From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758062AbaDWVQX (ORCPT ); Wed, 23 Apr 2014 17:16:23 -0400 Received: from lxorguk.ukuu.org.uk ([81.2.110.251]:56056 "EHLO lxorguk.ukuu.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754792AbaDWVQV (ORCPT ); Wed, 23 Apr 2014 17:16:21 -0400 Date: Wed, 23 Apr 2014 22:16:09 +0100 From: One Thousand Gnomes To: Andrew Morton Cc: Jan Kara , LKML , pmladek@suse.cz, Frederic Weisbecker , Steven Rostedt Subject: Re: [PATCH 8/8] printk: Add config option for disabling printk offloading Message-ID: <20140423221609.0d3b536f@alan.etchedpixels.co.uk> In-Reply-To: <20140423111750.88ad799bb271a1fe0bed119f@linux-foundation.org> References: <1395770101-24534-1-git-send-email-jack@suse.cz> <1395770101-24534-9-git-send-email-jack@suse.cz> <20140326172332.5f1e1bfb@alan.etchedpixels.co.uk> <20140326192815.GC18118@quack.suse.cz> <20140418115438.1e65e07af17e3ba6d7c554db@linux-foundation.org> <20140422112259.5f106a44@alan.etchedpixels.co.uk> <20140423110847.GB17824@quack.suse.cz> <20140423111750.88ad799bb271a1fe0bed119f@linux-foundation.org> Organization: Intel Corporation X-Mailer: Claws Mail 3.8.1 (GTK+ 2.24.20; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > The problem of course will be in sizing the driver's queue. Perhaps we > can have a driver which uses a huge queue, temporarily use that driver > during boot then switch over to a conventional console driver? It depends how early in boot you are but this is true equally of kernel threads and other stuff. The big problem doing it early is that we have no mechanism to "open" a tty port without having a tty. We do now have the split for this for queueing input and to some extent output, but not all of the "open" aspect. The output buffering is also quite limited for most devices but that bit *is* easy to fix if needed. Given this all only affects huge boxes and the only fix we have right now is hideously ugly I'd rather it stayed in the vendor hacks to their trees, until someone is inspired to sort it properly. On the plus side having a queueing console that can use tty methods would mean you could make any tty a console and special hacks for support would be improving reliability/early functionality rather than enabling. Essentially you can create tty_port_kopen()/tty_port_kclose() and use the tty device write methods with a tty_port, it's just a matter of continuing to clean up the tty/tty_port split properly. It's not a small job though. Alan