Re: [71attendees] ietf71 wireless problems

Jeffrey Hutzelman <> Tue, 11 March 2008 14:54 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id D2AAD3A6E1C; Tue, 11 Mar 2008 07:54:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -101.158
X-Spam-Status: No, score=-101.158 tagged_above=-999 required=5 tests=[AWL=-0.721, BAYES_00=-2.599, FH_RELAY_NODNS=1.451, HELO_MISMATCH_ORG=0.611, RDNS_NONE=0.1, USER_IN_WHITELIST=-100]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id hc4ob5o8daEl; Tue, 11 Mar 2008 07:54:00 -0700 (PDT)
Received: from (localhost []) by (Postfix) with ESMTP id C3EC628C493; Tue, 11 Mar 2008 07:52:57 -0700 (PDT)
Received: from localhost (localhost []) by (Postfix) with ESMTP id BDA6328C48D; Tue, 11 Mar 2008 07:52:56 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id x6fVYp1BgMJb; Tue, 11 Mar 2008 07:52:55 -0700 (PDT)
Received: from (JACKFRUIT.SRV.CS.CMU.EDU []) by (Postfix) with ESMTP id BC34A28C4F3; Tue, 11 Mar 2008 07:50:07 -0700 (PDT)
Received: from ( []) (authenticated bits=0) by (8.13.6/8.13.6) with ESMTP id m2BEldM3008531 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 11 Mar 2008 10:47:44 -0400 (EDT)
Date: Tue, 11 Mar 2008 10:47:38 -0400
From: Jeffrey Hutzelman <>
To: Michael Richardson <>,
Message-ID: <>
In-Reply-To: <>
References: <>
X-Mailer: Mulberry/4.0.8 (Linux/x86)
MIME-Version: 1.0
Content-Disposition: inline
Subject: Re: [71attendees] ietf71 wireless problems
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Discussion list for IETF Meeting 71 attendees <>
List-Unsubscribe: <>, <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit

--On Monday, March 10, 2008 10:56:00 AM -0400 Michael Richardson 
<> wrote:

> I've never had problems with wireless, and I'm pretty kernel competent,
> and the same cards were last used 2008-03-06.   I'm running debian
> "GNU/Linux" etch with a custom kernel.  I suspect that my cards
> might need firmware upgrades to deal with some new-fangled mechanism, but
> I am ignorant of what that might be.

It is worth noting that the Linux wireless extensions provide for at most 
64K-1 bytes of data in scan results (the length field is a 16-bit integer). 
In the breakfast room around the time you were having this problem, the 
scan data for the number of AP's visible was larger than 32K.  So if you 
have an application which starts with a 4K buffer and then doubles it, as 
NetworkManager does, then it ends up discovering 32K is not big enough, and 
doubling it to.... zero!  A patched NM which instead jumps from 32K to 
64K-1 (the maximum size) is able to collect scan results and connect to the 

Note that this is in addition to other issues which have been discussed on 
this thread, some of which have been resolved since Sunday.

-- Jeff
71attendees mailing list