[Prism54-devel] Re: Strange problem with wds in prism54

[Jean-Baptiste Note]

Hello Sergio,

It seems that i can rely on you to check on more things, so i'll take
advantage of this if you don't mind :)

> wds_net_send_packet
> wds_find_device
> wds_find_device: WDS from mac 01-80-c2-00-00-00 not in current config,
> returning NULL

Well, either this mac address is meaningfull for the device, or, well,
it looks like the beginning of a frame in softmac mode... Maybe a trap
response ? -- or rather an rx annex, that may be present in WDS mode,
but is not checked for.

The only way to know would be to display the *whole* buffer.

can you add, after

*** islpci_eth.c:
islpci_eth_transmit[441]       wds_dev = wds_find_device( skb->data, wdsp );

the following,
    
/* no sure about the memcmp check, remove if nothing appears */
if ( (wds_dev == NULL) && (memcmp(skb->data,ndev->broadcast,6) != 0) )
	{
                /* dumps the frames which have no wds entry and are
                non-broadcast. I think they are frames with an rx_annex */

		DEBUG(SHOW_BUFFER_CONTENTS, "\nrx %p ", skb->data);
		display_buffer((char *) skb->data, skb->len);
	}

so that we can see if the WDS frames with this strange mac really are
WDS frames ?

> wds_net_send_packet
> wds_find_device
> wds_find_device: WDS from mac ff-ff-ff-ff-ff-ff not in current config,
> returning NULL
> wds_net_send_packet

OK, those ones are more easily explained : I don't think they can come
from the WDS links, and they seem to be the ones explicitly directed to
the main device.

This is not a problem, as those packets are "by default" sent to the
primary interface. Yet we should certainly distinguish the
ff:ff:ff:ff:ff case which is "standard" from the other case which does
not seem to be... (i'll do the patch when i get time).

In order to go farther, i'd really appreciate a complete dump of the
packets starting with the strange MAC...

Thanks for all the work

JB

-- 
Jean-Baptiste Note
+33 (0)6 83 03 42 38
jean-baptiste.note at wanadoo.fr