Message-ID: <1334203438.3788.11.camel@jlt3.sipsolutions.net> (sfid-20120412_060443_349942_502ECEE8)
Subject: Re: [PATCH] mac80211: fix an issue in ieee80211_tx_info count field
 management
From: Johannes Berg <johannes@sipsolutions.net>
To: Lorenzo Bianconi <lorenzo.bianconi83@gmail.com>
Cc: John Linville <linville@tuxdriver.com>,
	linux-wireless@vger.kernel.org, Felix Fietkau <nbd@openwrt.org>
Date: Thu, 12 Apr 2012 06:03:58 +0200
In-Reply-To: <CAA2SeNJNan7uwJ_VELCb-rDX4ddq+=K0Ud3n5mrv6MJ+O-Bb9w@mail.gmail.com> (sfid-20120406_204847_916462_26A238BC)
References: <CAA2SeNJNan7uwJ_VELCb-rDX4ddq+=K0Ud3n5mrv6MJ+O-Bb9w@mail.gmail.com>
	 (sfid-20120406_204847_916462_26A238BC)
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Sender: linux-wireless-owner@vger.kernel.org

On Fri, 2012-04-06 at 20:48 +0200, Lorenzo Bianconi wrote:
> Hi,
> 
> I noticed a possible issue in the status count field management of the
> ieee80211_tx_info data structure. In particular, when the AGGR
> processing is employed,
> status.rates[].count is set just for the first frame and not for
> others belonging to the same burst, leading to wrong statistic data in
> the mac80211 debug file system.
> 
> Regards
> 
> Lorenzo
> 
> Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi83@gmail.com>
> ---
> --- a/net/mac80211/status.c
> +++ b/net/mac80211/status.c
> @@ -355,7 +355,13 @@
>  	int rtap_len;
> 
>  	for (i = 0; i < IEEE80211_TX_MAX_RATES; i++) {
> -		if (info->status.rates[i].idx < 0) {
> +		if ((info->flags & IEEE80211_TX_CTL_AMPDU) &&
> +		    !(info->flags & IEEE80211_TX_STAT_AMPDU)) {
> +			/* just the first aggr frame carry status info */
> +			info->status.rates[i].idx = -1;
> +			info->status.rates[i].count = 0;
> +			break;
> +		} else if (info->status.rates[i].idx < 0) {

Is that really the way all drivers do it? I know ath9k seems to do it
that way, but I don't think all drivers do. ath9k is probably the
exception rather than the rule, at least today

johannes