Return-path: Received: from w1.fi ([128.177.27.249]:58315 "EHLO jmalinen.user.openhosting.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751160Ab1FLRSu (ORCPT ); Sun, 12 Jun 2011 13:18:50 -0400 Date: Sun, 12 Jun 2011 10:18:43 -0700 From: Jouni Malinen To: hedwin Cc: linux-wireless Subject: Re: Support for Android for mac80211 / cfg80211 802.11 drivers Message-ID: <20110612171843.GA20860@jm.kir.nu> (sfid-20110612_191855_933304_FA66E283) References: <1296633772.3624.2.camel@jlt3.sipsolutions.net> <20110202113052.GA14126@jm.kir.nu> <1296646434.5671.1.camel@jlt3.sipsolutions.net> <20110202133839.GA14904@jm.kir.nu> <20110610200900.GE2665@tuxdriver.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: Sender: linux-wireless-owner@vger.kernel.org List-ID: On Fri, Jun 10, 2011 at 10:25:17PM +0200, hedwin wrote: > If I want to use wpa_supplicant it seems that I need to use libnl. If > not how does wpa_supplicant connect to mac80211? wpa_supplicant uses libnl for accessing nl80211. If libnl is not acceptable for whatever reason, I see two options: 1) re-implement libnl replacement that provides identical API or 2) modify wpa_supplicant src/drivers/driver_nl80211.c to process nl80211/netlink messages with internal implementation instead of libnl. I don't care particularly how that would be done, but trying to use WEXT is not really a reasonable approach here. > Well if it is google's problem to solve it, this discussion about > wireless support in android is on hold until they've done so. right? I have no reason to believe that Android support could not be discussed even if some Android distributions do not like libnl. People seem to be perfectly fine in using libnl on Android. Sure, some cases may not be able to include libnl, but that should not stop this discussion or implementation work and there are clear options described above on how to address the license concern. -- Jouni Malinen PGP id EFC895FA