From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from static-ip-62-75-166-246.inaddr.intergenia.de ([62.75.166.246]:48424 "EHLO vs166246.vserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752592AbXFTVBp (ORCPT ); Wed, 20 Jun 2007 17:01:45 -0400 From: Michael Buesch To: Johannes Berg Subject: Re: [PATCH] bcm43xx: honour promisc bit even in monitor mode Date: Wed, 20 Jun 2007 23:01:25 +0200 Cc: linux-wireless References: <1182295838.3714.14.camel@johannes.berg> In-Reply-To: <1182295838.3714.14.camel@johannes.berg> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Message-Id: <200706202301.25890.mb@bu3sch.de> Sender: linux-wireless-owner@vger.kernel.org List-ID: On Wednesday 20 June 2007 01:30:38 Johannes Berg wrote: > When bcm43xx goes into monitor mode, it always is a promisc monitor > mode. For the hostapd using monitor interfaces, we'll want to only go > into monitor mode if it's really required. I previously thought that > this requires changes to mac80211, but mac80211 does in fact already > seem to handle IFF_PROMISC appropriately. Hence, this change is the > only required to get proper semantics for a monitor interface that > isn't in promiscuous mode on bcm43xx. > > Signed-off-by: Johannes Berg > > --- > Since wireshark/tcpdump by default still set IFF_PROMISC everything will > work just fine, but if somebody really wants a monitor interface that > isn't promisc this makes it work too. However, you'll of course get all > packets still if any other interface is set promisc. We don't do any > promisc/non-promisc filtering in the stack when multiple monitor > interfaces are open. Applied, thanks. -- Greetings Michael.