[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4E9EFF51.6000104@broadcom.com>
Date: Wed, 19 Oct 2011 18:48:17 +0200
From: "Arend van Spriel" <arend@...adcom.com>
To: "Nico Schottelius" <nico-linux-20111017@...ottelius.org>,
"Eric Dumazet" <eric.dumazet@...il.com>,
LKML <linux-kernel@...r.kernel.org>,
"linux-wireless@...r.kernel.org" <linux-wireless@...r.kernel.org>,
netdev <netdev@...r.kernel.org>
Subject: Re: BUG: All network processes hang (brcmsmac/wpa_supplicant)
On 10/19/2011 04:55 PM, Nico Schottelius wrote:
> Hey Arend,
>
>
> Arend van Spriel [Wed, Oct 19, 2011 at 04:43:27PM +0200]:
>> I actually don't see the rtnl_lock or mutex_lock that was listed in your
>> original dmesg. There are a lot of warning traces from mac80211.c.
The warnings are probably a diversion. This is the warning in
ieee80211_rx():
if (WARN_ON(status->rate_idx < 0 ||
status->rate_idx >= sband->n_bitrates))
goto drop;
rate = &sband->bitrates[status->rate_idx];
This means that the driver (brcmsmac) provides a data packet with an
out-of-range rate index. This reminded me of a fix I made about a month
ago. Could you apply the attached patch file. It is based on:
commit bee709ab1d390afe69e4407bc86bb706c6fb2965
Merge: ad1c761 1f2c7e9
Author: Nico Schottelius <nico@...ethz.ch>
Date: Tue Oct 18 00:04:05 2011 +0200
Merge branch 'fix-edp-vdd-power' of ../keithp/linux
As it drops receive packets it may be your problem. Is your AP on 5GHz?
Gr. AvS
View attachment "0001-staging-brcm80211-fix-for-rate-index-in-receive-stat.patch" of type "text/x-patch" (1067 bytes)
Download attachment "signature.asc" of type "application/pgp-signature" (901 bytes)
Powered by blists - more mailing lists