From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benjamin Herrenschmidt Subject: Re: Console & FBDev vs. locking Date: 28 May 2003 15:02:58 +0200 Sender: linux-kernel-owner@vger.kernel.org Message-ID: <1054126978.541.3.camel@gaston> References: <1054122360.602.197.camel@gaston> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1054122360.602.197.camel@gaston> List-Id: Cc: Linux Fbdev development list , James Simmons On Wed, 2003-05-28 at 13:46, Benjamin Herrenschmidt wrote: > All printk originated console call should done with the console > semaphore held. The console-sem is the de-facto locking mecanism for the > console today, while not fine-grained, it's probably plenty enough for > what we need in 2.5 and unless previous implementations which ran with > irqs off, the console drivers can actually block and rely on HW > interrupts. Hrm... Of course, this is only true for things like resize/mode change, etc... (and soon blanking once I'm done with it). The way the console sem is used is so that normal output originating from printk can still happen at interrupt time. BTW. I know some people would prefer not doing so, but I'd like to eventually disable that 'feature', or at least add a flag to the the consw driver telling if it supports beeing called at interrupt/spinlock time or not, and if not, defer operations to process context. .. Ben.