Return-path: Received: from mail.ammma.de ([213.83.39.131]:3272 "EHLO ammma.de" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1755162AbZBPNjY (ORCPT ); Mon, 16 Feb 2009 08:39:24 -0500 Received: from ammma.net (hydra.ammma.mil [192.168.110.1]) by ammma.de (8.11.6/8.11.6/AMMMa AG) with ESMTP id n1GDf9I14307 for ; Mon, 16 Feb 2009 14:41:09 +0100 Received: from neo.wg.de (hydra.ammma.mil [192.168.110.1]) by ammma.net (8.12.11.20060308/8.12.11/AMMMa AG) with ESMTP id n1GDdKXI030039 for ; Mon, 16 Feb 2009 14:39:20 +0100 Received: from localhost (localhost [127.0.0.1]) by neo.wg.de (Postfix) with ESMTP id 8535C434FE4 for ; Mon, 16 Feb 2009 14:39:20 +0100 (CET) Received: from neo.wg.de ([127.0.0.1]) by localhost (neo.wg.de [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tzl+UvGlS1aH for ; Mon, 16 Feb 2009 14:39:15 +0100 (CET) Received: from localhost (localhost [127.0.0.1]) by neo.wg.de (Postfix) with ESMTP id 95CCF434FE6 for ; Mon, 16 Feb 2009 14:39:15 +0100 (CET) Message-ID: <20090216143915.18703m2wpmqic1ic@neo.wg.de> (sfid-20090216_143928_045937_879E5230) Date: Mon, 16 Feb 2009 14:39:15 +0100 From: Jan Schneider To: "linux-wireless@vger.kernel.org" Subject: Connection "interruptions" MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; DelSp="Yes"; format="flowed" Sender: linux-wireless-owner@vger.kernel.org List-ID: Hi, now that I got current wireless-compat snapshots running at all, I'm back to look into what I hoped had been fixed in newer versions: I see arbitrary "interruptions" of the wireless connection. The transfer just halts for a few seconds, no traffic is going over the ether at all. Then suddenly, the connection resumes as if nothing happened. Nothing is being logged by kernel, wpa_supplicant or syslog. This is a draft-n-only thing, it does not happen if I tell either the AP or the driver module to not use draft-n. This is an Intel 5300 board. I am wondering how to further track this down, or how to collect useful data for a bug report. Logging everything with the debug flag is probably not very helpful, since this produces huge logs, and it's not easy to determine the exact point when the connection drops. What might be useful flags for the debug50 parameter to get more information from the driver without getting lost in the sheer amount of debug messages? Jan. -- Do you need professional PHP or Horde consulting? http://horde.org/consulting/