From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pa0-f42.google.com (mail-pa0-f42.google.com [209.85.220.42]) by kanga.kvack.org (Postfix) with ESMTP id EEFDF6B00E8 for ; Fri, 22 May 2015 01:10:16 -0400 (EDT) Received: by pabru16 with SMTP id ru16so10033481pab.1 for ; Thu, 21 May 2015 22:10:16 -0700 (PDT) Received: from tyo202.gate.nec.co.jp (TYO202.gate.nec.co.jp. [210.143.35.52]) by mx.google.com with ESMTPS id lj8si1715108pbc.11.2015.05.21.22.10.15 for (version=TLSv1 cipher=RC4-SHA bits=128/128); Thu, 21 May 2015 22:10:16 -0700 (PDT) From: Naoya Horiguchi Subject: Re: [PATCH] hugetlb: Do not account hugetlb pages as NR_FILE_PAGES Date: Fri, 22 May 2015 05:09:34 +0000 Message-ID: <20150522050934.GA24376@hori1.linux.bs1.fc.nec.co.jp> References: <1432214842-22730-1-git-send-email-mhocko@suse.cz> In-Reply-To: <1432214842-22730-1-git-send-email-mhocko@suse.cz> Content-Language: ja-JP Content-Type: text/plain; charset="iso-2022-jp" Content-ID: Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Sender: owner-linux-mm@kvack.org List-ID: To: Michal Hocko Cc: Andrew Morton , Mel Gorman , "linux-mm@kvack.org" , LKML On Thu, May 21, 2015 at 03:27:22PM +0200, Michal Hocko wrote: > hugetlb pages uses add_to_page_cache to track shared mappings. This > is OK from the data structure point of view but it is less so from the > NR_FILE_PAGES accounting: > - huge pages are accounted as 4k which is clearly wrong > - this counter is used as the amount of the reclaimable page > cache which is incorrect as well because hugetlb pages are > special and not reclaimable > - the counter is then exported to userspace via /proc/meminfo > (in Cached:), /proc/vmstat and /proc/zoneinfo as > nr_file_pages which is confusing at least: > Cached: 8883504 kB > HugePages_Free: 8348 > ... > Cached: 8916048 kB > HugePages_Free: 156 > ... > thats 8192 huge pages allocated which is ~16G accounted as 32M >=20 > There are usually not that many huge pages in the system for this to > make any visible difference e.g. by fooling __vm_enough_memory or > zone_pagecache_reclaimable. >=20 > Fix this by special casing huge pages in both __delete_from_page_cache > and __add_to_page_cache_locked. replace_page_cache_page is currently > only used by fuse and that shouldn't touch hugetlb pages AFAICS but it > is more robust to check for special casing there as well. >=20 > Hugetlb pages shouldn't get to any other paths where we do accounting: > - migration - we have a special handling via > hugetlbfs_migrate_page > - shmem - doesn't handle hugetlb pages directly even for > SHM_HUGETLB resp. MAP_HUGETLB > - swapcache - hugetlb is not swapable >=20 > This has a user visible effect but I believe it is reasonable because > the previously exported number is simply bogus. >=20 > An alternative would be to account hugetlb pages with their real size > and treat them similar to shmem. But this has some drawbacks. >=20 > First we would have to special case in kernel users of NR_FILE_PAGES and > considering how hugetlb is special we would have to do it everywhere. We > do not want Cached exported by /proc/meminfo to include it because the > value would be even more misleading. > __vm_enough_memory and zone_pagecache_reclaimable would have to do > the same thing because those pages are simply not reclaimable. The > correction is even not trivial because we would have to consider all > active hugetlb page sizes properly. Users of the counter outside of the > kernel would have to do the same. > So the question is why to account something that needs to be basically > excluded for each reasonable usage. This doesn't make much sense to me. >=20 > It seems that this has been broken since hugetlb was introduced but I > haven't checked the whole history. >=20 > Signed-off-by: Michal Hocko looks good to me, Reviewed-by: Naoya Horiguchi = -- 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