Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757122Ab2BWV1l (ORCPT ); Thu, 23 Feb 2012 16:27:41 -0500 Received: from mail-qy0-f174.google.com ([209.85.216.174]:49444 "EHLO mail-qy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754369Ab2BWV1k convert rfc822-to-8bit (ORCPT ); Thu, 23 Feb 2012 16:27:40 -0500 Authentication-Results: mr.google.com; spf=pass (google.com: domain of bobbypowers@gmail.com designates 10.229.137.18 as permitted sender) smtp.mail=bobbypowers@gmail.com; dkim=pass header.i=bobbypowers@gmail.com MIME-Version: 1.0 In-Reply-To: <20120223195026.GA30895@merkur.ravnborg.org> References: <1330023731-14043-1-git-send-email-bobbypowers@gmail.com> <20120223195026.GA30895@merkur.ravnborg.org> Date: Thu, 23 Feb 2012 16:27:39 -0500 Message-ID: Subject: Re: [PATCH] linux headers: include linux/types.h where appropriate From: Bobby Powers To: Sam Ravnborg Cc: Dave Airlie , linux-kernel@vger.kernel.org, tglx@linutronix.de, davem@davemloft.net Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1856 Lines: 42 On Thu, Feb 23, 2012 at 2:50 PM, Sam Ravnborg wrote: > On Thu, Feb 23, 2012 at 07:28:10PM +0000, Dave Airlie wrote: >> >> > This silences the related header check warnings. >> >> please fix the bug in the correct place, the drm.h header includes >> linux/types.h and this is one on purpose. >> >> maybe expend some effort fixing the stupid tool to recurse like a normal >> preprocessor would. > > Before we fix the tool maybe we should agree on what the tool should check for. > One obvious thing to check is that we do not start to use user-land > types in our exported headers like uint32_t etc. > Another could be that no header file should depend oon the user to include other > header files. > Etc. I just sent out a patch that recursively checks for including types.h, as it was quite simple. Additional checks seem like they may be a good idea, but not something that has to be decided right here. > The checks we have implmented today are very conservative - but as you resist > to add a few includes in the drm/drm*.h files prevents us from hitting > zero warnings even with these conservative checks. > > If we started to be more strict we would see much more warnings/errors. > But all this fuzz about a handfull of includes in the drm header files > sure scare me away from ever starting this. > I can only image the uproar some other changes may cause then. Interestingly, with headers_check searching for types.h recursively (clearing up some noise), include/drm/drm_mode.h stands out as still needing to include , as it doesn't import drm.h. > > ? ? ? ?Sam -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/