Return-path: Received: from perninha.conectiva.com.br ([200.140.247.100]:48871 "EHLO perninha.conectiva.com.br" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753342AbYK0VwJ convert rfc822-to-8bit (ORCPT ); Thu, 27 Nov 2008 16:52:09 -0500 From: Herton Ronaldo Krzesinski To: "Stefanik =?iso-8859-1?q?G=E1bor?=" Subject: Re: [RFC] rtl8187: Do not wait for an ACK when IEEE80211_TX_CTL_NO_ACK is set Date: Thu, 27 Nov 2008 19:52:19 -0200 Cc: "linux-wireless" , "Hin-Tak Leung" , "Larry Finger" , "Johannes Berg" , "John W. Linville" References: <69e28c910811261431u26e341d3p51ebadca807f4b61@mail.gmail.com> <200811271543.59126.herton@mandriva.com.br> <69e28c910811271202s295bbf5bo3dc8c04c50b08a01@mail.gmail.com> In-Reply-To: <69e28c910811271202s295bbf5bo3dc8c04c50b08a01@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Message-Id: <200811271952.21408.herton@mandriva.com.br> (sfid-20081127_225215_066688_243428EB) Sender: linux-wireless-owner@vger.kernel.org List-ID: Em Quinta 27 Novembro 2008, =E0s 18:02:22, Stefanik G=E1bor escreveu: > On Thu, Nov 27, 2008 at 6:43 PM, Herton Ronaldo Krzesinski > > wrote: > > nack. > > > > I've seen other drivers set just a flag when hardware has a bit fie= ld to > > signalize that there isn't need to ack the tx packet to be sent, bu= t > > don't saw anyone changing retry in tx hw header. Also setting to 0 = in > > no_ack case looks wrong, for example see ath5k where it even checks= for > > 0 and returns an error if you try to do this. > > > > -- > > []'s > > Herton > > IMO the current implementation is even worse - completely disregard > NO_ACK, and when a packet that should not be ACKed according to IEEE > 802.11:2007 is sent (such as a broadcast), retry repeatedly until > timeout. In this case then shouldn't mac80211/rate control alg give a right rate= s[0].count? And other drivers have same bug? But I agree that we should set to 0 if to the hardware it means that 0 = is the same meaning to the no ack bitfield on others; only there is no referen= ce/docs mentioning this. > Also, v26.1010 of the vendor driver does exactly what my patch does. Is this newer than 26.1036.0708.2008 I have here? The version it never = sets 0, in code there is this: if( WLAN_FC_GET_STYPE(le16_to_cpu(frag_hdr->frame_ctl))= =3D=3D IEEE80211_STYPE_PROBE_RESP ) tx[5] |=3D (1<<8);//(priv->retry_data<<8); //re= try lim ; else tx[5] |=3D (11<<8);//(priv->retry_data<<8); //r= etry lim ; I didn't found this version 26.1010 to download, do you know where I ca= n get? -- []'s Herton -- To unsubscribe from this list: send the line "unsubscribe linux-wireles= s" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html