From mboxrd@z Thu Jan 1 00:00:00 1970 From: Corey Minyard Subject: Re: [PATCH] serial 8250: move push calls out of lock Date: Fri, 27 Apr 2007 08:09:23 -0500 Message-ID: <4631F603.40901@acm.org> References: <20070426205103.GA28569@localdomain> <20070426205847.GB11735@flint.arm.linux.org.uk> <46311E48.4070106@acm.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mta13.mail.adelphia.net ([68.168.78.44]:37405 "EHLO mta13.adelphia.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755806AbXD0NJ0 (ORCPT ); Fri, 27 Apr 2007 09:09:26 -0400 In-Reply-To: <46311E48.4070106@acm.org> Sender: linux-serial-owner@vger.kernel.org List-Id: linux-serial@vger.kernel.org To: Corey Minyard Cc: Linux Kernel , Alan Cox , linux-serial@vger.kernel.org, Andrew Morton Corey Minyard wrote: > Russell King wrote: >> >> Use of an uninitialised variable. The reason we use check_modem_status >> to return the current status here is because reading the MSR _clears_ >> interrupts. So, repeatedly calling get_mctrl via an ioctl is a great >> way to introduce a hardware race condition if you just read the MSR and >> never act on it. >> > "status" is used as a return variable; it is set by the call. I modified > check_modem_status() to return if it detected a change and set the > status if it was passed in. Ok, I see what I missed. Never mind. -corey