Broadcom 4318 not working with bleeding edge kernels

Broadcom 4318 not working with bleeding edge kernels

[Celejar]

[A version of this message was originally posted to the bcm43xx-dev
ML:

https://lists.berlios.de/pipermail/bcm43xx-dev/2009-October/006197.html

I received no response there, so I thought I'd try here.]

My Broadcom 4318 (details below) has been working fine for quite some
time with b43 in mainline.  As recently as 2.6.31, it was still working:

Oct 18 20:30:06 localhost wpa_supplicant[2892]: CTRL-EVENT-SCAN-RESULTS 
Oct 18 20:30:06 localhost wpa_supplicant[2892]: Trying to associate with 00:14:d1:3e:ff:30 (SSID='cranford' freq=2437 MHz)
Oct 18 20:30:06 localhost wpa_supplicant[2892]: Association request to the driver failed
Oct 18 20:30:07 localhost kernel: [   78.196304] wlan0: authenticate with AP 00:14:d1:3e:ff:30
Oct 18 20:30:07 localhost kernel: [   78.197716] wlan0: authenticated
Oct 18 20:30:07 localhost kernel: [   78.197719] wlan0: associate with AP 00:14:d1:3e:ff:30
Oct 18 20:30:07 localhost kernel: [   78.200084] wlan0: RX AssocResp from 00:14:d1:3e:ff:30 (capab=0x431 status=43 aid=0)
Oct 18 20:30:07 localhost kernel: [   78.200088] wlan0: AP denied association (code=43)
Oct 18 20:30:07 localhost kernel: [   78.397059] wlan0: associate with AP 00:14:d1:3e:ff:30
Oct 18 20:30:07 localhost wpa_supplicant[2892]: Associated with 00:14:d1:3e:ff:30
Oct 18 20:30:07 localhost kernel: [   78.399232] wlan0: RX AssocResp from 00:14:d1:3e:ff:30 (capab=0x431 status=0 aid=2)
Oct 18 20:30:07 localhost kernel: [   78.399237] wlan0: associated
Oct 18 20:30:07 localhost kernel: [   78.400055] ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
Oct 18 20:30:08 localhost dnsmasq[1968]: reading /var/run/dnsmasq/resolv.conf
Oct 18 20:30:08 localhost dnsmasq[1968]: using nameserver 192.168.0.1#53
Oct 18 20:30:09 localhost wpa_supplicant[2892]: WPA: Key negotiation completed with 00:14:d1:3e:ff:30 [PTK=CCMP GTK=CCMP]
Oct 18 20:30:09 localhost wpa_supplicant[2892]: CTRL-EVENT-CONNECTED - Connection to 00:14:d1:3e:ff:30 completed (auth) [id=0 id_str=]
Oct 18 20:30:09 localhost kernel: [   80.263676] b43-phy0 debug: Using hardware based encryption for keyidx: 0, mac: 00:14:d1:3e:ff:30
Oct 18 20:30:09 localhost kernel: [   80.263699] b43-phy0 debug: Using hardware based encryption for keyidx: 1, mac: ff:ff:ff:ff:ff:ff

The early 2.6.32 rc kernels were giving oopses, apparently because of
this:

https://lists.berlios.de/pipermail/bcm43xx-dev/2009-October/006129.html

The patch is apparently in -rc5, but that kernel, while it doesn't
oops, fails to authenticate:

Oct 17 19:44:05 localhost wpa_supplicant[3477]: CTRL-EVENT-SCAN-RESULTS 
Oct 17 19:44:07 localhost wpa_supplicant[3477]: CTRL-EVENT-SCAN-RESULTS 
Oct 17 19:44:07 localhost wpa_supplicant[3477]: Trying to associate with 00:14:d1:3e:ff:30 (SSID='cranford' freq=2437 MHz)
Oct 17 19:44:07 localhost kernel: [  198.914180] wlan0: deauthenticating from 00:14:d1:3e:ff:30 by local choice (reason=3)
Oct 17 19:44:07 localhost kernel: [  198.914320] wlan0: direct probe to AP 00:14:d1:3e:ff:30 (try 1)
Oct 17 19:44:07 localhost kernel: [  198.917136] wlan0: direct probe responded
Oct 17 19:44:07 localhost kernel: [  198.917141] wlan0: authenticate with AP 00:14:d1:3e:ff:30 (try 1)
Oct 17 19:44:07 localhost kernel: [  198.918857] wlan0: authenticated
Oct 17 19:44:07 localhost kernel: [  198.918873] wlan0: associate with AP 00:14:d1:3e:ff:30 (try 1)
Oct 17 19:44:07 localhost kernel: [  198.921192] wlan0: RX AssocResp from 00:14:d1:3e:ff:30 (capab=0x431 status=43 aid=0)
Oct 17 19:44:07 localhost kernel: [  198.921196] wlan0: AP denied association (code=43)
Oct 17 19:44:17 localhost wpa_supplicant[3477]: Authentication with 00:14:d1:3e:ff:30 timed out.
Oct 17 19:44:19 localhost wpa_supplicant[3477]: CTRL-EVENT-SCAN-RESULTS 
Oct 17 19:44:19 localhost wpa_supplicant[3477]: Trying to associate with 00:14:d1:3e:ff:30 (SSID='cranford' freq=2437 MHz)
Oct 17 19:44:19 localhost wpa_supplicant[3477]: Association request to the driver failed
Oct 17 19:44:24 localhost wpa_supplicant[3477]: Authentication with 00:14:d1:3e:ff:30 timed out.
Oct 17 19:44:25 localhost wpa_supplicant[3477]: CTRL-EVENT-SCAN-RESULTS 
Oct 17 19:44:25 localhost wpa_supplicant[3477]: Trying to associate with 00:14:d1:3e:ff:30 (SSID='cranford' freq=2437 MHz)
Oct 17 19:44:25 localhost wpa_supplicant[3477]: Association request to the driver failed
Oct 17 19:44:30 localhost wpa_supplicant[3477]: Authentication with 00:14:d1:3e:ff:30 timed out.
Oct 17 19:44:31 localhost wpa_supplicant[3477]: CTRL-EVENT-SCAN-RESULTS 
Oct 17 19:44:31 localhost wpa_supplicant[3477]: Trying to associate with 00:14:d1:3e:ff:30 (SSID='cranford' freq=2437 MHz)
Oct 17 19:44:31 localhost wpa_supplicant[3477]: Association request to the driver failed

Any idea what's going on here?  I can, of course, do a bisection if
necessary, but I'm asking first in case there's a known or suspected
issue that someone can point me to.

[Please cc me on responses.]

Celejar
-- 
mailmin.sourceforge.net - remote access via secure (OpenPGP) email
ssuds.sourceforge.net - A Simple Sudoku Solver and Generator

Re: Broadcom 4318 not working with bleeding edge kernels

[Pavel Roskin]

On Tue, 2009-10-20 at 21:47 -0400, Celejar wrote:

> Oct 17 19:44:19 localhost wpa_supplicant[3477]: Trying to associate with 00:14:d1:3e:ff:30 (SSID='cranford' freq=2437 MHz)
> Oct 17 19:44:19 localhost wpa_supplicant[3477]: Association request to the driver failed

Since the message is from wpa_supplicant, it must be something between
wpa_supplicant and mac80211.  I don't think any code in b43 is involved.

Perhaps you should include some information about wpa_supplicant, i.e.
its version and the command line.  Please try the latest wpa_supplicant.
Please try both wext and nl80211 drivers.  That would help understand
the extent of the problem.

-- 
Regards,
Pavel Roskin

Re: Broadcom 4318 not working with bleeding edge kernels

[Celejar]

On Wed, 21 Oct 2009 11:29:17 -0400
Pavel Roskin <proski@gnu.org> wrote:

> On Tue, 2009-10-20 at 21:47 -0400, Celejar wrote:
> 
> > Oct 17 19:44:19 localhost wpa_supplicant[3477]: Trying to associate
> > with 00:14:d1:3e:ff:30 (SSID='cranford' freq=2437 MHz) Oct 17
> > 19:44:19 localhost wpa_supplicant[3477]: Association request to the
> > driver failed
> 
> Since the message is from wpa_supplicant, it must be something between
> wpa_supplicant and mac80211.  I don't think any code in b43 is
> involved.
> 
> Perhaps you should include some information about wpa_supplicant, i.e.
> its version and the command line.  Please try the latest
> wpa_supplicant. Please try both wext and nl80211 drivers.  That would
> help understand the extent of the problem.

Version is 0.6.9, from the Debian package version 0.6.9.3.  I don't
know the exact command line, since I've been invoking it through the
Debian ifupdown scripts.  I haven't been specifying a backend, and the
Debian documentation states that wext is used in that case.

I have been playing with it some more, and it sometimes works, and
sometimes doesn't.  The only absolutely consistent pattern that I've
seen is that (in my limited testing) it seems to always work right
after unloading / reloading the driver (b43) (or a fresh boot).

As I mentioned, I've been using the Debian ifupdown scripts.  I have
now tried running wpa_supplicant directly, and so far it has worked
consistently, with a very minimal, mostly default config:

'wpa_supplicant -iwlan0 -c/wpa.conf'

where wpa.conf contains:

ctrl_interface=/var/run/wpa_supplicant

network={
        ssid="my-ssid"
        psk="my passphrase"
}


So at this point, I guess the problem is something that I need to
report to Debian, some sort of bugginess with their scripts.  Thanks
for the help ;)

Celejar
-- 
mailmin.sourceforge.net - remote access via secure (OpenPGP) email
ssuds.sourceforge.net - A Simple Sudoku Solver and Generator

Re: Broadcom 4318 not working with bleeding edge kernels

[Hin-Tak Leung]

On Wed, Oct 21, 2009 at 10:34 PM, Celejar <celejar@gmail.com> wrote:
> On Wed, 21 Oct 2009 11:29:17 -0400
> Pavel Roskin <proski@gnu.org> wrote:
>
>> On Tue, 2009-10-20 at 21:47 -0400, Celejar wrote:
>>
>> > Oct 17 19:44:19 localhost wpa_supplicant[3477]: Trying to associate
>> > with 00:14:d1:3e:ff:30 (SSID='cranford' freq=2437 MHz) Oct 17
>> > 19:44:19 localhost wpa_supplicant[3477]: Association request to the
>> > driver failed
>>
>> Since the message is from wpa_supplicant, it must be something between
>> wpa_supplicant and mac80211.  I don't think any code in b43 is
>> involved.
>>
>> Perhaps you should include some information about wpa_supplicant, i.e.
>> its version and the command line.  Please try the latest
>> wpa_supplicant. Please try both wext and nl80211 drivers.  That would
>> help understand the extent of the problem.
>
> Version is 0.6.9, from the Debian package version 0.6.9.3.  I don't
> know the exact command line, since I've been invoking it through the
> Debian ifupdown scripts.  I haven't been specifying a backend, and the
> Debian documentation states that wext is used in that case.

'ps -lax -w -w' should show the full command line of currently running
processes. Read the documentation of ps.

also, have a look inside /var/log/wpa_supplicant.log (or wherever
Debian puts it) for the log.

> I have been playing with it some more, and it sometimes works, and
> sometimes doesn't.  The only absolutely consistent pattern that I've
> seen is that (in my limited testing) it seems to always work right
> after unloading / reloading the driver (b43) (or a fresh boot).
>
> As I mentioned, I've been using the Debian ifupdown scripts.  I have
> now tried running wpa_supplicant directly, and so far it has worked
> consistently, with a very minimal, mostly default config:
>
> 'wpa_supplicant -iwlan0 -c/wpa.conf'
>
> where wpa.conf contains:
>
> ctrl_interface=/var/run/wpa_supplicant
>
> network={
>        ssid="my-ssid"
>        psk="my passphrase"
> }
>
>
> So at this point, I guess the problem is something that I need to
> report to Debian, some sort of bugginess with their scripts.  Thanks
> for the help ;)
>
> Celejar
> --
> mailmin.sourceforge.net - remote access via secure (OpenPGP) email
> ssuds.sourceforge.net - A Simple Sudoku Solver and Generator
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>