From mboxrd@z Thu Jan 1 00:00:00 1970 From: wujianguo Date: Fri, 24 Aug 2012 02:31:15 +0000 Subject: Re: [PATCH]mm: fix-up zone present pages Message-Id: <5036E773.6020609@gmail.com> List-Id: References: <5031DB52.9030806@gmail.com> <201208210555.41312.ptesarik@suse.cz> In-Reply-To: <201208210555.41312.ptesarik@suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset="windows-1252" Content-Transfer-Encoding: quoted-printable To: Petr Tesarik Cc: tony.luck@intel.com, fenghua.yu@intel.com, dhowells@redhat.com, tj@kernel.org, mgorman@suse.de, yinghai@kernel.org, minchan.kim@gmail.com, akpm@linux-foundation.org, viro@zeniv.linux.org.uk, aarcange@redhat.com, davem@davemloft.net, hannes@cmpxchg.org, liuj97@gmail.com, wency@cn.fujitsu.com, rientjes@google.com, kamezawa.hiroyu@jp.fujitsu.com, mhocko@suse.cz, linux-ia64@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, jiang.liu@huawei.com, guohanjun@huawei.com, qiuxishi@huawei.com On 2012-8-21 11:55, Petr Tesarik wrote: > Dne Po 20. srpna 2012 08:38:10 wujianguo napsal(a): >> From: Jianguo Wu >> >> Hi all, >> I think zone->present_pages indicates pages that buddy system can >> management, it should be: >> zone->present_pages =3D spanned pages - absent pages - bootmem pages, >> but now: >> zone->present_pages =3D spanned pages - absent pages - memmap pages. >> spanned pages=EF=BC=9Atotal size, including holes. >> absent pages: holes. >> bootmem pages: pages used in system boot, managed by bootmem allocator. >> memmap pages: pages used by page structs. >=20 > Absolutely. The memory allocated to page structs should be counted in. >=20 Hi Petr, Bootmem pages include the memory allocated to page structs. Thanks! Jianguo Wu >> This may cause zone->present_pages less than it should be. >> For example, numa node 1 has ZONE_NORMAL and ZONE_MOVABLE, >> it's memmap and other bootmem will be allocated from ZONE_MOVABLE, >> so ZONE_NORMAL's present_pages should be spanned pages - absent pages, >> but now it also minus memmap pages(free_area_init_core), which are actua= lly >> allocated from ZONE_MOVABLE. When offline all memory of a zone, This will >> cause zone->present_pages less than 0, because present_pages is unsigned >> long type, it is actually a very large integer, it indirectly caused >> zone->watermark[WMARK_MIN] become a large >> integer(setup_per_zone_wmarks()), than cause totalreserve_pages become a >> large integer(calculate_totalreserve_pages()), and finally cause memory >> allocating failure when fork process(__vm_enough_memory()). >> >> [root@localhost ~]# dmesg >> -bash: fork: Cannot allocate memory >> >> I think bug described in http://marc.info/?l=3Dlinux-mm&m=13450218271418= 6&w=3D2 >> is also caused by wrong zone present pages. >=20 > And yes, I can confirm that the bug I reported is caused by a too low num= ber=20 > for the present pages counter. Your patch does fix the bug for me. >=20 > Thanks! > Petr Tesarik >=20