Return-path: Received: from mail-yw0-f176.google.com ([209.85.211.176]:51775 "EHLO mail-yw0-f176.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753453Ab0AZCgy (ORCPT ); Mon, 25 Jan 2010 21:36:54 -0500 Received: by ywh6 with SMTP id 6so3711588ywh.4 for ; Mon, 25 Jan 2010 18:36:53 -0800 (PST) Message-ID: <4B5E5546.20105@twilightblue.net> Date: Mon, 25 Jan 2010 20:36:54 -0600 From: Don Darling MIME-Version: 1.0 To: Bob Copeland CC: =?ISO-8859-1?Q?G=E1bor_Stefanik?= , Larry Finger , linux-wireless@vger.kernel.org Subject: Re: Bug: 2.6.32 ath5k, associating with WEP128 AP is unreliable References: <4B5CD1FF.4090209@twilightblue.net> <4B5DE859.6080409@lwfinger.net> <69e28c911001251121i61be9c0bhed43648a12db42f6@mail.gmail.com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Sender: linux-wireless-owner@vger.kernel.org List-ID: Hello folks, Thanks for the suggestions -- here are a few questions I can answer from earlier today: 1) I didn't realize there was a subtle difference between WEP104 and WEP128. What I can tell you is that the WEP AP is running Tomato 1.23, and is set to WEP 128-bit. However, the key I'm using is only 104 bits. This probably means I'm actually using WEP104. 2) I'm definitely not using madwifi in either case -- I've always used the ath5k driver on this machine. 3) As far as I can tell, there are no modparams being used with ath5k. I checked both /etc/modprobe.d and /etc/sysctl.conf (I'm not sure if the latter affects modules or not, actually - I think it might depend on the module). 4) I realize WPA+TKIP is more secure and I would use it if I could, but I have some old devices that don't support WPA it so I'm limited to WEP. 5) I use -Dwext with wpa_supplicant when I *do* use WPA+TKIP networks. Also, FWIW: # lsmod | grep ath ath5k 127364 0 mac80211 155788 1 ath5k ath 7708 1 ath5k cfg80211 90364 3 ath5k,mac80211,ath led_class 4000 1 ath5k I'm not sure when I might have a chance to bisect, but will certainly let you know if I do. Best regards, Don Bob Copeland wrote: > 2010/1/25 G?bor Stefanik : > >> On Mon, Jan 25, 2010 at 7:52 PM, Larry Finger wrote: >> >>> For the record, WEP104 and WEP128 are the same thing. The smaller number >>> makes it explicit that the 24-bit IV is transmitted in the open. The same >>> thing applies to WEP40 and WEP64. >>> >>> >> Only one problem there is another, nonstandard WEP type where the >> actual key is 128-bit, and this is also sometimes referred to as >> WEP128 (with the same metric as WEP64, this one would be WEP152). Same >> goes for "WEP228/WEP256/WEP280". >> > > Yea, I think ath hw supports that too: > > #define AR5K_KEYTABLE_TYPE_40 0x00000000 > #define AR5K_KEYTABLE_TYPE_104 0x00000001 > #define AR5K_KEYTABLE_TYPE_128 0x00000003 > > at least that's what I assumed it meant. But there's no way to set it > from mac80211 (nor should there be). > >