Return-path: Received: from mail-pz0-f42.google.com ([209.85.210.42]:65401 "EHLO mail-pz0-f42.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753638Ab1HNUyU convert rfc822-to-8bit (ORCPT ); Sun, 14 Aug 2011 16:54:20 -0400 Received: by pzk37 with SMTP id 37so2037868pzk.1 for ; Sun, 14 Aug 2011 13:54:20 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <20110814214626.4413d7cf@boulder.homenet> References: <4CEAB969.20702@lwfinger.net> <1290451982.20888.2.camel@maggie> <4CEAC095.7020706@lwfinger.net> <4DC9853A.1090508@lwfinger.net> <20110601114839.433ae42d@boulder.homenet> <20110601130842.077da1c3@boulder.homenet> <20110601160101.75e30b3d@boulder.homenet> <20110601184931.499557c6@boulder.homenet> <20110814214626.4413d7cf@boulder.homenet> Date: Sun, 14 Aug 2011 22:54:20 +0200 Message-ID: (sfid-20110814_225424_136450_B4481DB5) Subject: Re: b43 error under heavy load From: =?UTF-8?B?UmFmYcWCIE1pxYJlY2tp?= To: Chris Vine Cc: Larry Finger , wireless , =?UTF-8?Q?Michael_B=C3=BCsch?= , b43-dev Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: 2011/8/14 Chris Vine : > On Sun, 14 Aug 2011 13:11:12 +0200 > Rafał Miłecki wrote: >> 2011/8/14 Rafał Miłecki : >> > Hey Chris, >> > >> > 2011/6/1 Chris Vine : >> >> Summary: Traffic sent up from the broadcom wireless device >> >> generates copious reports of "Stopped TX ring 1" but always >> >> carries on with its job and stays up, although its traffic is >> >> slower than on received packets.  Received traffic on the other >> >> hand reports no errors until the spate of "Out of order TX status >> >> report on DMA ring 1" errors occurs, which seems to happen at >> >> random (albeit accompanied on my failed transfer by a single >> >> "Stopped TX ring 1" log entry), and when it does happen brings the >> >> wireless link to a halt. Wireless traffic can be restarted simply >> >> by reassociating with the AP. >> > >> > So it seems heavy load of RX can cause problems with TX. I just >> > though of one another reason, could you test one more thing for me? >> > >> > Edit drivers/net/wireless/b43/dma.h and change define from value 64 >> > to in the line: >> > #define B43_RXRING_SLOTS                64 >> >> I meant "to 8": >> #define B43_RXRING_SLOTS                8 > > It is not an improvement.  Different errors occur earlier (approx 100MB > of download), upon which the wireless disassociates from the network. > dmesg produces the output at the end of this post (with MAC addresses > munged). I didn't expect improvement, I tried to make it worse :P I though our problems may be coming from not-fast-enough reading RX ring. So you just made that ring even smaller to allow faster/easier errors. What do you mean by different errors? The only one I can see is "b43-phy0 ERROR: PHY transmission error". Do you mean that or something else? Did you post all the log you wanted to? Maybe missed beginning? -- Rafał