From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755353AbaGDK1N (ORCPT ); Fri, 4 Jul 2014 06:27:13 -0400 Received: from mout.kundenserver.de ([212.227.17.13]:59241 "EHLO mout.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751958AbaGDK1L (ORCPT ); Fri, 4 Jul 2014 06:27:11 -0400 From: Arnd Bergmann To: Andrey Ryabinin Cc: linux-arm-kernel@lists.infradead.org, Russell King , Nicolas Pitre , Catalin Marinas , Will Deacon , linux-kernel@vger.kernel.org Subject: Re: [PATCH] arm: get rid of hardcoded assumptions about kernel stack size Date: Fri, 04 Jul 2014 12:27:04 +0200 Message-ID: <4429781.iuSvvPCxYj@wuerfel> User-Agent: KMail/4.11.5 (Linux/3.11.0-18-generic; KDE/4.11.5; x86_64; ; ) In-Reply-To: <53B6541B.6000702@samsung.com> References: <1403099422-850-1-git-send-email-a.ryabinin@samsung.com> <201407032224.28386.arnd@arndb.de> <53B6541B.6000702@samsung.com> MIME-Version: 1.0 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="us-ascii" X-Provags-ID: V02:K0:umN1ztHvhSOG/w7LYpvkxXO8gLuCk9sKMoRjdm0bd2F xFts66eXjHO7HUjMhMix7oFNk7ZJNDqRylX9fSeJpnKfbCY40Q idvTOv8i14fcv88Qn5rGSObTJReu9ofR00mnFm/9CHfS+NKtWW bTfGA8Vo6LqSh+4NR+SRfL0ke2wEoqZTTaj2yhhR1Gr73WrG8x Iju4OJZqzlgoEAV4UhLaM/SIqVbwYD56QuQ8/kFSgAG3IL0gTN ibZX6xtViCQtTwQRAIV5GLEytvZrb2UEMsS0msf95e9mEo76lH s690DSJIHqlmdzyDtNmfRKHi+W+2PgtJKKn7a7XxuN2F6JL/GK zZI8xxPtdjhrjHOOBWbk= Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Friday 04 July 2014 11:13:31 Andrey Ryabinin wrote: > > > > but I wonder if there is a way to avoid the extra include here, as it might also > > cause a general slowdown because of asm/memory.h getting pulled into more .c > > files. Would it be reasonable to hardcode PAGE_SIZE here? > > > > IMO it's a bug of iop13xx platform, that it includes "higlevel" linux/reboot.h > from a very "lowlevel" header mach/iop13xx.h. I think it should be fixed with a patch above. > Slowing down of kernel build for a few more seconds is not good enough reason for me to > hardcode PAGE_SIZE here. I don't think we can pinpoint a specific header that is "wrong" here, the fundamental problem is that our header files are a bit messy when it comes to recursive inclusion and we'd be better off if we generally were a little more careful about including headers from other headers. It's also very hard to retroactively clean this up on a large scale. Arnd