Return-Path: Date: Tue, 24 Jan 2012 00:14:30 -0600 From: Jonathan Nieder To: linux-bluetooth@vger.kernel.org Cc: Andres Cimmarusti , linux-kernel@vger.kernel.org, info@linuxcertified.com, Paul Fertser , "Gustavo F. Padovan" , Marcel Holtmann Subject: Re: firmware loading failure for AR3011 (0cf3:3000 or 0cf3:3005 ?) Message-ID: <20120124061430.GB30210@burratino> References: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: List-ID: Hi, Andres Cimmarusti wrote[1]: > This is my device: > > Bus 001 Device 003: ID 0cf3:3000 Atheros Communications, Inc. AR3011 Bluetooth (no firmware) > > It does not turn on and kern.log shows following errors: > > [ 9.335184] ath3k_load_firmware: Can't change to loading configuration err > [ 9.335413] ath3k: probe of 1-1.3:1.0 failed with error -110 Further investigation revealed that the problem seems to be in userspace: various kernels including a 2.6.32.y-based and 3.1.8-based one reproducibly exhibit this behavior with a Debian wheezy userspace and do _not_ exhibit the problem in Debian squeeze, Ubuntu 11.10, or Fedora 16. More results: * The problem does not seem to be udev: downgrading udev in wheezy does not make the problem go away, and upgrading udev in squeeze does not provoke trouble. * The problem does not seem to be firmware-atheros: upgrading the firmware in squeeze does not provoke trouble. * The problem does not seem to be usb-modeswitch, for similar reasons. (You can see us grasping at straws here.) I guess the next natural component to blame is bluez. Any hints (e.g., debugging switches) for tracking this down? Thanks, Jonathan [1] http://bugs.debian.org/641749