From mboxrd@z Thu Jan 1 00:00:00 1970 From: Theodore Tso Subject: Re: is_real_interrupt in 8250.c Date: Wed, 29 Nov 2006 13:11:49 -0500 Message-ID: <20061129181148.GA31874@thunk.org> References: <200611291744.kATHixEN006240@oogie-boogie.ics.uci.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from thunk.org ([69.25.196.29]:8138 "EHLO thunker.thunk.org") by vger.kernel.org with ESMTP id S967497AbWK2SLv (ORCPT ); Wed, 29 Nov 2006 13:11:51 -0500 Content-Disposition: inline In-Reply-To: <200611291744.kATHixEN006240@oogie-boogie.ics.uci.edu> Sender: linux-serial-owner@vger.kernel.org List-Id: linux-serial@vger.kernel.org To: Dan Nicolaescu Cc: linux-serial@vger.kernel.org On Wed, Nov 29, 2006 at 09:44:59AM -0800, Dan Nicolaescu wrote: > > Hi, > > In 8250.c is_real_interrupt is defined like this: > > /* > * We default to IRQ0 for the "no irq" hack. Some > * machine types want others as well - they're free > * to redefine this in their header file. > */ > #define is_real_interrupt(irq) ((irq) != 0) > > on my platform the UART IRQ is 0, so is_real_interrupt returns > false. > > In order to allow machines to override is_real_interrupt, and for the > code to match the comment shouldn't this patch be applied? No, IRQ 0 for linux is always "no IRQ". We should probably add this to the LKML FAQ, since it keeps coming up. Per Linus, if there are any architectures that has a valid IRQ #0, then the architecture-specific kernel code should remap IRQ numbers. See: http://comments.gmane.org/gmane.linux.kernel/468449 - Ted