ralphy wrote:
> 33 is the only length I've seen.
> <snip>
> I've added this for all players to 8.0.1r16852 which is now on
> sourceforge and removed the patch above which I had only applied to the
> radio firmware.
Thank you very much. As I raised the point, I thought I'd better have a
look... :)
This is on a Radio:
Code:
--------------------
Mar 30 11:37:58 kernel: [ 180.444967] eth1 (WE) : Wireless Event
(cmd=0x8B1A) too big (33)
Mar 30 11:37:59 kernel: [ 180.459621] AR6000 disconnected from
78:dd:12:cd:a2:f2
Mar 30 11:37:59 kernel: [ 180.602091] eth1 (WE) : Wireless Event
(cmd=0x8B1A) too big (33)
Mar 30 11:37:59 kernel: [ 180.632777] AR6000 disconnected
Mar 30 11:37:59 kernel: [ 181.255741] AR6000 disconnected
Mar 30 11:37:59 kernel: [ 181.264227] eth1 (WE) : Wireless Event
(cmd=0x8B1A) too big (33)
Mar 30 11:38:09 kernel: [ 191.271217] AR6000 disconnected
--------------------
Command -0x8B1A- is -SIOCSIWESSID-. It sets the SSID. SSIDs have a
maximum length of 32 bytes.
After running through the current hostap sources (which incorporate
-wpa_supplicant-), I find a plausible cause:
-wpa_driver_wext_disconnect-
Amongst other things:
> Set a random SSID to make sure the driver will not be trying to
> associate with something even if it does not understand...
Note that the random SSID is of length 32.
https://w1.fi/cgit/hostap/tree/src/drivers/driver_wext.c?h=hostap_2_9#n1985
-wpa_driver_wext_set_ssid-
Will add on extra character (NUL) to the SSID if
-drv->we_version_compiled < 21-.
https://w1.fi/cgit/hostap/tree/src/drivers/driver_wext.c?h=hostap_2_9#n182
So, possibly we have some kind of mismatch here if the wireless driver
is compiled for wireless extensions < 21 (is it ?). Because setting a
random SSID, or any SSID, of length 32 won't work, because
-wpa_supplicant- will bump the length by one, to 33. I won't pretend
that it makes 100% sense to me, I would have thought that it might have
been spotted some time ago. But perhaps there aren't any current bits of
wireless kit that use wireless extensions < 21.
I have always associated the kernel message with
connection/disconnection events, although that's not saying much. But I
think I spin a plausible tale, even though it may not be supported by
hard facts. Hard facts require a more rigorous investigation.
------------------------------------------------------------------------
mrw's Profile: http://forums.slimdevices.com/member.php?userid=38299
View this thread: http://forums.slimdevices.com/showthread.php?t=111663
_______________________________________________
Radio mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/radio
