From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755362AbYIKMx5 (ORCPT ); Thu, 11 Sep 2008 08:53:57 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753016AbYIKMxs (ORCPT ); Thu, 11 Sep 2008 08:53:48 -0400 Received: from shadow.wildlava.net ([67.40.138.81]:46181 "EHLO shadow.wildlava.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752951AbYIKMxs (ORCPT ); Thu, 11 Sep 2008 08:53:48 -0400 Message-ID: <48C914D8.6080806@skyrush.com> Date: Thu, 11 Sep 2008 06:53:44 -0600 From: Joe Peterson User-Agent: Thunderbird 2.0.0.16 (X11/20080727) MIME-Version: 1.0 To: Andrew Morton CC: alan@lxorguk.ukuu.org.uk, linux-kernel@vger.kernel.org Subject: Re: [PATCH] TTY: Fix loss of echoed characters (2nd follow-on PATCH attached) References: <200807252257.m6PMvieO003213@imap1.linux-foundation.org> <48AC3A16.4080209@skyrush.com> <48B3F9F7.2050503@skyrush.com> <48C54EC2.4060901@skyrush.com> <20080908173250.2452c5b8.akpm@linux-foundation.org> <20080909115536.1664d57a@lxorguk.ukuu.org.uk> <20080909104325.4653190b.akpm@linux-foundation.org> <48C6DFA4.5040406@skyrush.com> <20080910163902.2b7ac22b.akpm@linux-foundation.org> In-Reply-To: <20080910163902.2b7ac22b.akpm@linux-foundation.org> X-Enigmail-Version: 0.95.6 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Andrew Morton wrote: > On Tue, 09 Sep 2008 14:42:12 -0600 > Joe Peterson wrote: > >> 1) Fix tab erasure handling >> 2) Improve locking when working with the echo buffer >> 3) Remove the big kernel lock (BKL) from n_tty >> >> Tab erasure handling is now more robust and able to handle non-zero >> canon column cases more correctly. This is done by making correct use >> of what is known in the eraser function (read buffer contents) and what >> is known at the time of processing the tab erasure (column state). >> >> Also, better locking of the echo buffer will now prevent any attempts to >> process partial multi-byte echo operations. And since the echo buffer >> code now isolates the tty column state code to the process_out* and >> process_echoes functions, we can remove the big kernel lock (BKL) >> and replace it with more modern mutex locks. > > Boy. Has this been carefully tested with lockdep enabled? Yes, I have the following in my kernel config: CONFIG_KALLSYMS_ALL=y CONFIG_RT_MUTEX_TESTER=y CONFIG_DEBUG_SPINLOCK=y CONFIG_DEBUG_LOCK_ALLOC=y CONFIG_PROVE_LOCKING=y CONFIG_LOCKDEP=y CONFIG_LOCK_STAT=y CONFIG_DEBUG_LOCKDEP=y CONFIG_TRACE_IRQFLAGS=y CONFIG_STACKTRACE=y I have run the various tests that I used to find/investigate the echo loss problem to begin with (i.e. that make output compete with echo). Also some manual tests that fill the echo buffer during a stopped tty, test the eraser interaction, etc. Things look fine so far. -Joe