From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail-ve0-f169.google.com ([209.85.128.169]:45643 "EHLO mail-ve0-f169.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751634AbaCUOW7 (ORCPT ); Fri, 21 Mar 2014 10:22:59 -0400 Received: by mail-ve0-f169.google.com with SMTP id pa12so2669845veb.28 for ; Fri, 21 Mar 2014 07:22:59 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: References: <1395396014-24631-1-git-send-email-yeohchunyeow@gmail.com> Date: Fri, 21 Mar 2014 22:22:59 +0800 Message-ID: (sfid-20140321_152314_340730_26B943A5) Subject: Re: [PATCH] ath10k: add the Rx rate in FW stats From: Yeoh Chun-Yeow To: Michal Kazior Cc: "ath10k@lists.infradead.org" , linux-wireless , Kalle Valo Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-wireless-owner@vger.kernel.org List-ID: > Unfortunately this isn't as simple as it seems. > > Stats are received via wmi_stats_event. This structure contains fields > describing number of pdev/vdev/beer/bcn stat entries. All entries > (structures) are appended in certain order after this header. This > means if you change length of structure type you skew the calculation > of all succeeding entries. > > IOW this means this patch breaks peer stats for 999.999.0.636. > > I'm quite positive wmi_pdev_stats is also broken. This implies vdev > stats and peer stats position calculation is skewed in the first place > and you get garbage. > > The main problem here is there are subtle yet crazy binary > incompatibilities between these firmwares. > > The best way would probably be to implement wmi as an abstraction > layer with completely different backends for different firmware > branches/revisions. Otherwise you sign up for some pain.. Thanks for your explanation. I will definitely take a look on this if I am able to figure out. This patch seems work well with patch from Ben Greear on latest AP firmware. ---- Chun-Yeow From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail-ve0-x22e.google.com ([2607:f8b0:400c:c01::22e]) by merlin.infradead.org with esmtps (Exim 4.80.1 #2 (Red Hat Linux)) id 1WR0Re-0007Ve-Qt for ath10k@lists.infradead.org; Fri, 21 Mar 2014 14:29:15 +0000 Received: by mail-ve0-f174.google.com with SMTP id oz11so2571249veb.5 for ; Fri, 21 Mar 2014 07:28:52 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: References: <1395396014-24631-1-git-send-email-yeohchunyeow@gmail.com> Date: Fri, 21 Mar 2014 22:22:59 +0800 Message-ID: Subject: Re: [PATCH] ath10k: add the Rx rate in FW stats From: Yeoh Chun-Yeow List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "ath10k" Errors-To: ath10k-bounces+kvalo=adurom.com@lists.infradead.org To: Michal Kazior Cc: Kalle Valo , linux-wireless , "ath10k@lists.infradead.org" > Unfortunately this isn't as simple as it seems. > > Stats are received via wmi_stats_event. This structure contains fields > describing number of pdev/vdev/beer/bcn stat entries. All entries > (structures) are appended in certain order after this header. This > means if you change length of structure type you skew the calculation > of all succeeding entries. > > IOW this means this patch breaks peer stats for 999.999.0.636. > > I'm quite positive wmi_pdev_stats is also broken. This implies vdev > stats and peer stats position calculation is skewed in the first place > and you get garbage. > > The main problem here is there are subtle yet crazy binary > incompatibilities between these firmwares. > > The best way would probably be to implement wmi as an abstraction > layer with completely different backends for different firmware > branches/revisions. Otherwise you sign up for some pain.. Thanks for your explanation. I will definitely take a look on this if I am able to figure out. This patch seems work well with patch from Ben Greear on latest AP firmware. ---- Chun-Yeow _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k