From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263235AbTDRUr0 (ORCPT ); Fri, 18 Apr 2003 16:47:26 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263239AbTDRUr0 (ORCPT ); Fri, 18 Apr 2003 16:47:26 -0400 Received: from e35.co.us.ibm.com ([32.97.110.133]:62646 "EHLO e35.co.us.ibm.com") by vger.kernel.org with ESMTP id S263235AbTDRUrZ (ORCPT ); Fri, 18 Apr 2003 16:47:25 -0400 Date: Fri, 18 Apr 2003 13:49:05 -0700 From: "Martin J. Bligh" Reply-To: linux-kernel To: linux-kernel Subject: [Bug 602] New: warnings on hcd rmmod: "dangling refs(N) to bus B" (fwd) Message-ID: <1382940000.1050698945@flay> X-Mailer: Mulberry/2.1.2 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org http://bugme.osdl.org/show_bug.cgi?id=602 Summary: warnings on hcd rmmod: "dangling refs(N) to bus B" Kernel Version: 2.5.66 Status: NEW Severity: low Owner: greg@kroah.com Submitter: dbrownell@users.sourceforge.net The bus refcounting mechanism is broken, it often produces warning messages like that one when modules are removed. The problem appears to be an extremely long-standing one, at least in terms of the how-to-reproduce I saw: - connect a device that won't immediately enumerate, so that setting its address needs to be retried (or similar error, like set_configuration failing) - unplug that device ... at this point the refcount is wrong, but you can't tell until - rmmod the relevant HCD. This particular cause of that message is because the bus refcount isn't released on error. The fix isn't as straightforward as it should be because of funky calling conventions ("longstanding", likely since 2.4), but it's not complicated either.