From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1162417Ab3DEScM (ORCPT ); Fri, 5 Apr 2013 14:32:12 -0400 Received: from mail-ye0-f174.google.com ([209.85.213.174]:34171 "EHLO mail-ye0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1162396Ab3DEScK (ORCPT ); Fri, 5 Apr 2013 14:32:10 -0400 Message-ID: <515F18A8.7030102@gmail.com> Date: Fri, 05 Apr 2013 14:32:08 -0400 From: KOSAKI Motohiro User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:17.0) Gecko/20130307 Thunderbird/17.0.4 MIME-Version: 1.0 To: Naoya Horiguchi CC: Andrew Morton , Mel Gorman , Hugh Dickins , Rik van Riel , KOSAKI Motohiro , Konstantin Khlebnikov , Michal Hocko , HATAYAMA Daisuke , linux-mm@kvack.org, linux-kernel@vger.kernel.org, kosaki.motohiro@gmail.com Subject: Re: [PATCH v3 1/3] hugetlbfs: stop setting VM_DONTDUMP in initializing vma(VM_HUGETLB) References: <1365014138-19589-1-git-send-email-n-horiguchi@ah.jp.nec.com> <1365014138-19589-2-git-send-email-n-horiguchi@ah.jp.nec.com> In-Reply-To: <1365014138-19589-2-git-send-email-n-horiguchi@ah.jp.nec.com> Content-Type: text/plain; charset=ISO-2022-JP Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org (4/3/13 2:35 PM), Naoya Horiguchi wrote: > Currently we fail to include any data on hugepages into coredump, > because VM_DONTDUMP is set on hugetlbfs's vma. This behavior was recently > introduced by commit 314e51b98 "mm: kill vma flag VM_RESERVED and > mm->reserved_vm counter". This looks to me a serious regression, > so let's fix it. I don't think this is enough explanations. Let's explain the code meaning time to time order. First, we had no madvice(DONTDUMP) nor coredump_filter(HUGETLB). then hugetlb pages were never dumped. Second, I added coredump_filter(HUGETLB). and then vm_dump_size became.. vm_dump_size() { /* Hugetlb memory check */ if (vma->vm_flags & VM_HUGETLB) { .. goto whole; } if (vma->vm_flags & VM_RESERVED) return 0; The point is, hugetlb was checked before VM_RESERVED. i.e. hugetlb core dump ignored VM_RESERVED. At this time, "if (vma->vm_flags & VM_HUGETLB)" statement don't need return 0 because VM_RESERVED prevented to go into the subsequent flag checks. Third, Jason added madvise(DONTDUMP). then vm_dump_size became... vm_dump_size() { if (vma->vm_flags & VM_NODUMP) return 0; /* Hugetlb memory check */ if (vma->vm_flags & VM_HUGETLB) { .. goto whole; } if (vma->vm_flags & VM_RESERVED) return 0; Look, VM_NODUMP and VM_RESERVED had similar and different meanings at this time. Finally, Konstantin removed VM_RESERVED and hugetlb coredump behavior has been changed. Thus, patch [1/3] and [2/3] should be marked [stable for v3.6 or later]. Anyway, this patch is correct. Thank you! Acked-by: KOSAKI Motohiro