From mboxrd@z Thu Jan 1 00:00:00 1970 From: Felix Fietkau Date: Thu, 10 Mar 2011 01:40:07 +0100 Subject: [ath9k-devel] DMA issues with ar9280 cards In-Reply-To: <201102261037.56348.bschmidt@techwires.net> References: <201102230955.35674.bschmidt@techwires.net> <201102251037.07890.bschmidt@techwires.net> <20110225171209.GS19293@tux> <201102261037.56348.bschmidt@techwires.net> Message-ID: <4D781DE7.40102@openwrt.org> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ath9k-devel@lists.ath9k.org On 2011-02-26 10:37 AM, Bernhard Schmidt wrote: > On Friday 25 February 2011 18:12:09 Luis R. Rodriguez wrote: >> On Fri, Feb 25, 2011 at 01:37:07AM -0800, Bernhard Schmidt wrote: >> > On Thursday, February 24, 2011 19:18:25 Luis R. Rodriguez wrote: >> > > On Thu, Feb 24, 2011 at 01:33:18AM -0800, Bernhard Schmidt wrote: >> > > > I've stripped everything not necessary, this one is use for the >> > > > log below. >> > > > >> > > > # cat /etc/hostapd.conf >> > > > interface=wlan0 >> > > > driver=nl80211 >> > > > ssid=testtest >> > > > hw_mode=a >> > > > channel=44 >> > > > >> > > > [ 574.605956] ath: Enable TXE on queue: 9 >> > > > [ 574.708386] ath: Enable TXE on queue: 9 >> > > > [ 574.810752] ath: Enable TXE on queue: 9 >> > > > [ 574.913151] ath: Enable TXE on queue: 9 >> > > > [ 575.015541] ath: Enable TXE on queue: 9 >> > > > [ 575.117938] ath: Enable TXE on queue: 9 >> > > > [ 575.220335] ath: Enable TXE on queue: 9 >> > > > [ 575.322732] ath: Enable TXE on queue: 9 >> > > > [ 576.248395] ath: ath9k_hw_stoptxdma: Num of pending TX >> > > > Frames 1 on Q 9 [ 576.259188] ath: Failed to stop TX DMA in >> > > > 100 msec after killing last frame >> > > >> > > Can you confirm what kernel ? >> > > >> > > I see 2.6.38-rc6-wl-12466-g09f3227 >> > > >> > > So I take it you are using wireless-testing, is this right? Can >> > > you reproduce with wireless-testing as of today? This would be >> > > helpful for our own testing. >> > >> > Yes, the logs were posted using wireless-testing, always pulled >> > before doing the tests. >> > >> > Just to confirm, with wireless-testing as of a few minutes ago, >> > same behavior. >> > >> > # uname -a >> > Linux meshnode 2.6.38-rc6-wl-12527-g8ccd3c8 #25 SMP Fri Feb 25 >> > 10:19:49 CET 2011 i686 GNU/Linux # git log --pretty=oneline | head >> > -n1 >> > 8ccd3c862a39d46626ba89777e59d1775a579b62 Merge >> > ssh://master.kernel.org/pub/scm/linux/kernel/git/linville/wireless >> > -next-2.6 >> > >> > [ 176.675020] ath: Enable TXE on queue: 9 >> > [ 176.777416] ath: Enable TXE on queue: 9 >> > [ 176.827134] ath: qnum: 0, txq depth: 0 >> > [ 176.831007] ath: Enable TXE on queue: 0 >> > [ 176.834979] ath: tx queue 0 (2f1b01f4), link ef1b01f4 >> > [ 176.840244] ath: tx queue 0 (2f1b01f4), link ef1b01f4 >> > [ 176.879857] ath: Enable TXE on queue: 9 >> > [ 176.982217] ath: Enable TXE on queue: 9 >> > .. >> > [ 202.786342] ath: Enable TXE on queue: 9 >> > [ 202.888732] ath: Enable TXE on queue: 9 >> > [ 203.814434] ath: ath9k_hw_stoptxdma: Num of pending TX Frames 1 >> > on Q 9 [ 203.825260] ath: Failed to stop TX DMA in 100 msec after >> > killing last frame >> > >> > [ 203.836898] ath: Reset TX queue: 0 >> > [ 203.840352] ath: Reset TX queue: 1 >> > [ 203.843794] ath: Reset TX queue: 2 >> > [ 203.847238] ath: Reset TX queue: 3 >> > [ 203.850683] ath: Reset TXQ, inactive queue: 4 >> > [ 203.855077] ath: Reset TXQ, inactive queue: 5 >> > [ 203.859469] ath: Reset TXQ, inactive queue: 6 >> > [ 203.863864] ath: Reset TXQ, inactive queue: 7 >> > [ 203.868258] ath: Reset TX queue: 8 >> > [ 203.871709] ath: Reset TX queue: 9 >> > [ 203.877917] ath: Set queue properties for: 9 >> > [ 203.882244] ath: Reset TX queue: 9 >> > [ 203.978048] ath: Enable TXE on queue: 9 >> > [ 204.903691] ath: ath9k_hw_stoptxdma: Num of pending TX Frames 1 >> > on Q 9 [ 204.914536] ath: Failed to stop TX DMA in 100 msec after >> > killing last frame [ 204.933408] ath: DMA failed to stop in 10 ms >> > AR_CR=0x00000024 AR_DIAG_SW=0x42000020 [ 204.941106] ath: Could >> > not stop RX, we could be confusing the DMA engine when we start RX >> > up [ 204.949576] ------------[ cut here ]------------ >> > [ 204.954249] WARNING: at >> > drivers/net/wireless/ath/ath9k/recv.c:509 ath_stoprecv+0xb7/0xc9 >> > [ath9k]() >> >> Ok neat.. how soon after starting hostapd does this come up, >> immeidiately? > > 2 mails back is a complete log including timestamps, about 30 seconds > from hostapd start until the error. > >> What version of hostapd? > > hostap.git @cbcf92b4 I just sent out a 4-patch series to linux-wireless@ which almost completely fixes all the tx dma stop issues that I could reproduce in my setup. Please give it a try and let me know if it fixes your issues as well. - Felix