic into
> > wpa_supplicant so that it can first perform the Open Authentication and only
> > then perform the EAPOL key exchange. Of course, it a peer does not support the
> > Open Auth System, wpa_supplicant will still work but will not be able to perform
> > the reboot detection.
> 
> Thanks, applied with some changes. I made the use of Authentication
> frame exchange conditional on the driver being able to accept the
> command to transmit an Authentication frame. Without this, the patch
> broke IBSS RSN for all drivers that did not support this new mechanism.

Ah right (even if IBSS/RSN will probably not work properly on those drivers
anyway).

Thanks a lot!

Cheers,

-- 
Antonio Quartulli

..each of us alone is worth nothing..
Ernesto "Che" Guevara
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.shmoo.com/pipermail/hostap/attachments/20130807/ad54c0ab/attachment.pgp>



More information about the Hostap mailing list