[ath9k-devel] Kernel panic + HT question

[ath9k-devel] Kernel panic + HT question

[Kostas Nakos]

Hello list,

I have a pair of ath9k wireless adapters. One of them is on a 
powerpc-based board, with compat-wireless cross compiled, acting as an 
AP with hostapd; the other one is on a x86 pc using a mini-pci to pci 
card, also running compat-wireless. Both cards are identical, as ath9k 
reports the following info:

phy0: Atheros AR9160 MAC/BB Rev:1 AR5133 RF Rev:b0:

The host and the station run the 2.6.28 kernel. The host (powerpc) has 
compat-wireless-20090205 while the client (pc) has compat-wireless-20090211.

The client connects and works OK when the AP is set to use channel 9 and 
  ieee80211n=1 in hostapd. When we additionally supply ht_capab=[HT40-] 
to hostapd, the client crashes with a kernel panic. Please find 
attached: messages.txt which contains the panic message of such a crash 
from /var/log/messages, and ath9kcrash.txt which contains the first few 
functions as reported in such a crash in the console of the pc and 
written down by hand by me.

If it is any help, searching around, as this issue has been around for 
at least a month in compat-wireless, I stumbled upon a similar bug 
report here:
https://bugzilla.redhat.com/show_bug.cgi?id=472372
which John has marked as a duplicate bug.



As an aside, it'd be nice to know if such a setup can achieve actual 
tx/rx rates that are in the neighborhood of 11n, e.g. greater than g 
speeds, using the setup we have: 2xath9k+hostapd+wpa_supplicant. I may 
be jumping the gun here, but we've been trying to test for this the last 
month and the crashes deter us.

Best Regards,
Kostas
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: messages.txt
Url: http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20090212/e627d842/attachment.txt 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: ath9kcrash.txt
Url: http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20090212/e627d842/attachment-0001.txt 

[ath9k-devel] Kernel panic + HT question

[Luis R. Rodriguez]

On Thu, Feb 12, 2009 at 03:19:42AM -0800, Kostas Nakos wrote:
> Hello list,
> 
> I have a pair of ath9k wireless adapters. One of them is on a
> powerpc-based board, with compat-wireless cross compiled, acting as an
> AP with hostapd; the other one is on a x86 pc using a mini-pci to pci
> card, also running compat-wireless. Both cards are identical, as ath9k
> reports the following info:
> 
> phy0: Atheros AR9160 MAC/BB Rev:1 AR5133 RF Rev:b0:
> 
> The host and the station run the 2.6.28 kernel. The host (powerpc) has
> compat-wireless-20090205 while the client (pc) has compat-wireless-20090211.
> 
> The client connects and works OK when the AP is set to use channel 9 and
>   ieee80211n=1 in hostapd. When we additionally supply ht_capab=[HT40-]
> to hostapd, the client crashes with a kernel panic. Please find
> attached: messages.txt which contains the panic message of such a crash
> from /var/log/messages, and ath9kcrash.txt which contains the first few
> functions as reported in such a crash in the console of the pc and
> written down by hand by me.
> 
> If it is any help, searching around, as this issue has been around for
> at least a month in compat-wireless, I stumbled upon a similar bug
> report here:
> https://bugzilla.redhat.com/show_bug.cgi?id=472372
> which John has marked as a duplicate bug.
> 
> 
> 
> As an aside, it'd be nice to know if such a setup can achieve actual
> tx/rx rates that are in the neighborhood of 11n, e.g. greater than g
> speeds, using the setup we have: 2xath9k+hostapd+wpa_supplicant. I may
> be jumping the gun here, but we've been trying to test for this the last
> month and the crashes deter us.
> 
> Best Regards,
> Kostas

Content-Description: messages.txt
> Feb 12 12:15:40 bender kernel: [   73.168441] ------------[ cut here ]------------
> Feb 12 12:15:40 bender kernel: [   73.168444] WARNING: at /home/knakos/Work/compat-wireless-2009-02-11/include/net/mac80211.h:2048 ath_get_rate+0x9f/0x600 [ath9k]()

I don't see a kernel panic, I see a kernel warning, which is non-fatal,
it should not crash your box. This rate control related so will let someone
more familiar with rate control comment on that.

  Luis

[ath9k-devel] Kernel panic + HT question

[Kostas Nakos]

Luis R. Rodriguez wrote:
> I don't see a kernel panic, I see a kernel warning, which is non-fatal,
> it should not crash your box.

Whoops, you're right. I included that log from messages at the last 
moment before sending the email and didn't check it much. In fact I was 
unable to get a log of the panic all this time and was surprised to see, 
what seemed to me at the time, as related info. So this may not be 
related too much/at all. But the kernel panic is real, and the other 
attachment describes the stack dump as output on the console.

Thanks,
Kostas