From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754107AbbDRQRN (ORCPT ); Sat, 18 Apr 2015 12:17:13 -0400 Received: from aserp1040.oracle.com ([141.146.126.69]:25803 "EHLO aserp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753960AbbDRQRH (ORCPT ); Sat, 18 Apr 2015 12:17:07 -0400 Message-ID: <55328375.20407@oracle.com> Date: Sat, 18 Apr 2015 09:16:53 -0700 From: Mike Kravetz User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.5.0 MIME-Version: 1.0 To: Hillf Danton , "'Dave Hansen'" CC: linux-kernel , linux-mm@kvack.org, Michal Hocko Subject: Re: [RFC PATCH 4/4] mm: madvise allow remove operation for hugetlbfs References: <00e601d078da$9e762190$db6264b0$@alibaba-inc.com> <00ef01d078dd$96bfc480$c43f4d80$@alibaba-inc.com> <55313ECD.3050604@oracle.com> In-Reply-To: <55313ECD.3050604@oracle.com> Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit X-Source-IP: userv0021.oracle.com [156.151.31.71] Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 04/17/2015 10:11 AM, Mike Kravetz wrote: > On 04/17/2015 12:10 AM, Hillf Danton wrote: >>> >>> Now that we have hole punching support for hugetlbfs, we can >>> also support the MADV_REMOVE interface to it. >>> >>> Signed-off-by: Dave Hansen >>> Signed-off-by: Mike Kravetz >>> --- >>> mm/madvise.c | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/mm/madvise.c b/mm/madvise.c >>> index d551475..c4a1027 100644 >>> --- a/mm/madvise.c >>> +++ b/mm/madvise.c >>> @@ -299,7 +299,7 @@ static long madvise_remove(struct vm_area_struct >>> *vma, >>> >>> *prev = NULL; /* tell sys_madvise we drop mmap_sem */ >>> >>> - if (vma->vm_flags & (VM_LOCKED | VM_HUGETLB)) >>> + if (vma->vm_flags & VM_LOCKED) >>> return -EINVAL; >>> >>> f = vma->vm_file; >>> -- >>> 2.1.0 >> >> After the above change offset is computed, >> >> offset = (loff_t)(start - vma->vm_start) >> + ((loff_t)vma->vm_pgoff << PAGE_SHIFT); >> >> and I wonder if it is correct for huge page mapping. > > I think it will be correct. > > The above will be a (base) page size aligned offset into the file. > This offset will be huge page aligned in the fallocate hole punch > code. > > /* > * For hole punch round up the beginning offset of the hole and > * round down the end. > */ > hole_start = (offset + hpage_size - 1) & ~huge_page_mask(h); > hole_end = (offset + len - (hpage_size - 1)) * ~huge_page_mask(h); > > Was the alignment your concern, or something else? Well, that alignment code in fallocate hole punch obviously wrong. :( Sorry about that. I'll send out an updated RFC with working hole punch. -- Mike Kravetz