From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ben Greear Date: Tue, 01 Feb 2011 22:55:51 -0800 Subject: [ath9k-devel] Weird error messages in logs In-Reply-To: References: Message-ID: <4D48FFF7.1050908@candelatech.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ath9k-devel@lists.ath9k.org On 02/01/2011 09:57 PM, Mohammed Shafi wrote: > On Wed, Feb 2, 2011 at 11:18 AM, Nikolay Martynov wrote: >> Hi. >> >> Have ath9k AP and intel 5300 client. AP has latest openwrt, intel5300 >> has latest linux driver. Network operates in 801.11g mode. When client >> uploads to the AP for a long time I get these errors in logs on AP: >> >> ath: txq: 80c3d7fc axq_qnum: 2, mac80211_qnum: 2 axq_link: a0d575f8 >> pending frames: 7 axq_acq empty: 0 stopped: 0 axq_depth: 0 Attempting >> to restart tx logic. >> ath: txq: 80c3d7fc axq_qnum: 2, mac80211_qnum: 2 axq_link: a0d58f5c >> pending frames: 124 axq_acq empty: 0 stopped: 1 axq_depth: 0 >> Attempting to restart tx logic. >> >> Transmission seems to die after this. >> Could you please tell me what these messages mean and is it possible >> to fix this issue on ath9k side? > > Looks like a tx hang and this commit will provide us more details > > commit 60f2d1d506195803fa6e1dcf3972637b740fdd60 > Author: Ben Greear > Date: Sun Jan 9 23:11:52 2011 -0800 > > ath9k: Restart xmit logic in xmit watchdog. > > The system can get into a state where the xmit queue > is stopped, but there are no packets pending, so > the queue will not be restarted. > > Add logic to the xmit watchdog to attempt to restart > the xmit logic if this situation is detected. > > Example 'dmesg' output: > > ath: txq: f4e723e0 axq_qnum: 2, mac80211_qnum: 2 axq_link: > f4e996c8 pending frames: 1 axq_acq empty: 1 stopped: 0 axq_depth: 0 > Attempting to restart tx logic. > > sudo modprobe ath9k debug=0x481(FATAL | XMIT | RESET) will provide us > some useful debug messages Looks like my attempt to restart is not working, for whatever reason. There are several ways it could get stuck...my hack worked around some of them, but obviously not all of them. You might also look at the debugfs files while the system is in the stuck state: cat /debug/ieee80211/wiphy0/ath9k/stations cat /debug/ieee80211/wiphy0/ath9k/xmit Those show info that *might* help debug the problem... Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com