From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from psmtp.com (na3sys010amx108.postini.com [74.125.245.108]) by kanga.kvack.org (Postfix) with SMTP id 6255E6B0078 for ; Mon, 14 Jan 2013 13:43:11 -0500 (EST) Date: Mon, 14 Jan 2013 19:43:08 +0100 From: Michal Hocko Subject: Re: [PATCH] memory-hotplug: revert register_page_bootmem_info_node() to empty when platform related code is not implemented Message-ID: <20130114184308.GD5126@dhcp22.suse.cz> References: <1358160835-30617-1-git-send-email-linfeng@cn.fujitsu.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1358160835-30617-1-git-send-email-linfeng@cn.fujitsu.com> Sender: owner-linux-mm@kvack.org List-ID: To: Lin Feng Cc: akpm@linux-foundation.org, linux-mm@kvack.org, wency@cn.fujitsu.com, jiang.liu@huawei.com, laijs@cn.fujitsu.com, isimatu.yasuaki@jp.fujitsu.com, linux-kernel@vger.kernel.org, tangchen@cn.fujitsu.com On Mon 14-01-13 18:53:55, Lin Feng wrote: > Memory-hotplug codes for x86_64 have been implemented by patchset: > https://lkml.org/lkml/2013/1/9/124 > While other platforms haven't been completely implemented yet. > > If we enable both CONFIG_MEMORY_HOTPLUG_SPARSE and CONFIG_SPARSEMEM_VMEMMAP, > register_page_bootmem_info_node() may be buggy, which is a hotplug generic > function but falling back to call platform related function > register_page_bootmem_memmap(). > > Other platforms such as powerpc it's not implemented, so on such platforms, > revert them as empty as they were before. > > Reported-by: Michal Hocko > Signed-off-by: Lin Feng > --- > mm/memory_hotplug.c | 9 +++++++++ > 1 files changed, 9 insertions(+), 0 deletions(-) > > diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c > index 8aa2b56..bd93c2e 100644 > --- a/mm/memory_hotplug.c > +++ b/mm/memory_hotplug.c > @@ -189,6 +189,7 @@ static void register_page_bootmem_info_section(unsigned long start_pfn) > } > #endif > > +#ifdef CONFIG_X86_64 > void register_page_bootmem_info_node(struct pglist_data *pgdat) > { > unsigned long i, pfn, end_pfn, nr_pages; > @@ -230,6 +231,14 @@ void register_page_bootmem_info_node(struct pglist_data *pgdat) > register_page_bootmem_info_section(pfn); > } > } > +#else > +static inline void register_page_bootmem_info_node(struct pglist_data *pgdat) > +{ > + /* > + * Todo: platforms other than X86_64 haven't been implemented yet. > + */ > +} > +#endif This is just ugly. Could you please add something like HAVE_BOOTMEM_INFO_NODE or something with a bettern name and let CONFIG_MEMORY_HOTPLUG select it for supported architectures and configurations (e.g. CONFIG_SPARSEMEM_VMEMMAP doesn't need a special arch support, right?). These Todo things are just too messy. -- Michal Hocko SUSE Labs -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org