Return-path: Received: from bar.sig21.net ([80.81.252.164]:49881 "EHLO bar.sig21.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753584Ab0KPQAI (ORCPT ); Tue, 16 Nov 2010 11:00:08 -0500 Date: Tue, 16 Nov 2010 16:59:51 +0100 From: Johannes Stezenbach To: Walter Goldens Cc: "John W. Linville" , Ivo van Doorn , jay_hung@ralinktech.com, linux-wireless@vger.kernel.org, users@rt2x00.serialmonkey.com Subject: Re: [PATCH 7/9] rt2x00: Fix rt2800 USB TX Path DMA issue Message-ID: <20101116155951.GA21461@sig21.net> References: <201011131911.47346.IvDoorn@gmail.com> <310488.65157.qm@web56802.mail.re3.yahoo.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <310488.65157.qm@web56802.mail.re3.yahoo.com> Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, Nov 15, 2010 at 01:45:24AM -0800, Walter Goldens wrote: > > These patches appear to have finally brought some life in to the rt307x chips > in terms of stability and usability. There are no more duplicate (very few) > or lost packets and overall, the driver is improved immensely, however the > latency is increased compared to the STA driver. Pinging my gateway with STA > yields about 2-3ms average, whereas the rt2800usb is about 8-9ms. The > throughput is also about 20% less. I'm runnung rt2800usb (RT3070) in AP mode, and when pinging from the client I get: [12631.727724] phy0 -> rt2x00usb_watchdog_tx_status: Warning - TX queue 2 status timed out, invoke forced tx handler [12632.727451] phy0 -> rt2x00usb_watchdog_tx_status: Warning - TX queue 2 status timed out, invoke forced tx handler [12638.728027] phy0 -> rt2x00usb_watchdog_tx_status: Warning - TX queue 2 status timed out, invoke forced tx handler [12674.732633] phy0 -> rt2x00usb_watchdog_tx_status: Warning - TX queue 0 status timed out, invoke forced tx handler [12676.732855] phy0 -> rt2x00usb_watchdog_tx_status: Warning - TX queue 0 status timed out, invoke forced tx handler [12679.733581] phy0 -> rt2x00usb_watchdog_tx_status: Warning - TX queue 0 status timed out, invoke forced tx handler [12779.745991] phy0 -> rt2x00usb_watchdog_tx_status: Warning - TX queue 0 status timed out, invoke forced tx handler etc. I wonder why it is using TX queue 2? Isn't this QID_AC_VI, i.e. it should only be used when some programs uses setsockopt SO_PRIORITY to request QoS for video data? Lokking through the code, in rt2x00mac_tx(): enum data_queue_qid qid = skb_get_queue_mapping(skb); is passed to rt2x00queue_get_queue() which uses enum data_queue_qid has 2 == QID_AC_VI, but net/core/dev.c dev_pick_tx() calls skb_set_queue_mapping() with 2 == BE (from ieee802_1d_to_ac, return by ieee80211_netdev_select_queue()) Maybe I'm just confused? Queue stats after some flood pinging: # cat /sys/kernel/debug/ieee80211/phy0/rt2800usb/queue/queue qid count limit length index dma done done 14 199365 128 128 69 69 69 0 785 64 1 18 18 17 1 0 64 0 0 0 0 2 196406 64 0 54 54 54 3 0 64 0 0 0 0 16 0 8 0 0 0 0 Johannes