From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [Bugme-new] [Bug 8107] New: dev->header_cache_update has a random value Date: Thu, 01 Mar 2007 17:38:43 -0800 (PST) Message-ID: <20070301.173843.104646614.davem@davemloft.net> References: <20070301143727.4ba00d04@freekitty> <20070301145423.6ffc52e5.akpm@linux-foundation.org> <20070301153025.6559a53e@freekitty> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: akpm@linux-foundation.org, netdev@vger.kernel.org, bugme-daemon@bugzilla.kernel.org, loveminix@yahoo.com.cn, khc@pm.waw.pl To: shemminger@linux-foundation.org Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:49566 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1422706AbXCBBip (ORCPT ); Thu, 1 Mar 2007 20:38:45 -0500 In-Reply-To: <20070301153025.6559a53e@freekitty> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org From: Stephen Hemminger Date: Thu, 1 Mar 2007 15:30:25 -0800 > Your right, but so far there is no ioctl to take it out of this mode. > So it is a one way door. You can keep attaching a different new protocol to an HDLC device until it is brought up. > This device never calls hdlc->detach?? The HDLC layer calls the protocol ->detach, but the Cisco HDLC code does not have a detach method. I'm starting to feel that this is the bug, because as it stands the Cisco HDLC code leaves stale pointers around and then if you unload the Cisco HDLC module things can explode.