From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754163AbaEHOem (ORCPT ); Thu, 8 May 2014 10:34:42 -0400 Received: from cam-admin0.cambridge.arm.com ([217.140.96.50]:39304 "EHLO cam-admin0.cambridge.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753205AbaEHOel (ORCPT ); Thu, 8 May 2014 10:34:41 -0400 Date: Thu, 8 May 2014 15:34:07 +0100 From: Will Deacon To: One Thousand Gnomes Cc: Andrew Morton , Jan Kara , "mm-commits@vger.kernel.org" , "peterz@infradead.org" , "kay@vrfy.org" , LKML Subject: Re: + printk-print-initial-logbuf-contents-before-re-enabling-interrupts.patch added to -mm tree Message-ID: <20140508143407.GB8981@arm.com> References: <53640c8c.5++0zeO0pmfqKMwm%akpm@linux-foundation.org> <20140502224651.GG23636@quack.suse.cz> <20140506120648.GA30234@arm.com> <20140506122958.GG9291@quack.suse.cz> <20140506131234.GD30234@arm.com> <20140506150553.a12959fd71d83eddcb4b325d@linux-foundation.org> <20140507095850.GC18456@arm.com> <20140507174151.3c28f15a@alan.etchedpixels.co.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20140507174151.3c28f15a@alan.etchedpixels.co.uk> User-Agent: Mutt/1.5.23 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Alan, On Wed, May 07, 2014 at 05:41:51PM +0100, One Thousand Gnomes wrote: > > Possibly, but I fear we'd incur the wrath of Alan after reading that other > > thread. Having a CONFIG_ option or similar to control the amount of printing > > we do is very similar to the command-line option Jan proposed in his series. > > I've nothing against a configuration option, and having a printk that > queued most stuff to the serial IRQ handler on overflow and a boot option > of printk=synchronous for debug work would be awesome in my book. Many end > production boxes don't give a toss about losing the odd bit of data they > just need to shift the logs somewhere for filing. Ok, I'll revisit this with a command-line option for the debug case. Thanks. > All these "clever" approaches just seem to me to be ever more convoluted > attempts to fail to deal with the simple reality that if you put more > down the sewage pipe than fits it has to overflow somewhere. > > Our tty drivers have a fifo, our tty drivers have an IRQ driven write > operation. It seems silly to be adding magic to solve the problem rather > than just using it (and in turn getting a ton of other consoles for free, > and being able to take GregKH's current usb console hack out) Hmm, I'm not at all familiar with the tty layer, so I'd have to go and take a look. Jan, did you look at this at all? Will