From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755734Ab2IZMb5 (ORCPT ); Wed, 26 Sep 2012 08:31:57 -0400 Received: from cantor2.suse.de ([195.135.220.15]:36524 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753549Ab2IZMb4 (ORCPT ); Wed, 26 Sep 2012 08:31:56 -0400 Date: Wed, 26 Sep 2012 13:31:34 +0100 From: Mel Gorman To: Mark Brown Cc: Andrew Morton , Rik van Riel , Rafael Aquini , linux-kernel@vger.kernel.org Subject: Re: [PATCH] mm: Fix build with CMA && !CONFIG_COMPACTION Message-ID: <20120926123134.GJ11266@suse.de> References: <1348661586-22970-1-git-send-email-broonie@opensource.wolfsonmicro.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline In-Reply-To: <1348661586-22970-1-git-send-email-broonie@opensource.wolfsonmicro.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Sep 26, 2012 at 01:13:06PM +0100, Mark Brown wrote: > In -next 20120926 "mm: compaction: cache if a pageblock was scanned and > no pages were isolated" adds pageblock skipping operations used when CMA > is enabled but only provides them if CONFIG_COMPACTION is defined, breaking > the build. Fix this by also providing them when only CONFIG_CMA is enabled. > > Signed-off-by: Mark Brown Thanks but I've already pushed an alternative patch towards Andrew for this particular build problem. The patch explains why the fix is different. ---8<--- mm: compaction: cache if a pageblock was scanned and no pages were isolated -fix Fengguang Wu reported the following tree: git://git.kernel.org/pub/scm/linux/kernel/git/mhocko/mm.git since-3.5 head: 2f641f902ca76711e47e8d3b18004f0e46ca3d9b commit: 7faeb2a39c789f1bac69014cc468677a60b73395 [184/186] mm: compaction: cache if a pageblock was scanned and no pages were isolated config: i386-randconfig-b083 (attached as .config) All error/warnings: mm/compaction.c: In function 'isolation_suitable': mm/compaction.c:60:2: error: implicit declaration of function 'get_pageblock_skip' [-Werror=implicit-function-declaration] mm/compaction.c: In function 'reset_isolation_suitable': mm/compaction.c:94:3: error: implicit declaration of function 'clear_pageblock_skip' [-Werror=implicit-function-declaration] mm/compaction.c: In function 'update_pageblock_skip': mm/compaction.c:108:3: error: implicit declaration of function 'set_pageblock_skip' [-Werror=implicit-function-declaration] mm/compaction.c: At top level: mm/compaction.c:68:13: warning: 'reset_isolation_suitable' defined but not used [-Wunused-function] mm/compaction.c:177:13: warning: 'compact_capture_page' defined but not used [-Wunused-function] cc1: some warnings being treated as errors Michal Hocko suggested implementing !CONFIG_COMPACTION versions of these functions but that still leaves the dead version of reset_isolation_suitable in the !CONFIG_COMPACTION && CONFIG_CMA case. Create !CONFIG_COMPACTION versions of isolation_suitable() and update_pageblock_skip() instead. This is a fix for mm-compaction-cache-if-a-pageblock-was-scanned-and-no-pages-were-isolated.patch Reported-by: Fengguang Wu Signed-off-by: Mel Gorman --- mm/compaction.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/mm/compaction.c b/mm/compaction.c index 5ffe9a5..fb07abb 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -50,6 +50,7 @@ static inline bool migrate_async_suitable(int migratetype) return is_migrate_cma(migratetype) || migratetype == MIGRATE_MOVABLE; } +#ifdef CONFIG_COMPACTION /* Returns true if the pageblock should be scanned for pages to isolate. */ static inline bool isolation_suitable(struct compact_control *cc, struct page *page) @@ -107,6 +108,17 @@ static void update_pageblock_skip(struct page *page, unsigned long nr_isolated) if (!nr_isolated) set_pageblock_skip(page); } +#else +static inline bool isolation_suitable(struct compact_control *cc, + struct page *page) +{ + return true; +} + +static void update_pageblock_skip(struct page *page, unsigned long nr_isolated) +{ +} +#endif /* CONFIG_COMPACTION */ static inline bool should_release_lock(spinlock_t *lock) {