> De : Bob Copeland [mailto:[email protected]]
> Envoyé : vendredi 28 novembre 2014 15:05
> À : Cédric Debarge - ACKSYS; [email protected]
> Cc : 'Yeoh Chun-Yeow'
> Objet : Re: 80211s communication problem
>
> Hmm, you're not allowed to change the supported rates in
> NL80211_CMD_SET_STATION in at least recent kernel versions. We already
> set them in peer_created().
>
> So, I'd try this completely untested patch:
>
> diff --git a/linux/meshd-nl80211.c b/linux/meshd-nl80211.c index
> e361de8..76b67c5 100644
> --- a/linux/meshd-nl80211.c
> +++ b/linux/meshd-nl80211.c
> @@ -1148,7 +1148,6 @@ void estab_peer_link(unsigned char *peer,
> /* to check integrity of multicast mgmt frames from this peer */
> install_key(&nlcfg, peer, CIPHER_AES_CMAC,
> NL80211_KEYTYPE_GROUP, 4, peer_mgtk);
>
> - set_supported_rates(&nlcfg, peer, rates, rates_len);
> }
> }
Thanks for this patch, it works great and suppresses my first error
(NL80211_CMD_SET_STATION in new_unauthenticated_peer). I am now able to
communicate between my two products. However, this does not correct the
second
error message nor the weird malformed packets during peering (the malformed
packet are the same than previously).
But it made me realise that the products are using 802.11a rate to
communicate
(displayed in the PPI header of each frame) despite the fact that I
configured
802.11na (channel 132 + HT20) into meshd configuration file (see the
configuration in my first email). Some traces in the meshd deamon show that
the
the right channel and 802.11 mode are extracted correctly.
I don't if it is related or not but when I try a bidirectional UDP iperf at
15Mbps (see diagram below). After a few seconds, one of two iperf stops
communicating. A few seconds later, the second iperf also stops and I am not
able to communicate anymore between the two products. I am aware that
2*15Mpbs
is excessive for a 802.11a connexion but I cannot understand why the
products
suddenly stop to communicate. I was just expecting receiving less than
2*15Mpbs.
__ _ Ethernet |_|_| |_|_| Ethernet __ _
[__]|=|--------->[____°] [____°]--------->[__]|=|
/::/|_| /::/|_|
iperf iperf
I made the same test without SAE, and did not encounter any of this
problems.
Thanks for your help.
Cédric
_______________________________________________
Devel mailing list
[email protected]
http://lists.open80211s.org/cgi-bin/mailman/listinfo/devel
