From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758602Ab3GPB4B (ORCPT ); Mon, 15 Jul 2013 21:56:01 -0400 Received: from LGEMRELSE6Q.lge.com ([156.147.1.121]:53018 "EHLO LGEMRELSE6Q.lge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754606Ab3GPBz7 (ORCPT ); Mon, 15 Jul 2013 21:55:59 -0400 X-AuditID: 9c930179-b7c49ae000000e68-b1-51e4a82d5a8b Date: Tue, 16 Jul 2013 10:56:00 +0900 From: Joonsoo Kim To: "Aneesh Kumar K.V" Cc: Andrew Morton , Rik van Riel , Mel Gorman , Michal Hocko , KAMEZAWA Hiroyuki , Hugh Dickins , Davidlohr Bueso , David Gibson , linux-mm@kvack.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH 6/9] mm, hugetlb: do not use a page in page cache for cow optimization Message-ID: <20130716015600.GH2430@lge.com> References: <1373881967-16153-1-git-send-email-iamjoonsoo.kim@lge.com> <1373881967-16153-7-git-send-email-iamjoonsoo.kim@lge.com> <87d2qkj5b7.fsf@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <87d2qkj5b7.fsf@linux.vnet.ibm.com> User-Agent: Mutt/1.5.21 (2010-09-15) X-Brightmail-Tracker: AAAAAA== Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, Jul 15, 2013 at 07:25:40PM +0530, Aneesh Kumar K.V wrote: > Joonsoo Kim writes: > > > Currently, we use a page with mapped count 1 in page cache for cow > > optimization. If we find this condition, we don't allocate a new > > page and copy contents. Instead, we map this page directly. > > This may introduce a problem that writting to private mapping overwrite > > hugetlb file directly. You can find this situation with following code. > > > > size = 20 * MB; > > flag = MAP_SHARED; > > p = mmap(NULL, size, PROT_READ|PROT_WRITE, flag, fd, 0); > > if (p == MAP_FAILED) { > > fprintf(stderr, "mmap() failed: %s\n", strerror(errno)); > > return -1; > > } > > p[0] = 's'; > > fprintf(stdout, "BEFORE STEAL PRIVATE WRITE: %c\n", p[0]); > > munmap(p, size); > > > > flag = MAP_PRIVATE; > > p = mmap(NULL, size, PROT_READ|PROT_WRITE, flag, fd, 0); > > if (p == MAP_FAILED) { > > fprintf(stderr, "mmap() failed: %s\n", strerror(errno)); > > } > > p[0] = 'c'; > > munmap(p, size); > > > > flag = MAP_SHARED; > > p = mmap(NULL, size, PROT_READ|PROT_WRITE, flag, fd, 0); > > if (p == MAP_FAILED) { > > fprintf(stderr, "mmap() failed: %s\n", strerror(errno)); > > return -1; > > } > > fprintf(stdout, "AFTER STEAL PRIVATE WRITE: %c\n", p[0]); > > munmap(p, size); > > > > We can see that "AFTER STEAL PRIVATE WRITE: c", not "AFTER STEAL > > PRIVATE WRITE: s". If we turn off this optimization to a page > > in page cache, the problem is disappeared. > > Do we need to trun of the optimization for page cache completely ? > Can't we check for MAP_PRIVATE case ? IMO, we need to turn off the optimization for page cache completely. This optimization works just for MAP_PRIVATE case. If we map with MAP_SHARED, we map a page from page cache directly, so we don't need this cow optimization. > > Also, we may want to add the above test into libhugetlbfs I will submit the above test into libhugetlbfs. Thanks. > > > > > Signed-off-by: Joonsoo Kim > > > > diff --git a/mm/hugetlb.c b/mm/hugetlb.c > > index d4a1695..6c1eb9b 100644 > > --- a/mm/hugetlb.c > > +++ b/mm/hugetlb.c > > @@ -2512,7 +2512,6 @@ static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma, > > { > > struct hstate *h = hstate_vma(vma); > > struct page *old_page, *new_page; > > - int avoidcopy; > > int outside_reserve = 0; > > unsigned long mmun_start; /* For mmu_notifiers */ > > unsigned long mmun_end; /* For mmu_notifiers */ > > @@ -2522,10 +2521,8 @@ static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma, > > retry_avoidcopy: > > /* If no-one else is actually using this page, avoid the copy > > * and just make the page writable */ > > - avoidcopy = (page_mapcount(old_page) == 1); > > - if (avoidcopy) { > > - if (PageAnon(old_page)) > > - page_move_anon_rmap(old_page, vma, address); > > + if (page_mapcount(old_page) == 1 && PageAnon(old_page)) { > > + page_move_anon_rmap(old_page, vma, address); > > set_huge_ptep_writable(vma, address, ptep); > > return 0; > > } > > -aneesh > > -- > 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 From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from psmtp.com (na3sys010amx126.postini.com [74.125.245.126]) by kanga.kvack.org (Postfix) with SMTP id 6F4496B0032 for ; Mon, 15 Jul 2013 21:55:59 -0400 (EDT) Date: Tue, 16 Jul 2013 10:56:00 +0900 From: Joonsoo Kim Subject: Re: [PATCH 6/9] mm, hugetlb: do not use a page in page cache for cow optimization Message-ID: <20130716015600.GH2430@lge.com> References: <1373881967-16153-1-git-send-email-iamjoonsoo.kim@lge.com> <1373881967-16153-7-git-send-email-iamjoonsoo.kim@lge.com> <87d2qkj5b7.fsf@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <87d2qkj5b7.fsf@linux.vnet.ibm.com> Sender: owner-linux-mm@kvack.org List-ID: To: "Aneesh Kumar K.V" Cc: Andrew Morton , Rik van Riel , Mel Gorman , Michal Hocko , KAMEZAWA Hiroyuki , Hugh Dickins , Davidlohr Bueso , David Gibson , linux-mm@kvack.org, linux-kernel@vger.kernel.org On Mon, Jul 15, 2013 at 07:25:40PM +0530, Aneesh Kumar K.V wrote: > Joonsoo Kim writes: > > > Currently, we use a page with mapped count 1 in page cache for cow > > optimization. If we find this condition, we don't allocate a new > > page and copy contents. Instead, we map this page directly. > > This may introduce a problem that writting to private mapping overwrite > > hugetlb file directly. You can find this situation with following code. > > > > size = 20 * MB; > > flag = MAP_SHARED; > > p = mmap(NULL, size, PROT_READ|PROT_WRITE, flag, fd, 0); > > if (p == MAP_FAILED) { > > fprintf(stderr, "mmap() failed: %s\n", strerror(errno)); > > return -1; > > } > > p[0] = 's'; > > fprintf(stdout, "BEFORE STEAL PRIVATE WRITE: %c\n", p[0]); > > munmap(p, size); > > > > flag = MAP_PRIVATE; > > p = mmap(NULL, size, PROT_READ|PROT_WRITE, flag, fd, 0); > > if (p == MAP_FAILED) { > > fprintf(stderr, "mmap() failed: %s\n", strerror(errno)); > > } > > p[0] = 'c'; > > munmap(p, size); > > > > flag = MAP_SHARED; > > p = mmap(NULL, size, PROT_READ|PROT_WRITE, flag, fd, 0); > > if (p == MAP_FAILED) { > > fprintf(stderr, "mmap() failed: %s\n", strerror(errno)); > > return -1; > > } > > fprintf(stdout, "AFTER STEAL PRIVATE WRITE: %c\n", p[0]); > > munmap(p, size); > > > > We can see that "AFTER STEAL PRIVATE WRITE: c", not "AFTER STEAL > > PRIVATE WRITE: s". If we turn off this optimization to a page > > in page cache, the problem is disappeared. > > Do we need to trun of the optimization for page cache completely ? > Can't we check for MAP_PRIVATE case ? IMO, we need to turn off the optimization for page cache completely. This optimization works just for MAP_PRIVATE case. If we map with MAP_SHARED, we map a page from page cache directly, so we don't need this cow optimization. > > Also, we may want to add the above test into libhugetlbfs I will submit the above test into libhugetlbfs. Thanks. > > > > > Signed-off-by: Joonsoo Kim > > > > diff --git a/mm/hugetlb.c b/mm/hugetlb.c > > index d4a1695..6c1eb9b 100644 > > --- a/mm/hugetlb.c > > +++ b/mm/hugetlb.c > > @@ -2512,7 +2512,6 @@ static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma, > > { > > struct hstate *h = hstate_vma(vma); > > struct page *old_page, *new_page; > > - int avoidcopy; > > int outside_reserve = 0; > > unsigned long mmun_start; /* For mmu_notifiers */ > > unsigned long mmun_end; /* For mmu_notifiers */ > > @@ -2522,10 +2521,8 @@ static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma, > > retry_avoidcopy: > > /* If no-one else is actually using this page, avoid the copy > > * and just make the page writable */ > > - avoidcopy = (page_mapcount(old_page) == 1); > > - if (avoidcopy) { > > - if (PageAnon(old_page)) > > - page_move_anon_rmap(old_page, vma, address); > > + if (page_mapcount(old_page) == 1 && PageAnon(old_page)) { > > + page_move_anon_rmap(old_page, vma, address); > > set_huge_ptep_writable(vma, address, ptep); > > return 0; > > } > > -aneesh > > -- > 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 -- 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