Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261929AbTILXoF (ORCPT ); Fri, 12 Sep 2003 19:44:05 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261957AbTILXoF (ORCPT ); Fri, 12 Sep 2003 19:44:05 -0400 Received: from dsl092-233-042.phl1.dsl.speakeasy.net ([66.92.233.42]:56983 "EHLO whisper.qrpff.net") by vger.kernel.org with ESMTP id S261929AbTILXoC (ORCPT ); Fri, 12 Sep 2003 19:44:02 -0400 X-All-Your-Base: Are Belong To Us!!! X-Envelope-Recipient: linux-kernel@vger.kernel.org X-Envelope-Sender: oliver@klozoff.com Message-ID: <3F625996.2070306@klozoff.com> Date: Fri, 12 Sep 2003 19:41:10 -0400 From: Stevie-O User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 X-Accept-Language: en-us, en MIME-Version: 1.0 To: linux-kernel@vger.kernel.org Subject: Serial port problem with SiS 730 chispet? X-Enigmail-Version: 0.76.1.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1959 Lines: 43 I apologize if this problem has been reported before; I tried googling but came up empty. I just upgraded my system and put in a EGS K7SEM motherboard with a SiS 730 chipset. I have Debian installed. For those not in the know, Debian provides a utility called 'discover' that does simple things to detect hardware, such as scanning the pci bus to find modules that need to be loaded. One of the things it can do is probe serial ports to find modems and mice and other such things. Well, to make a long story short, after some time stracing and gdbing and other experimenting, I think I have determined that discover is stalling when it tries to open /dev/ttyS0. discover is called twice Debian's startup scripts, and both times it blocks waiting to open the port. I modified the single-user-mode startup to run 'openvt sulogin' right away. Things got stranger as I proceeded: (1) strace -p wakes it up at 'open(.../dev/ttyS0)'. I think the signal being sent when strace does ptrace(PTRACE_ATTACH) kickstarts it into working. (2) stty -F/dev/ttyS0 (while discover is still blocked) also blocks. (3) However, when I hit ctrl-C to kill stty, discover unblocks. (4) To top it all off, everything begins to work at some point. After the machine has completely booted and I can ssh in (the machine is *intended* to be a headless server, so normally the only access is via ssh), I can happily run discover as often as I want and it will proceed normally. Can anyone out there please help me figure out exactly what's happening (and more importantly, how to fix it)? -- Please CC me, I'm not subscribed to the list. -- - Stevie-O Real Programmers use COPY CON PROGRAM.EXE - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/