From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752058Ab0HUM65 (ORCPT ); Sat, 21 Aug 2010 08:58:57 -0400 Received: from pfepa.post.tele.dk ([195.41.46.235]:41053 "EHLO pfepa.post.tele.dk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751459Ab0HUM6z (ORCPT ); Sat, 21 Aug 2010 08:58:55 -0400 Date: Sat, 21 Aug 2010 14:58:53 +0200 From: Sam Ravnborg To: Wu Fengguang Cc: Andrew Morton , Tiago Maluta , Michal Marek , LKML Subject: Re: [PATCH][RESEND] fix compilation error with CROSS_COMPILE Message-ID: <20100821125853.GA25430@merkur.ravnborg.org> References: <20100821062133.GA809@localhost> <20100821072454.GA25054@merkur.ravnborg.org> <20100821072955.GA4950@localhost> <20100821074021.GA25099@merkur.ravnborg.org> <20100821092652.GA9967@localhost> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20100821092652.GA9967@localhost> User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > > However I confirmed that the linux-2.6 tree is OK, the above patch that caused > the problem is still sitting in -mm tree: > > commit 7f48a0a8c15c41f382f2265088bc96007836e608 > Author: Tiago Maluta > Date: Mon Jun 21 22:00:17 2010 +0200 > > Linux has an recent option to define CROSS_COMPILE in menuconfig. > > Typically, one of the options when building for another target is > ('arm' is just an example): > > # make ARCH=arm CROSS_COMPILE=arm-cc O=../build > > If I omit CROSS_COMPILE and use the new way in .config: > CONFIG_CROSS_COMPILE="arm-cc" > > I got: > > cc1: error: unrecognized command line option "-mlittle-endian" > cc1: error: unrecognized command line option "-mabi=aapcs-linux" > cc1: error: unrecognized command line option "-mno-thumb-interwork" > /usr/src/linux/kernel/bounds.c:1: error: bad value (armv5t) for -march= switch > /usr/src/linux/kernel/bounds.c:1: error: bad value (armv5t) for -mtune= switch > make[2]: *** [kernel/bounds.s] Error 1 > make[1]: *** [prepare0] Error 2 > make: *** [sub-make] Error 2 > > Verbose output clearly says that I'm using the host compiler instead cross > compiler. This patch fixed my problem. > > Cc: Michal Marek > Signed-off-by: Andrew Morton > > diff --git a/Makefile b/Makefile > index 7431c28..13126ae 100644 > --- a/Makefile > +++ b/Makefile > @@ -190,7 +190,7 @@ SUBARCH := $(shell uname -m | sed -e s/i.86/i386/ -e s/sun4u/sparc64/ \ > export KBUILD_BUILDHOST := $(SUBARCH) > ARCH ?= $(SUBARCH) > CROSS_COMPILE ?= > -CROSS_COMPILE ?= $(CONFIG_CROSS_COMPILE:"%"=%) > +CROSS_COMPILE := $(shell grep CONFIG_CROSS_COMPILE .config | cut -d'"' -f2) So this patch does two things: 1) It grep for CONFIG_CROSS_COMPILE because we may not have incldued .config 2) It no longer pick up CROSS_COMPILE from the environment. Traditionally we have picked up ARCH and CROSS_COMPILE from the environmnet, hence the use of "?=". "?=" says - perfrom this assignmnet unless lhs is already defined. And lhs may be already defined is we have the ysmbol defined in the environmnet. I think the original submitter chenged from ?= to := without realising this. So the patch should be adjusted to use ?= again as you suggest before it is applied. I assume Michal / Andrew will take care. Subject on patch could also be better. Suggestion: kbuild: fix reading CROSS_COMPILE from .config Sam